From owner-pro64-support@oss.sgi.com Thu Mar 1 10:31:05 2001 Received: by oss.sgi.com id ; Thu, 1 Mar 2001 10:30:55 -0800 Received: from beta.dmz-eu.st.com ([164.129.1.35]:61154 "HELO beta.dmz-eu.st.com") by oss.sgi.com with SMTP id ; Thu, 1 Mar 2001 10:30:41 -0800 Received: from zeta.dmz-eu.st.com (zeta.dmz-eu.st.com [164.129.230.9]) by beta.dmz-eu.st.com (STMicroelectronics) with SMTP id EC8D64A2C for ; Thu, 1 Mar 2001 18:30:32 +0000 (GMT) Received: by zeta.dmz-eu.st.com (STMicroelectronics, from userid 0) id 8EAB14C4E; Thu, 1 Mar 2001 18:28:34 +0000 (GMT) Received: from thistle.bri.st.com (localhost [127.0.0.1]) by zeta.dmz-eu.st.com (STMicroelectronics) with ESMTP id 0BB91184B for ; Thu, 1 Mar 2001 18:28:34 +0000 (GMT) Received: from [164.129.14.43] (helo=absolut) by thistle.bristol.st.com with esmtp (Exim 3.03 #5) id 14YXoN-00017K-00 for pro64-support@oss.sgi.com; Thu, 01 Mar 2001 18:28:23 +0000 Received: from bowersa (helo=localhost) by absolut with local-esmtp (Exim 3.03 #5) id 14YXoM-0005RL-00 for pro64-support@oss.sgi.com; Thu, 01 Mar 2001 18:28:22 +0000 Date: Thu, 1 Mar 2001 18:28:21 +0000 (GMT) From: Antony Bowers X-Sender: bowersa@absolut To: pro64-support@oss.sgi.com Subject: Build tools? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing How do people build the Pro64 compiler from source, on Linux? I'm having trouble finding a workable method (I'm looking into adding a new target). The README says that gcc 2.96 (experimental) from November 1999 is known to work. Presumably it's possible to fetch and build a gcc of this vintage from the CVS repository, but I don't have easy access to it. That leaves either gcc 2.95.2 or a recent gcc snapshot (pre-3.0) as possibilities. We have established that gcc 2.95.2 will build the compiler with fairly minor modifications to the source; but depending forever on a compiler that will soon be superceded is not ideal. I've since tried building with the pre-3.0 gcc of 20010212, but there are difficulties. 1. The Pro64 source depends on SGI STL extensions, but some time after 2.96 gcc dropped its SGI-based STL. The experimental iostreams version of the STL from SGI can be used to fill the gap though, with some fiddling about. 2. A few code modifications are needed where the Pro64 C++ source is not standards conforming (the new gcc is a bit closer to enforcing the language standard). I haven't got this build to go through yet; currently there are mysterious problems linking "be". I'd prefer to avoid the forked version of gcc 2.96 that comes with Red Hat 7.0; that compiler is rumoured to be buggy and is surely no less doomed to obsolescence than 2.95. Is anything being done to ensure that Pro64 will remain buildable on x86 Linux in the long term? What is the right approach? Thanks for any suggestions, Tony -- Antony Bowers, STMicroelectronics, Bristol, UK From owner-pro64-support@oss.sgi.com Fri Mar 2 22:46:37 2001 Received: by oss.sgi.com id ; Fri, 2 Mar 2001 22:46:28 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:33294 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Fri, 2 Mar 2001 22:46:03 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Fri, 2 Mar 2001 23:45:31 -0700 Date: Fri, 2 Mar 2001 23:45:30 -0700 (MST) From: Peng Zhao To: sgi Subject: Some error messages and questions Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing HI, 1. When running "sgicc -Ofast test.c", I was responded with following messages: "sgicc ERROR: cannot exec /usr/lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/ipa_link" When using "sgicc -O3 -INLINE:must=callit:list test.c", the thing is ok. WHat is the reason? Should I use "-Y" things? 2. I also meet this kind of error messages: " tmp/ccs.Q9rHEG:285: Error: no such 386 instruction: `data1 0x01,0x11,0x01,0x03,0x08,0x25,0x08,0x13' /tmp/ccs.Q9rHEG:286: Error: no such 386 instruction: `data1 0x0b,0x42,0x0b,0x10,0x06,0x00,0x00,0x02' /tmp/ccs.Q9rHEG:287: Error: no such 386 instruction: `data1 0x2e,0x00,0x3a,0x0b,0x " 3. How to check the result of function inlining? If I switch on the "-S" or "-keep" option, the IPA will not be applied, am I right? Thanks Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Sat Mar 3 06:09:00 2001 Received: by oss.sgi.com id ; Sat, 3 Mar 2001 06:08:51 -0800 Received: from lslsun.epfl.ch ([128.178.150.20]:3010 "EHLO lslsun.epfl.ch") by oss.sgi.com with ESMTP id ; Sat, 3 Mar 2001 06:08:35 -0800 Received: from epfl.ch (johaenni@lslsun8 [128.178.150.31]) by lslsun.epfl.ch (8.8.X/EPFL-8.1a) with ESMTP id PAA03696; Sat, 3 Mar 2001 15:08:29 +0100 (MET) Message-ID: <3AA0FADC.404A8127@epfl.ch> Date: Sat, 03 Mar 2001 15:08:28 +0100 From: Jacques-Olivier Haenni Organization: Swiss Federal Institute of Technology, Lausanne (EPFL) X-Mailer: Mozilla 4.75 [en] (X11; U; SunOS 5.6 sun4u) X-Accept-Language: en, fr-CH, fr MIME-Version: 1.0 To: Peng Zhao CC: sgi Subject: Re: Some error messages and questions References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi, Peng Zhao wrote: > > HI, > > 1. When running "sgicc -Ofast test.c", I was responded with following > messages: > "sgicc ERROR: cannot exec > /usr/lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/ipa_link" > > When using "sgicc -O3 -INLINE:must=callit:list test.c", the thing > is ok. > > WHat is the reason? Should I use "-Y" things? On February 11th, Mike Murphy wrote : : -Ofast is not supported yet. It uses ipa, which has not been released yet : (but will be soon). Use -O3 for now. This answers the first question. > 2. I also meet this kind of error messages: > " tmp/ccs.Q9rHEG:285: Error: no such 386 instruction: `data1 > 0x01,0x11,0x01,0x03,0x08,0x25,0x08,0x13' > > /tmp/ccs.Q9rHEG:286: Error: no > such 386 instruction: `data1 0x0b,0x42,0x0b,0x10,0x06,0x00,0x00,0x02' > > /tmp/ccs.Q9rHEG:287: Error: no such 386 instruction: `data1 > 0x2e,0x00,0x3a,0x0b,0x > " sgicc generates an assembly file which is then processed by /usr/bin/as. If you run sgicc under Linux on a IA-32 machine, /usr/bin/as is the IA-32 assembler and you'll get the above-reproduced messages. A solution could be to run sgicc under nue (or, even better, to have an Itanium machine...). However, on my Linux machine, I am not able to run sgicc (the one compiled on my Linux IA-32 machine) under nue, because of a problem with glibc versions. What I do now is to compile sgicc on my IA-32 linux machine. Then, I use it to generate assembly files (sgicc -S toto.c); next, I copy the assembly files on an Itanium machine (some are available on the web) where I assemble and link them. Has somebody a better solution ? > 3. How to check the result of function inlining? If I switch on the > "-S" or "-keep" option, the IPA will not be applied, am I right? IPA is never applied (yet), isn't it ? If some part of my answer is wrong, I request the Pro64 gurus to correct me... Jacques-Olivier -- Jacques-Olivier Haenni http://lslwww.epfl.ch/~johaenni/ Jacques-Olivier.Haenni@epfl.ch Logic Systems Laboratory Swiss Federal Institute of Technology (EPFL) | Tel: (+41 21) 693 66 30 1015 Lausanne - Switzerland | Fax: (+41 21) 693 37 05 From owner-pro64-support@oss.sgi.com Sat Mar 3 16:04:50 2001 Received: by oss.sgi.com id ; Sat, 3 Mar 2001 16:04:41 -0800 Received: from fmfdns02.fm.intel.com ([132.233.247.11]:30192 "EHLO thalia.fm.intel.com") by oss.sgi.com with ESMTP id ; Sat, 3 Mar 2001 16:04:21 -0800 Received: from SMTP (fmsmsxvs01-1.fm.intel.com [132.233.42.201]) by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id AAA11998; Sun, 4 Mar 2001 00:03:52 GMT Received: from fmsmsx17.intel.com ([132.233.48.17]) by 132.233.48.201 (Norton AntiVirus for Internet Email Gateways 1.0) ; Sun, 04 Mar 2001 00:03:52 0000 (GMT) Received: by fmsmsx17.fm.intel.com with Internet Mail Service (5.5.2650.21) id ; Sat, 3 Mar 2001 16:03:50 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F05B1AC16@FMSMSX41> From: "Chan, Sun C" To: Jacques-Olivier Haenni , Peng Zhao Cc: sgi Subject: RE: Some error messages and questions Date: Sat, 3 Mar 2001 16:03:49 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing to do cross compile, do "setenv TOOLROOT "path of sgicc binary", then $TOOLROOT/usr/bin/sgicc file.c will do. To debug, set LD_LIBRARY_PATH to where your backend components are. > -----Original Message----- > From: Jacques-Olivier Haenni [mailto:Jacques-Olivier.Haenni@epfl.ch] > Sent: Saturday, March 03, 2001 6:08 AM > To: Peng Zhao > Cc: sgi > Subject: Re: Some error messages and questions > > > Hi, > > Peng Zhao wrote: > > > > HI, > > > > 1. When running "sgicc -Ofast test.c", I was responded > with following > > messages: > > "sgicc ERROR: cannot exec > > /usr/lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/ipa_link" > > > > When using "sgicc -O3 -INLINE:must=callit:list > test.c", the thing > > is ok. > > > > WHat is the reason? Should I use "-Y" things? > > > On February 11th, Mike Murphy wrote : > > : -Ofast is not supported yet. It uses ipa, which has not > been released yet > : (but will be soon). Use -O3 for now. > > This answers the first question. > > > > 2. I also meet this kind of error messages: > > " tmp/ccs.Q9rHEG:285: Error: no such 386 instruction: `data1 > > 0x01,0x11,0x01,0x03,0x08,0x25,0x08,0x13' > > > > /tmp/ccs.Q9rHEG:286: Error: no > > such 386 instruction: `data1 > 0x0b,0x42,0x0b,0x10,0x06,0x00,0x00,0x02' > > > > /tmp/ccs.Q9rHEG:287: Error: no such 386 instruction: `data1 > > 0x2e,0x00,0x3a,0x0b,0x > > " > > > sgicc generates an assembly file which is then processed by > /usr/bin/as. If > you run sgicc under Linux on a IA-32 machine, /usr/bin/as is the IA-32 > assembler and you'll get the above-reproduced messages. > > A solution could be to run sgicc under nue (or, even better, > to have an Itanium > machine...). However, on my Linux machine, I am not able to > run sgicc (the one > compiled on my Linux IA-32 machine) under nue, because of a > problem with glibc > versions. > > What I do now is to compile sgicc on my IA-32 linux machine. > Then, I use it to > generate assembly files (sgicc -S toto.c); next, I copy the > assembly files on > an Itanium machine (some are available on the web) where I > assemble and link > them. Has somebody a better solution ? > > > > 3. How to check the result of function inlining? If I > switch on the > > "-S" or "-keep" option, the IPA will not be applied, am I right? > > IPA is never applied (yet), isn't it ? > > If some part of my answer is wrong, I request the Pro64 gurus > to correct me... > > Jacques-Olivier > > -- > Jacques-Olivier Haenni http://lslwww.epfl.ch/~johaenni/ Jacques-Olivier.Haenni@epfl.ch Logic Systems Laboratory Swiss Federal Institute of Technology (EPFL) | Tel: (+41 21) 693 66 30 1015 Lausanne - Switzerland | Fax: (+41 21) 693 37 05 From owner-pro64-support@oss.sgi.com Sat Mar 3 23:19:46 2001 Received: by oss.sgi.com id ; Sat, 3 Mar 2001 23:19:36 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:53310 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Sat, 3 Mar 2001 23:19:19 -0800 Received: from rohi.engr.sgi.com (rohi.engr.sgi.com [130.62.180.74]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id XAA06217 for ; Sat, 3 Mar 2001 23:28:59 -0800 (PST) mail_from (mpm@rohi.engr.sgi.com) Received: (from mpm@localhost) by rohi.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id XAA87765; Sat, 3 Mar 2001 23:17:26 -0800 (PST) Date: Sat, 3 Mar 2001 23:17:26 -0800 (PST) From: mpm@rohi.engr.sgi.com (Michael Murphy) Message-Id: <200103040717.XAA87765@rohi.engr.sgi.com> To: Peng Zhao , Jacques-Olivier Haenni Subject: Re: Some error messages and questions Cc: sgi References: Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing From: Jacques-Olivier Haenni What I do now is to compile sgicc on my IA-32 linux machine. Then, I use it to generate assembly files (sgicc -S toto.c); next, I copy the assembly files on an Itanium machine (some are available on the web) where I assemble and link them. Has somebody a better solution ? You can build a cross gas by doing something like: configure --host=i586-linux-gnu --target=ia64-sgi-linux and then using -Ya to use this gas from sgicc. This is basically what we do in our cross environment, except we put everything into a cross toolroot so that the -Y isn't needed. You can do the same for ld too, except you'll also need proper libraries, which gets more difficult. > 3. How to check the result of function inlining? If I switch on the > "-S" or "-keep" option, the IPA will not be applied, am I right? IPA is never applied (yet), isn't it ? -S and -keep do not prevent IPA, but IPA is not released yet. With -O you will get inlining within a file, just no cross-file inlining. -- Mike Murphy -- mpm@sgi.com -- quote of the day: -- "You can choose your choices, but you can't choose the consequences -- of your choices." From owner-pro64-support@oss.sgi.com Sun Mar 4 18:00:25 2001 Received: by oss.sgi.com id ; Sun, 4 Mar 2001 18:00:15 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:9990 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Sun, 4 Mar 2001 17:59:54 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Sun, 4 Mar 2001 18:59:39 -0700 Date: Sun, 4 Mar 2001 18:59:39 -0700 (MST) From: Peng Zhao To: Michael Murphy cc: Jacques-Olivier Haenni , sgi Subject: Re: Some error messages and questions In-Reply-To: <200103040717.XAA87765@rohi.engr.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Thanks for your response. I am confused by the error messages encountered these days. I use "make build -fMake.cross" and "INSTALL" to generate the sgicc. Should I run sgicc under "nue"? Now that it is a cross compiler, I had thought I can run sgicc directly without nue but run the generated code under NUE. Let's look at the problem in my sgicc under NUE. following is the error message: sgicc test.c sgicc: /usr/ia32-pc-linux/lib/libc.so.6: version `GLIBC_2.1.3' not found (required by sgicc) Thanks. Peng On Sat, 3 Mar 2001, Michael Murphy wrote: Jacques-Olivier Haenni platforms but they produce IA-64 code.> > > What I do now is to compile sgicc on my IA-32 linux machine. Then, I use it to > generate assembly files (sgicc -S toto.c); next, I copy the assembly files on > an Itanium machine (some are available on the web) where I assemble and link > them. Has somebody a better solution ? > > You can build a cross gas by doing something like: > configure --host=i586-linux-gnu --target=ia64-sgi-linux > and then using -Ya to use this gas from sgicc. > This is basically what we do in our cross environment, except > we put everything into a cross toolroot so that the -Y isn't needed. > You can do the same for ld too, except you'll also need proper libraries, > which gets more difficult. > > > 3. How to check the result of function inlining? If I switch on the > > "-S" or "-keep" option, the IPA will not be applied, am I right? > > IPA is never applied (yet), isn't it ? > > -S and -keep do not prevent IPA, but IPA is not released yet. > With -O you will get inlining within a file, just no cross-file inlining. > > -- Mike Murphy > -- mpm@sgi.com > -- quote of the day: > -- "You can choose your choices, but you can't choose the consequences > -- of your choices." > Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Mon Mar 5 09:48:21 2001 Received: by oss.sgi.com id ; Mon, 5 Mar 2001 09:48:01 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:2323 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Mon, 5 Mar 2001 09:47:35 -0800 Received: from rohi.engr.sgi.com (rohi.engr.sgi.com [130.62.180.74]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA05067 for ; Mon, 5 Mar 2001 09:57:17 -0800 (PST) mail_from (mpm@rohi.engr.sgi.com) Received: (from mpm@localhost) by rohi.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id JAA93353; Mon, 5 Mar 2001 09:45:47 -0800 (PST) Date: Mon, 5 Mar 2001 09:45:47 -0800 (PST) From: mpm@rohi.engr.sgi.com (Michael Murphy) Message-Id: <200103051745.JAA93353@rohi.engr.sgi.com> To: Peng Zhao Subject: Re: Some error messages and questions Cc: sgi , Jacques-Olivier Haenni Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing From pengzhao@cs.ualberta.ca Sun Mar 4 18:00:02 2001 I am confused by the error messages encountered these days. I use "make build -fMake.cross" and "INSTALL" to generate the sgicc. Should I run sgicc under "nue"? Now that it is a cross compiler, I had thought I can run sgicc directly without nue but run the generated code under NUE. Let's look at the problem in my sgicc under NUE. following is the error message: sgicc test.c sgicc: /usr/ia32-pc-linux/lib/libc.so.6: version `GLIBC_2.1.3' not found (required by sgicc) You can compile apart from NUE, but you can't link apart from NUE (or on a IA64 machine). The above error is a link error, because the wrong library is being used. -- Mike Murphy -- mpm@sgi.com -- quote of the day: -- "Consider the postage stamp. It secures success through its ability -- to stick to one thing till it gets there." (Josh Billings) From owner-pro64-support@oss.sgi.com Mon Mar 5 13:14:32 2001 Received: by oss.sgi.com id ; Mon, 5 Mar 2001 13:14:13 -0800 Received: from sgi.SGI.COM ([192.48.153.1]:20997 "EHLO sgi.com") by oss.sgi.com with ESMTP id ; Mon, 5 Mar 2001 13:14:08 -0800 Received: from harpoon.engr.sgi.com ([130.62.37.31]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id NAA04249 for ; Mon, 5 Mar 2001 13:13:59 -0800 (PST) mail_from (jkingdon@cthulhu.engr.sgi.com) Received: from engr.sgi.com (kingdon.engr.sgi.com [130.62.180.80]) by harpoon.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA93320; Mon, 5 Mar 2001 13:12:36 -0800 (PST) Message-ID: <3AA40123.EFFB3755@engr.sgi.com> Date: Mon, 05 Mar 2001 13:12:03 -0800 From: Jim Kingdon X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: Antony Bowers CC: pro64-support@oss.sgi.com Subject: Re: Build tools? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Antony Bowers wrote: > 2. A few code modifications are needed where the Pro64 C++ source is not > standards conforming (the new gcc is a bit closer to enforcing the > language standard). Well, care to send in some diffs (or at least pointers to what needs to be fixed)? The Party Line (as I understand it) is that Pro64 is written in standard C++ but if the reality doesn't live up to it, well, maybe we can fix that. > I'd prefer to avoid the forked version of gcc 2.96 that comes with Red Hat > 7.0; that compiler is rumoured to be buggy and is surely no less doomed to > obsolescence than 2.95. Shrug. If you apply http://www.redhat.com/support/errata/RHBA-2000-132.html I'm not sure it is as buggy as the rumors say. But I don't think we've been using it for Pro64, so I'm not sure there is any particular reason to go with that version. From owner-pro64-support@oss.sgi.com Tue Mar 6 09:32:56 2001 Received: by oss.sgi.com id ; Tue, 6 Mar 2001 09:32:36 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:26632 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Tue, 6 Mar 2001 09:32:22 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Tue, 6 Mar 2001 10:32:04 -0700 Date: Tue, 6 Mar 2001 10:32:04 -0700 (MST) From: Peng Zhao To: sgi Subject: Some error messages and questions Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi, I built sgicc on ia32 and run it under NUE. I also think it is because of some problem of the version of glibc E.g., I have two workstations: peers and sakwatamau, following is the result of command of "locate libc.so", notice that the error message complains about libc.so.6 +++++++++++++++++++++++++++++++++++++ peers:/usr/brule12/grad/pengzhao/oss/osprey1.0> locate libc.so /usr/lib/libc.so /usr/i486-linux-libc5/lib/libc.so.5.3.12 /usr/i486-linux-libc5/lib/libc.so.5 /lib/libc.so.6 ++++++++++++++++++++++++++++++++++++ sakwatamau:/usr/brule12/grad/pengzhao/oss> locate libc.so /usr/lib/libc.so /usr/i486-linux-libc5/lib/libc.so.5.3.12 /usr/i486-linux-libc5/lib/libc.so.5 /lib/libc.so.6 /nue/usr/ia32-pc-linux/lib/libc.so.6 /nue/usr/ia32-pc-linux/usr/lib/libc.so /nue/usr/lib/libc.so /nue/lib/libc.so.0 +++++++++++++++++++++++++++++++++++++ After building, I copy the sgicc to the direcotory accessable by NUE. the sgicc generated on peers has no problem but sakwatamau one fails Now, the question is: which need the libc.so thing? sgicc or the generation of my "program.c" if it is the process of generating my application that need libc.so.6, what is the sequence that sgicc search this share object? I notice that libc.so.6 doesn't appear the first one of the result of locate on peers. Another question is that why the libc.so.6 on ia32 system is needed by the generation of a ia64 application? Thanks Peng From owner-pro64-support@oss.sgi.com Tue Mar 6 10:46:45 2001 Received: by oss.sgi.com id ; Tue, 6 Mar 2001 10:46:35 -0800 Received: from beta.dmz-eu.st.com ([164.129.1.35]:23755 "HELO beta.dmz-eu.st.com") by oss.sgi.com with SMTP id ; Tue, 6 Mar 2001 10:46:11 -0800 Received: from zeta.dmz-eu.st.com (zeta.dmz-eu.st.com [164.129.230.9]) by beta.dmz-eu.st.com (STMicroelectronics) with SMTP id 091F84961; Tue, 6 Mar 2001 18:46:04 +0000 (GMT) Received: by zeta.dmz-eu.st.com (STMicroelectronics, from userid 0) id A68A9491E; Tue, 6 Mar 2001 18:46:14 +0000 (GMT) Received: from thistle.bri.st.com (localhost [127.0.0.1]) by zeta.dmz-eu.st.com (STMicroelectronics) with ESMTP id 27C701847; Tue, 6 Mar 2001 18:46:14 +0000 (GMT) Received: from [164.129.14.43] (helo=absolut) by thistle.bristol.st.com with esmtp (Exim 3.03 #5) id 14aMTC-0003M1-00; Tue, 06 Mar 2001 18:46:02 +0000 Received: from bowersa (helo=localhost) by absolut with local-esmtp (Exim 3.03 #5) id 14aMTC-0004sv-00; Tue, 06 Mar 2001 18:46:02 +0000 Date: Tue, 6 Mar 2001 18:46:02 +0000 (GMT) From: Antony Bowers X-Sender: bowersa@absolut To: Jim Kingdon Cc: pro64-support@oss.sgi.com Subject: Re: Build tools? In-Reply-To: <3AA40123.EFFB3755@engr.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi Jim, On Mon, 5 Mar 2001, Jim Kingdon wrote: > Well, care to send in some diffs (or at least pointers to what needs to > be fixed)? The Party Line (as I understand it) is that Pro64 is written > in standard C++ but if the reality doesn't live up to it, well, > maybe we can fix that. Sorry, no diffs, but I can point out a couple of things (admittedly not all language problems). I'm using gcc version 2.97 20010212 (experimental) obtained from the gcc-3_0-branch of the CVS repository, on Red Hat 6.2. 1. be/opt/opt_cfg_trans.h line 652 The specialisation of struct less<> needs to be done in the std namespace. 2. common/com/ia64/targ_sim_body.h line 142 The declaration static SIM sim_info[]; is rejected (a defining declaration needs a complete type). This also clashes with static SIM sim_info[] = { /* big initialiser */ }; in common/com/ia64/targ_sim.cxx 3. be/com/emulate.cxx line 69 trunc() is already provided by the libm that comes with glibc2.1 (at least on RH6.2). > > I'd prefer to avoid the forked version of gcc 2.96 that comes with Red Hat > > 7.0; > I don't think we've been using it for Pro64, so I'm not sure there is > any particular reason to go with that version. Great, I was half expecting to hear that everybody is using that one :) Question: after tinkering with these files, I wanted to start the build again from scratch; running 'make clean' didn't seem to have that effect, so I tried 'make clobber' and now the build process seems badly broken. Does the clobber target remove something that make can't rebuild? If I can get going again I will probably find more that needs fixing. Cheers, Tony -- Antony Bowers, STMicroelectronics, Bristol, UK From owner-pro64-support@oss.sgi.com Wed Mar 7 14:40:10 2001 Received: by oss.sgi.com id ; Wed, 7 Mar 2001 14:40:00 -0800 Received: from [38.170.141.29] ([38.170.141.29]:19441 "EHLO heart.hq.tensilica.com") by oss.sgi.com with ESMTP id ; Wed, 7 Mar 2001 14:39:51 -0800 Received: (from goodwin@localhost) by heart.hq.tensilica.com (8.9.3/8.9.3) id OAA20550; Wed, 7 Mar 2001 14:39:50 -0800 X-Authentication-Warning: heart.hq.tensilica.com: goodwin set sender to goodwin@tensilica.com using -f From: David Goodwin MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15014.47286.765847.933570@heart.hq.tensilica.com> Date: Wed, 7 Mar 2001 14:39:50 -0800 (PST) To: pro64-support@oss.sgi.com Cc: goodwin@tensilica.com Subject: c++ fe bugs X-Mailer: VM 6.75 under Emacs 20.5.1 Reply-to: goodwin@tensilica.com Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing I've mostly completed work on adding exception handling to the compiler, but I've run across several c++ fe bugs that are holding me up. Before trying to tackle them myself, I wanted to see if you had already fixed them... The attached code and fe whirl show some of the problems. 1. There are object copies introduced without using the ECLASS copy constructor. Some of these are trivial to optimize: MMLDID 0 <2,3,anon2> T<28,ECLASS,4> MSTID 0 <2,3,anon2> T<28,ECLASS,4> but others are introduced by the MCOMMA expressions: MMLDID 0 <2,2,anon1> T<28,ECLASS,4> MCOMMA MRETURN_VAL 2. The calling standard for passing/returning objects seems incorrect. I'm not good at reading ia64 code, but looking as the assembly file, it seems the ECLASS object is passed and returned by value (i.e. by copying into registers and or memory). g++ seems to pass a class by reference if it has no methods. Mips mongoose seems a little smarter in that it only passes a class by reference if it has a copy constructor. David class ECLASS { private: int _i[10]; public: static int _cnt; ECLASS (); ECLASS (int i); ECLASS (const ECLASS& ec); ~ECLASS(); int val (void) const; }; extern ECLASS foo (ECLASS); ECLASS eh1 (int sel) { return foo(ECLASS(1)); } LOC 0 0 source files: 1 "/usr/local/home/goodwin/simp.ii" LOC 0 0 source files: 2 "/usr/local/home/goodwin/simp.cc" LOC 2 19 eh1 (int sel) LOC 2 20 { FUNC_ENTRY <1,21,eh1__Fi> IDNAME 0 <2,1,sel> BODY BLOCK END_BLOCK BLOCK END_BLOCK BLOCK PRAGMA 0 120 0 (0x0) # PREAMBLE_END LOC 2 21 return foo(ECLASS(1)); BLOCK U8LDA 0 <2,3,anon2> T<33,anon_ptr.,8> U8PARM 2 T<33,anon_ptr.,8> # by_value I4INTCONST 1 (0x1) I4PARM 2 T<4,.predef_I4,4> # by_value VCALL 126 <1,26,__6ECLASSi> # flags 0x7e MMLDID 0 <2,3,anon2> T<28,ECLASS,4> MSTID 0 <2,3,anon2> T<28,ECLASS,4> BLOCK MMLDID 0 <2,3,anon2> T<28,ECLASS,4> MPARM 2 T<28,ECLASS,4> # by_value MCALL 126 <1,25,foo__FG6ECLASS> # flags 0x7e MMLDID -1 <1,19,.preg_return_val> T<28,ECLASS,4> MSTID 0 <2,4,.Mreturn.foo__FG6ECLASS_temp_0> T<28,ECLASS,4> END_BLOCK MMLDID 0 <2,4,.Mreturn.foo__FG6ECLASS_temp_0> T<28,ECLASS,4> MCOMMA MSTID 0 <2,2,anon1> T<28,ECLASS,4> U8LDA 0 <2,3,anon2> T<33,anon_ptr.,8> U8PARM 2 T<33,anon_ptr.,8> # by_value I4INTCONST 2 (0x2) I4PARM 2 T<4,.predef_I4,4> # by_value VCALL 126 <1,27,_._6ECLASS> # flags 0x7e END_BLOCK MMLDID 0 <2,2,anon1> T<28,ECLASS,4> MCOMMA MRETURN_VAL LOC 2 22 } LABEL L1 0 END_BLOCK From owner-pro64-support@oss.sgi.com Thu Mar 8 21:23:24 2001 Received: by oss.sgi.com id ; Thu, 8 Mar 2001 21:23:14 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:25360 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Thu, 8 Mar 2001 21:22:56 -0800 Received: (from localhost user: 'mcnaught' uid#11891 fake: STDIN (mcnaught@tees.cs.ualberta.ca)) by scapa.cs.ualberta.ca id ; Thu, 8 Mar 2001 22:22:39 -0700 Date: Thu, 8 Mar 2001 22:22:39 -0700 (MST) From: Matthew McNaughton To: pro64-support@oss.sgi.com Subject: osprey design web page Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi, I'm doing some research on hyperblock formation strategies, and I noticed in osprey1.0/be/cg/hb_id_candidates.h around line 51, a mention of an "osprey design web page for hyperblocks", but no URL. I also couldn't find any such a thing on the pro64 page, nor with a web search. Is this an internal document to sgi? If so, could someone please sneak me a peek at it? ;) Thanks, -- Matthew McNaughton From owner-pro64-support@oss.sgi.com Thu Mar 8 22:22:34 2001 Received: by oss.sgi.com id ; Thu, 8 Mar 2001 22:22:24 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:15374 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Thu, 8 Mar 2001 22:21:56 -0800 Received: from sgitokyo.nsg.sgi.com (sgitokyo.nsg.sgi.com [134.14.128.2]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id WAA08466 for ; Thu, 8 Mar 2001 22:31:41 -0800 (PST) mail_from (ds-takeda@nsg.sgi.com) Received: from phoenix.nsg.sgi.com (phoenix.nsg.sgi.com [134.14.128.19]) by sgitokyo.nsg.sgi.com (8.8.8/3.6W-99043013) id PAA08001 for ; Fri, 9 Mar 2001 15:20:38 +0900 (JST) Received: from wonder.nsg.sgi.com.nsg.sgi.com ([134.14.130.235]) by phoenix.nsg.sgi.com (Netscape Messaging Server 3.6) with ESMTP id AAA96CC7; Fri, 9 Mar 2001 15:19:44 +0900 Date: Fri, 09 Mar 2001 15:16:53 +0900 Message-ID: <8fly9ufqxey.wl@nsg.sgi.com> From: "Daisuke Takeda" To: pro64-support@oss.sgi.com Subject: OpenMP and libmp User-Agent: Wanderlust/2.2.15 (More Than Words) WEMI/1.13.7 (Shimada) FLIM/1.13.2 (Kasanui) MULE XEmacs/21.1 (patch 10) (Capitol Reef) (i386-debian-linux) Organization: SGI Japan, Ltd. MIME-Version: 1.0 (generated by WEMI 1.13.7 - "Shimada") Content-Type: text/plain; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hello, We tried OpenMP feature of SGI Pro64 compiler (0.01.0-12). However, we could not succeed parallelize with it because of the lack of libmp, as shown below: [ds-takeda@troon02 ~]# sgicc -O3 -mp mxm.c /usr/bin/ld: cannot find -lmp collect2: ld returned 1 exit status [ds-takeda@troon02 ~]# sgicc -O3 -apo mxm.c /usr/bin/ld: cannot find -lmp collect2: ld returned 1 exit status [ds-takeda@troon02 ~]# sgicc --version sgicc ERROR: -- not allowed in non XPG4 environment sgicc ERROR parsing --version: unknown flag SGIcc Compilers: Version 0.01.0-12 We looked for libmp but could not find it. Is it available now? If it is, would you please show me where to get it? If not yet, I hope to know the schedule to release it. Thanks in advance, -- // Daisuke TAKEDA Scalable Systems Technology Center // email: ds-takeda@sgi.co.jp Products and Technology // phone: 81-3-5488-1838 SGI Japan, Ltd. From owner-pro64-support@oss.sgi.com Fri Mar 9 01:55:57 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 01:55:47 -0800 Received: from [159.226.39.4] ([159.226.39.4]:3335 "HELO mail.ict.ac.cn") by oss.sgi.com with SMTP id ; Fri, 9 Mar 2001 01:55:23 -0800 Received: (qmail 1371 invoked from network); 9 Mar 2001 09:50:19 -0000 Received: from unknown (HELO ACT116) (159.226.40.174) by 159.226.39.4 with SMTP; 9 Mar 2001 09:50:19 -0000 Message-ID: <000701c0a87f$56301ac0$ae28e29f@ACT116> From: "zhang kexin" To: Subject: Date: Fri, 9 Mar 2001 17:57:21 +0800 MIME-Version: 1.0 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: base64 X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing c3Vic2NyaWJlIG0tbGlzdCBrZXhpbnpoQGljdC5hYy5jbg0K From owner-pro64-support@oss.sgi.com Fri Mar 9 01:57:17 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 01:57:07 -0800 Received: from [159.226.39.4] ([159.226.39.4]:5639 "HELO mail.ict.ac.cn") by oss.sgi.com with SMTP id ; Fri, 9 Mar 2001 01:57:05 -0800 Received: (qmail 1506 invoked from network); 9 Mar 2001 09:52:06 -0000 Received: from unknown (HELO ACT116) (159.226.40.174) by 159.226.39.4 with SMTP; 9 Mar 2001 09:52:06 -0000 Message-ID: <000d01c0a87f$95a9a590$ae28e29f@ACT116> From: "zhang kexin" To: Subject: Date: Fri, 9 Mar 2001 17:59:08 +0800 MIME-Version: 1.0 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: base64 X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing DQogc3Vic2NyaWJlIHBybzY0LXN1cHBvcnQga2V4aW56aEBpY3QuYWMuY24NCiANCg== From owner-pro64-support@oss.sgi.com Fri Mar 9 06:25:41 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 06:25:31 -0800 Received: from [38.138.181.132] ([38.138.181.132]:52752 "EHLO gate.init.com") by oss.sgi.com with ESMTP id ; Fri, 9 Mar 2001 06:25:19 -0800 Received: from denali.abinitio.com (localmail.init.com [10.50.30.91]) by gate.init.com (8.9.3/8.9.3/ab-gate-1.1) with ESMTP id KAA11689; Fri, 9 Mar 2001 10:20:40 -0500 Received: from localhost.localdomain (neo.init.com [10.50.20.32]) by denali.abinitio.com (8.9.3/8.9.3/ab-hub-1.2) with ESMTP id JAA14255; Fri, 9 Mar 2001 09:20:14 -0500 (EST) Received: (from rshapiro@localhost) by localhost.localdomain (8.9.3/8.8.8/ab-cli-1.1) id JAA03954; Fri, 9 Mar 2001 09:20:14 -0500 Date: Fri, 9 Mar 2001 09:20:14 -0500 Message-Id: <200103091420.JAA03954@localhost.localdomain> X-Authentication-Warning: localhost.localdomain: rshapiro set sender to rshapiro@localhost.localdomain using -f From: Richard Shapiro To: mcnaught@cs.ualberta.ca CC: pro64-support@oss.sgi.com In-reply-to: (message from Matthew McNaughton on Thu, 8 Mar 2001 22:22:39 -0700 (MST)) Subject: Re: osprey design web page References: Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Date: Thu, 8 Mar 2001 22:22:39 -0700 (MST) From: Matthew McNaughton X-Orcpt: rfc822;pro64-support@oss.sgi.com Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Length: 433 Hi, I'm doing some research on hyperblock formation strategies, and I noticed in osprey1.0/be/cg/hb_id_candidates.h around line 51, a mention of an "osprey design web page for hyperblocks", but no URL. I also couldn't find any such a thing on the pro64 page, nor with a web search. Is this an internal document to sgi? If so, could someone please sneak me a peek at it? ;) I don;t think you would actually find that page all that helpful. When I did a re-write of the hyperblock code, a lot of the original ideas in the referenced document turned out to be bad ideas, but we never (as of a year ago) updated the document. Rich -- Richard Shapiro Ab Initio Software Corporation rshapiro@abinitio.com (781) 301-2311 From owner-pro64-support@oss.sgi.com Fri Mar 9 07:25:31 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 07:25:21 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:40747 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Fri, 9 Mar 2001 07:25:05 -0800 Received: from gaea.engr.sgi.com (gaea.engr.sgi.com [130.62.180.97]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id HAA02851 for ; Fri, 9 Mar 2001 07:34:52 -0800 (PST) mail_from (murthy@gaea.engr.sgi.com) Received: (from murthy@localhost) by gaea.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id HAA11272; Fri, 9 Mar 2001 07:23:19 -0800 (PST) Date: Fri, 9 Mar 2001 07:23:19 -0800 (PST) From: murthy@gaea.engr.sgi.com (Chandrasekhar Murthy) Message-Id: <200103091523.HAA11272@gaea.engr.sgi.com> To: pro64-support@oss.sgi.com, "Daisuke Takeda" Subject: Re: OpenMP and libmp Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Currently there is no support for Open MP in sgicc or sgiCC. Although sgif90 does support Open MP, libmp is not currently released. Murthy From owner-pro64-support@oss.sgi.com Fri Mar 9 09:41:32 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 09:41:22 -0800 Received: from sgi.SGI.COM ([192.48.153.1]:50964 "EHLO sgi.com") by oss.sgi.com with ESMTP id ; Fri, 9 Mar 2001 09:41:12 -0800 Received: from cchkms.engr.sgi.com ([130.62.180.48]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id JAA00655 for ; Fri, 9 Mar 2001 09:41:11 -0800 (PST) mail_from (rat@cchkms.engr.sgi.com) Received: (from rat@localhost) by cchkms.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id JAA84239 for pro64-support@oss.sgi.com; Fri, 9 Mar 2001 09:39:50 -0800 (PST) From: "Ross A. Towle" Message-Id: <10103090939.ZM84525@cchkms.engr.sgi.com> Date: Fri, 9 Mar 2001 09:39:50 -0800 In-Reply-To: "Daisuke Takeda" "OpenMP and libmp" (Mar 9, 3:16pm) References: <8fly9ufqxey.wl@nsg.sgi.com> X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: pro64-support@oss.sgi.com Subject: Re: OpenMP and libmp Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing To fully support OpenMP we need to implement THREADPRIVATE which requires some linker work. This will happen when Cygnus/RedHat provides an ELF based linker (instead of bfd based). -Ross From owner-pro64-support@oss.sgi.com Fri Mar 9 14:05:52 2001 Received: by oss.sgi.com id ; Fri, 9 Mar 2001 14:05:32 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:58695 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Fri, 9 Mar 2001 14:05:06 -0800 Received: from harpoon.engr.sgi.com (harpoon.engr.sgi.com [130.62.41.49]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id OAA24328 for ; Fri, 9 Mar 2001 14:03:58 -0800 (PST) mail_from (jkingdon@cthulhu.engr.sgi.com) Received: from engr.sgi.com (kingdon.engr.sgi.com [130.62.180.80]) by harpoon.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id OAA22082 for ; Fri, 9 Mar 2001 14:03:47 -0800 (PST) Message-ID: <3AA95321.868045C4@engr.sgi.com> Date: Fri, 09 Mar 2001 14:03:13 -0800 From: Jim Kingdon X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: pro64-support@oss.sgi.com Subject: Re: OpenMP and libmp References: <8fly9ufqxey.wl@nsg.sgi.com> <10103090939.ZM84525@cchkms.engr.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing "Ross A. Towle" wrote: > To fully support OpenMP we need to implement THREADPRIVATE which requires > some linker work. This will happen when Cygnus/RedHat provides an ELF based > linker (instead of bfd based). :-). Don't know anything about THREADPRIVATE, but Jack Carter and I have been making good progress on getting the IPA stuff merged in to GNU ld. See http://sources.redhat.com/ml/binutils/2001-03/msg00072.html http://sources.redhat.com/ml/binutils/2001-02/msg00566.html Jack is going to get some kind of medal once we get through this one.... From owner-pro64-support@oss.sgi.com Sun Mar 11 13:47:38 2001 Received: by oss.sgi.com id ; Sun, 11 Mar 2001 13:47:18 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:3335 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Sun, 11 Mar 2001 13:47:09 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Sat, 10 Mar 2001 22:21:45 -0700 Date: Sat, 10 Mar 2001 22:21:44 -0700 (MST) From: Peng Zhao To: Jacques-Olivier Haenni cc: sgi Subject: Re: Some error messages and questions In-Reply-To: <3AA0FADC.404A8127@epfl.ch> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi, > > : -Ofast is not supported yet. It uses ipa, which has not been released yet > : (but will be soon). Use -O3 for now. In the source code of Pro64, there is a directory named ipa. Isn't it the ipa you have meant? What is the difference of the light-weight inliner, standalone inliner and IPA? Although I have read several times of the mannual of the IPA(5), I am still confused on this. When I use "-INLINE:dfe=ON", it seems that it doesn't work: the dead function is not eliminated while it would never be called after inlining.. I have a very simple program in which the main call the only function. The function is inlined, but the function is still there in the assembly program (.s). Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Sun Mar 11 13:55:08 2001 Received: by oss.sgi.com id ; Sun, 11 Mar 2001 13:54:49 -0800 Received: from dynamic-addr-83-177.resnet.rochester.edu ([128.151.83.177]:32640 "EHLO dynamic-addr-83-177.resnet.rochester.edu") by oss.sgi.com with ESMTP id ; Sun, 11 Mar 2001 13:54:46 -0800 Received: (from dberlin@localhost) by dynamic-addr-83-177.resnet.rochester.edu (8.11.2/8.11.0/SuSE Linux 8.11.0-0.4) id f2BLsTf17957; Sun, 11 Mar 2001 16:54:29 -0500 X-Authentication-Warning: dynamic-addr-83-177.resnet.rochester.edu: dberlin set sender to dberlin@redhat.com using -f To: Peng Zhao Cc: Jacques-Olivier Haenni , sgi Subject: Re: Some error messages and questions References: From: Daniel Berlin Date: 11 Mar 2001 16:54:28 -0500 In-Reply-To: (Peng Zhao's message of "Sat, 10 Mar 2001 22:21:44 -0700 (MST)") Message-ID: Lines: 24 User-Agent: Gnus/5.090001 (Oort Gnus v0.01) XEmacs/21.2 (Thalia) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Peng Zhao writes: > Hi, > > > > : -Ofast is not supported yet. It uses ipa, which has not been released yet > > : (but will be soon). Use -O3 for now. > > In the source code of Pro64, there is a directory named ipa. Isn't it the > ipa you have meant? > > What is the difference of the light-weight inliner, standalone inliner and > IPA? Although I have read several times of the mannual of the IPA(5), I > am still confused on this. > > When I use "-INLINE:dfe=ON", it seems that it doesn't work: the dead > function is not eliminated while it would never be called after inlining.. > I have a very > simple program in which the main call the only function. The function is > inlined, but the function is still there in the assembly program (.s). > Is the function static? If not, it can't eliminate it. --Dan From owner-pro64-support@oss.sgi.com Mon Mar 12 07:11:30 2001 Received: by oss.sgi.com id ; Mon, 12 Mar 2001 07:11:10 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:9576 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Mon, 12 Mar 2001 07:11:07 -0800 Received: from gaea.engr.sgi.com (gaea.engr.sgi.com [130.62.180.97]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id HAA11183 for ; Mon, 12 Mar 2001 07:09:59 -0800 (PST) mail_from (murthy@gaea.engr.sgi.com) Received: (from murthy@localhost) by gaea.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id HAA35965; Mon, 12 Mar 2001 07:09:19 -0800 (PST) Date: Mon, 12 Mar 2001 07:09:19 -0800 (PST) From: murthy@gaea.engr.sgi.com (Chandrasekhar Murthy) Message-Id: <200103121509.HAA35965@gaea.engr.sgi.com> To: Jacques-Olivier Haenni , Peng Zhao Subject: Re: Some error messages and questions Cc: sgi Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Although the ipa directory is there, some of the changes in ld which are needed by ipa were not yet released. The light-weight inliner and the standalone inliner both operate only on the current file. IPA takes a more global approach and does cross file inlining. The following gives you a rough idea of the icreasing amount of work done 1. light-weight inliner 2. standalone inliner (more than light-weight) 3. IPA (more than standalone) Could you mail us your sample program. Thanks, Murthy From owner-pro64-support@oss.sgi.com Mon Mar 12 10:20:20 2001 Received: by oss.sgi.com id ; Mon, 12 Mar 2001 10:20:10 -0800 Received: from scapa.cs.ualberta.ca ([129.128.4.44]:9994 "EHLO scapa.cs.ualberta.ca") by oss.sgi.com with ESMTP id ; Mon, 12 Mar 2001 10:19:50 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Mon, 12 Mar 2001 11:19:30 -0700 Date: Mon, 12 Mar 2001 11:19:30 -0700 (MST) From: Peng Zhao To: Shin-Ming Liu cc: sgi Subject: Re: Some error messages and questions In-Reply-To: <20010312073400.12154.qmail@web3504.mail.yahoo.com> Message-ID: MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="-2138965737-1674250855-984421170=:5082" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. Send mail to mime@docserver.cac.washington.edu for more info. ---2138965737-1674250855-984421170=:5082 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, Following is my command to compile my program and eliminate the dead code. ../bin/sgicc -Yf,../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0 -Yb,../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0 -YL,../lib -YL,../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0 -Yi,../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0 -INLINE:must=callit:list:dfe=ON -S test.c And the attatched file is the simple program. Thanks Peng On Sun, 11 Mar 2001, Shin-Ming Liu wrote: > At which optimization level did you try the flag > INLINE:dfe=ON? > > > --- Peng Zhao wrote: > > Hi, > > > > > > : -Ofast is not supported yet. It uses ipa, which > > has not been released yet > > > : (but will be soon). Use -O3 for now. > > > > In the source code of Pro64, there is a directory > > named ipa. Isn't it the > > ipa you have meant? > > > > What is the difference of the light-weight inliner, > > standalone inliner and > > IPA? Although I have read several times of the > > mannual of the IPA(5), I > > am still confused on this. > > > > When I use "-INLINE:dfe=ON", it seems that it > > doesn't work: the dead > > function is not eliminated while it would never be > > called after inlining.. > > I have a very > > simple program in which the main call the only > > function. The function is > > inlined, but the function is still there in the > > assembly program (.s). > > > > > > Peng > > -- > > Peng Zhao pengzhao@cs.ualberta.ca > > http://www.cs.ualberta.ca/~pengzhao > > TEL (Lab): (780)492-3725 Lab: > > CSC251 > > > > > > > __________________________________________________ > Do You Yahoo!? > Yahoo! Auctions - Buy the things you want at great prices. > http://auctions.yahoo.com/ > Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 ---2138965737-1674250855-984421170=:5082 Content-Type: TEXT/PLAIN; charset=US-ASCII; name="test.c" Content-Transfer-Encoding: BASE64 Content-ID: Content-Description: Content-Disposition: attachment; filename="test.c" I2luY2x1ZGUgPHN0ZGlvLmg+DQoNCnZvaWQgY2FsbGl0KGludCk7DQoNCm1h aW4oKQ0Kew0KCWludCBpLGosazsNCg0KCWkgPSAxMDsNCgljYWxsaXQoaSk7 DQoJaj1pOw0KfQ0KDQp2b2lkDQpjYWxsaXQoaW50IGopDQp7DQoJaW50IGk7 DQoJaSA9IGo7DQoJaSArKzsNCglpICsrOw0KCWkgKys7DQoJaSArKzsNCglp ICsrOw0KCXByaW50ZigiJWRcbiIsIGkpOw0KDQp9DQo= ---2138965737-1674250855-984421170=:5082-- From owner-pro64-support@oss.sgi.com Mon Mar 12 13:51:51 2001 Received: by oss.sgi.com id ; Mon, 12 Mar 2001 13:51:32 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:53523 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Mon, 12 Mar 2001 13:51:15 -0800 Received: from harpoon.engr.sgi.com (harpoon.engr.sgi.com [130.62.41.49]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id OAA06405 for ; Mon, 12 Mar 2001 14:01:04 -0800 (PST) mail_from (jkingdon@cthulhu.engr.sgi.com) Received: from engr.sgi.com (kingdon.engr.sgi.com [130.62.180.80]) by harpoon.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA36428; Mon, 12 Mar 2001 13:49:45 -0800 (PST) Message-ID: <3AAD4456.5E517194@engr.sgi.com> Date: Mon, 12 Mar 2001 13:49:10 -0800 From: Jim Kingdon X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 CC: Jacques-Olivier Haenni , Peng Zhao , sgi Subject: Re: Some error messages and questions References: <200103121509.HAA35965@gaea.engr.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: unlisted-recipients:; (no To-header on input) Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Chandrasekhar Murthy wrote: > Although the ipa directory is there, some of the changes in ld which are > needed by ipa were not yet released. Well, if you really want to work on it, the changes in ld are at http://sources.redhat.com/ml/binutils/2001-02/msg00553.html The disclaimer is that this is more of a work in progress than something which is necessarily ready to go as-is. A few of us at SGI have been working on the issues discussed in http://sources.redhat.com/ml/binutils/2001-02/msg00566.html but if anyone makes any progress, please do report it to this list and/or the binutils list. We're not above getting help ;-). From owner-pro64-support@oss.sgi.com Mon Mar 12 17:43:01 2001 Received: by oss.sgi.com id ; Mon, 12 Mar 2001 17:42:41 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:8752 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Mon, 12 Mar 2001 17:42:13 -0800 Received: from sgitokyo.nsg.sgi.com (sgitokyo.nsg.sgi.com [134.14.128.2]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id RAA04204 for ; Mon, 12 Mar 2001 17:52:02 -0800 (PST) mail_from (ds-takeda@nsg.sgi.com) Received: from phoenix.nsg.sgi.com (phoenix.nsg.sgi.com [134.14.128.19]) by sgitokyo.nsg.sgi.com (8.8.8/3.6W-99043013) id KAA21491 for ; Tue, 13 Mar 2001 10:40:55 +0900 (JST) Received: from wonder.nsg.sgi.com.nsg.sgi.com ([134.14.130.235]) by phoenix.nsg.sgi.com (Netscape Messaging Server 3.6) with ESMTP id AAAA3CFB for ; Tue, 13 Mar 2001 10:39:56 +0900 Date: Tue, 13 Mar 2001 10:36:59 +0900 Message-ID: <8flpufm5u10.wl@nsg.sgi.com> From: "Daisuke Takeda" To: pro64-support@oss.sgi.com Subject: Re: OpenMP and libmp In-Reply-To: In your message of "Fri, 09 Mar 2001 14:03:13 -0800" <3AA95321.868045C4@engr.sgi.com> References: <8fly9ufqxey.wl@nsg.sgi.com> <10103090939.ZM84525@cchkms.engr.sgi.com> <3AA95321.868045C4@engr.sgi.com> User-Agent: Wanderlust/2.2.15 (More Than Words) WEMI/1.13.7 (Shimada) FLIM/1.13.2 (Kasanui) MULE XEmacs/21.1 (patch 10) (Capitol Reef) (i386-debian-linux) Organization: SGI Japan, Ltd. MIME-Version: 1.0 (generated by WEMI 1.13.7 - "Shimada") Content-Type: text/plain; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hello, Thank you for information. We hope OpenMP will get available as soon as possible. Thanks again. -- // Daisuke TAKEDA Scalable Systems Technology Center // email: ds-takeda@sgi.co.jp Products and Technology // phone: 81-3-5488-1838 SGI Japan, Ltd. From owner-pro64-support@oss.sgi.com Tue Mar 13 13:09:01 2001 Received: by oss.sgi.com id ; Tue, 13 Mar 2001 13:08:42 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:5443 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Tue, 13 Mar 2001 13:08:30 -0800 Received: from cchkms.engr.sgi.com (cchkms.engr.sgi.com [130.62.180.48]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id NAA12618 for ; Tue, 13 Mar 2001 13:07:19 -0800 (PST) mail_from (rat@cchkms.engr.sgi.com) Received: (from rat@localhost) by cchkms.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id NAA91148 for pro64-support@oss.sgi.com; Tue, 13 Mar 2001 13:06:58 -0800 (PST) From: "Ross A. Towle" Message-Id: <10103131306.ZM91503@cchkms.engr.sgi.com> Date: Tue, 13 Mar 2001 13:06:57 -0800 X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: pro64-support@oss.sgi.com Subject: Pro64 release 0.13 is available Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Release 0.13 of Pro64 is available from http://oss.sgi.com/projects/Pro64. Additional features include: - interprocedural optimization is enabled. The needed ipa-link is now included. - flush-to-zero is available via command line argument - numerous bug fixes and performance enhancements - the runtime libraries work with glibc version 2.2 - the website now has an archive of the pro64-support email - the compiler emits additional assembler directives Caveats: Because the runtime depends/requires glibc version 2.2 and the assembly code contains additional assembler directives, this release will not work with HP's NUE until they upgrade the assembler, glibc, and include files. We are working on native version of the compilers. As soon as we can achieve the same level of robutness then we will release those compilers as release 0.14. -Ross A. Towle Director of Engineering High Performance Programming Environment From owner-pro64-support@oss.sgi.com Tue Mar 13 14:17:01 2001 Received: by oss.sgi.com id ; Tue, 13 Mar 2001 14:16:42 -0800 Received: from sgi.SGI.COM ([192.48.153.1]:48767 "EHLO sgi.com") by oss.sgi.com with ESMTP id ; Tue, 13 Mar 2001 14:16:12 -0800 Received: from relay1.corp.sgi.com (corp.sgi.com [198.29.75.13]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id OAA00283 for ; Tue, 13 Mar 2001 14:16:10 -0800 (PST) mail_from (beverlyb@engr.sgi.com) Received: from engr.sgi.com (sshgate.corp.sgi.com [169.238.216.146]) by relay1.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id OAA25098; Tue, 13 Mar 2001 14:15:38 -0800 (PST) Message-ID: <3AAE9BEA.26E22F9@engr.sgi.com> Date: Tue, 13 Mar 2001 14:15:06 -0800 From: Beverly Bernard Organization: SGI X-Mailer: Mozilla 4.73 [en]C-CCK-MCD SGI (WinNT; I) X-Accept-Language: en MIME-Version: 1.0 To: "Ross A. Towle" CC: pro64-support@oss.sgi.com Subject: Re: Pro64 release 0.13 is available References: <10103131306.ZM91503@cchkms.engr.sgi.com> Content-Type: multipart/mixed; boundary="------------89844ECE96B5225C70781262" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing This is a multi-part message in MIME format. --------------89844ECE96B5225C70781262 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Great going, guys! --Beverly "Ross A. Towle" wrote: > > Release 0.13 of Pro64 is available from > http://oss.sgi.com/projects/Pro64. Additional > features include: > - interprocedural optimization is enabled. > The needed ipa-link is now included. > - flush-to-zero is available via command line argument > - numerous bug fixes and performance enhancements > - the runtime libraries work with glibc version 2.2 > - the website now has an archive of the pro64-support > email > - the compiler emits additional assembler directives > > Caveats: Because the runtime depends/requires glibc > version 2.2 and the assembly code contains additional > assembler directives, this release will not work with > HP's NUE until they upgrade the assembler, glibc, and > include files. > > We are working on native version of the compilers. As > soon as we can achieve the same level of robutness then > we will release those compilers as release 0.14. > > -Ross A. Towle > Director of Engineering > High Performance Programming Environment -- Beverly Bernard Product Marketing Manager - Development Tools Vnet 933-4239 Home Office: 775-852-6305 Cell: 775-843-5941 --------------89844ECE96B5225C70781262 Content-Type: text/x-vcard; charset=us-ascii; name="beverlyb.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Beverly Bernard Content-Disposition: attachment; filename="beverlyb.vcf" begin:vcard n:Bernard;Beverly tel;cell:775-843-5941 tel;home:775-852-6305 tel;work:VNet 933-4239 x-mozilla-html:FALSE adr:;;;;;; version:2.1 email;internet:beverlyb@engr.sgi.com fn:Beverly Bernard end:vcard --------------89844ECE96B5225C70781262-- From owner-pro64-support@oss.sgi.com Tue Mar 13 23:59:45 2001 Received: by oss.sgi.com id ; Tue, 13 Mar 2001 23:59:25 -0800 Received: from elin.scali.no ([195.139.250.10]:10256 "EHLO elin.scali.no") by oss.sgi.com with ESMTP id ; Tue, 13 Mar 2001 23:58:59 -0800 Received: from scali.no (pc81-218.aasen.avidi.online.no [148.122.218.81]) by elin.scali.no (8.9.3/8.8.7) with ESMTP id HAA22119 for ; Wed, 14 Mar 2001 07:58:59 +0100 Message-ID: <3AAF24B5.225034B3@scali.no> Date: Wed, 14 Mar 2001 08:58:45 +0100 From: Ole W Saastad Organization: Scali A/S X-Mailer: Mozilla 4.06 [en] (X11; U; Linux 2.0.39 i586) MIME-Version: 1.0 To: pro64-support@oss.sgi.com Subject: IA 64 fortran compiler and getargs Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Dear SGI open source team, I am trying to get hold of the arguments on the command line in a fortran program. The usualy way to do this is by calling getarg. I did not find it in the FAq and nor did I find manuals in the download area, which forces me to bother you. How do I get hold of the command line arguments ? ./Linux2.ia64.gnu/producer_f.o:/home/swtest/work/ScaMPItst/src/apps/producer_f/producer_f.f:124: undefined reference to `iargc_' ./Linux2.ia64.gnu/producer_f.o: In function `MAIN__': ./Linux2.ia64.gnu/producer_f.o(.text+0x1e2): undefined reference to `getarg_' ./Linux2.ia64.gnu/producer_f.o(.text+0x372): undefined reference to `getarg_' ./Linux2.ia64.gnu/producer_f.o(.text+0x3f2): undefined reference to `getarg_' collect2: ld returned 1 exit status Best regards Ole W Saastad Scali AS. http://www.scali.no From owner-pro64-support@oss.sgi.com Wed Mar 14 09:45:38 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 09:45:28 -0800 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:46094 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 09:45:24 -0800 Received: from cchkms.engr.sgi.com (cchkms.engr.sgi.com [130.62.180.48]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id JAA04038 for ; Wed, 14 Mar 2001 09:55:15 -0800 (PST) mail_from (rat@cchkms.engr.sgi.com) Received: (from rat@localhost) by cchkms.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id JAA49418 for pro64-support@oss.sgi.com; Wed, 14 Mar 2001 09:44:03 -0800 (PST) From: "Ross A. Towle" Message-Id: <10103140944.ZM90327@cchkms.engr.sgi.com> Date: Wed, 14 Mar 2001 09:44:02 -0800 In-Reply-To: Ole W Saastad "IA 64 fortran compiler and getargs" (Mar 14, 8:58am) References: <3AAF24B5.225034B3@scali.no> X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: pro64-support@oss.sgi.com Subject: Re: IA 64 fortran compiler and getargs Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing getarg and iargc are part of the Pro64 release 0.13. -Ross A snippet from the man page for them (man pages are being written) subroutine getarg(i, c) character*N c integer*4 i integer*4 function iargc() integer j call getarg(i, c) j = iargc() DESCRIPTION getarg returns the i-th command-line argument of the current process. iargc returns the index of the last argument. foo arg1 arg2 arg3 getarg(2, c) would return the string ``arg2'' in the character variable c. iargc would return 3 as the value of the function call. From owner-pro64-support@oss.sgi.com Wed Mar 14 09:58:08 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 09:57:58 -0800 Received: from probity.mcc.ac.uk ([130.88.200.94]:43528 "EHLO probity.mcc.ac.uk") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 09:57:49 -0800 Received: from wallace.mvc.mcc.ac.uk ([130.88.1.130] helo=man.ac.uk) by probity.mcc.ac.uk with esmtp (Exim 2.05 #4) id 14dFWt-0009JO-00; Wed, 14 Mar 2001 17:57:47 +0000 Message-ID: <3AAFB146.62C8AD21@man.ac.uk> Date: Wed, 14 Mar 2001 17:58:30 +0000 From: Dan Kidger Organization: Manchester Computing X-Mailer: Mozilla 4.76 [en] (Windows NT 5.0; U) X-Accept-Language: en MIME-Version: 1.0 To: "Ross A. Towle" CC: pro-64 Subject: Re: IA 64 fortran compiler and getargs References: <3AAF24B5.225034B3@scali.no> <10103140944.ZM90327@cchkms.engr.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Dear Ross, "Ross A. Towle" wrote: > > getarg and iargc are part of the Pro64 release 0.13. > > -Ross > > A snippet from the man page for them (man pages are being written) The functions getarg and iargc in pro-64 appear to be the same as on most Unix platforms. The question is which library are they in: For example in the following code, f90 under Irix doesn't need a library to resolve getarg and iargc, but I have never found how to lik to them on our Troon boxes.... program test_command_line_args integer :: i,nargs, iargc character :: argument*30 nargs=iargc() print*,' The ',narg,' arguments you gave were...' do i=1,nargs call getarg(i,argument) print*,i,argument enddo end Yours, Daniel ----------------------------------------------------------------------- Dr. Daniel Kidger | E: d.kidger@man.ac.uk High Performance Computing Group | W: www.csar.cfs.ac.uk Manchester Computing, University of Manchester, | T: +44 161 275 7038 Oxford Road, Manchester, M13 9PL, UK | F: +44 161 275 6800 --------------------Q: what's up ? A: X cross Z --------------------- From owner-pro64-support@oss.sgi.com Wed Mar 14 10:06:48 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 10:06:38 -0800 Received: from naghost.nag.co.uk ([62.232.54.245]:18306 "EHLO egham.nag.co.uk") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 10:06:33 -0800 Received: (from arnaud@localhost) by egham.nag.co.uk (8.8.7/8.8.7) id SAA15829; Wed, 14 Mar 2001 18:10:24 GMT From: Arnaud Desitter Message-Id: <200103141810.SAA15829@egham.nag.co.uk> Subject: Re: IA 64 fortran compiler and getargs To: rat@cchkms.engr.sgi.com (Ross A. Towle) Date: Wed, 14 Mar 2001 18:10:24 +0000 () Cc: pro64-support@oss.sgi.com In-Reply-To: <10103140944.ZM90327@cchkms.engr.sgi.com> from "Ross A. Towle" at Mar 14, 2001 09:44:02 AM Reply-To: Arnaud Desitter Organization: Nag Ltd. X-Mailer: ELM [version 2.5 PL2] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Hi, I would be nice to have Pro64 support in the F2k proposal implemented in: http://www.winteracter.com/f2kcli/ Cheers, Arnaud > > getarg and iargc are part of the Pro64 release 0.13. > > -Ross > From owner-pro64-support@oss.sgi.com Wed Mar 14 10:11:38 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 10:11:28 -0800 Received: from sgi.SGI.COM ([192.48.153.1]:24366 "EHLO sgi.com") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 10:11:24 -0800 Received: from cchkms.engr.sgi.com ([130.62.180.48]) by sgi.com (980327.SGI.8.8.8-aspam/980304.SGI-aspam: SGI does not authorize the use of its proprietary systems or networks for unsolicited or bulk email from the Internet.) via ESMTP id KAA09571 for ; Wed, 14 Mar 2001 10:11:22 -0800 (PST) mail_from (rat@cchkms.engr.sgi.com) Received: (from rat@localhost) by cchkms.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id KAA69612 for pro64-support@oss.sgi.com; Wed, 14 Mar 2001 10:10:02 -0800 (PST) From: "Ross A. Towle" Message-Id: <10103141010.ZM93216@cchkms.engr.sgi.com> Date: Wed, 14 Mar 2001 10:10:01 -0800 In-Reply-To: Dan Kidger "Re: IA 64 fortran compiler and getargs" (Mar 14, 5:58pm) References: <3AAF24B5.225034B3@scali.no> <10103140944.ZM90327@cchkms.engr.sgi.com> <3AAFB146.62C8AD21@man.ac.uk> X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: pro-64 Subject: Re: IA 64 fortran compiler and getargs Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Dan, getarg and iargc for Fortran90 appear in libfortran. The actual entry points are named "getarg_" and "iargc_". These are wrappers around the calls to the c routines. These are part of 0.13 Pro64 release. They were not part of earlier releases which is why you would have had problems prior to yesterday's release. This wrapper technique is how lots of other common "c language" functions are implemented. And sorry to burst your bubble about IRIX and getarg. It works EXACTLY the same as in Pro64. There is a wrapper. The wrapper is in a library. And the library is automatically linked in by the f90 command. -Ross From owner-pro64-support@oss.sgi.com Wed Mar 14 10:19:28 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 10:19:18 -0800 Received: from probity.mcc.ac.uk ([130.88.200.94]:12811 "EHLO probity.mcc.ac.uk") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 10:19:12 -0800 Received: from wallace.mvc.mcc.ac.uk ([130.88.1.130] helo=man.ac.uk) by probity.mcc.ac.uk with esmtp (Exim 2.05 #4) id 14dFra-0009cg-00; Wed, 14 Mar 2001 18:19:10 +0000 Message-ID: <3AAFB64A.EBA1989A@man.ac.uk> Date: Wed, 14 Mar 2001 18:19:54 +0000 From: Dan Kidger Organization: Manchester Computing X-Mailer: Mozilla 4.76 [en] (Windows NT 5.0; U) X-Accept-Language: en MIME-Version: 1.0 To: "Ross A. Towle" CC: pro-64 Subject: Re: IA 64 fortran compiler and getargs References: <3AAF24B5.225034B3@scali.no> <10103140944.ZM90327@cchkms.engr.sgi.com> <3AAFB146.62C8AD21@man.ac.uk> <10103141010.ZM93216@cchkms.engr.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing "Ross A. Towle" wrote: > > Dan, > > getarg and iargc for Fortran90 appear in libfortran. The actual entry > points are named "getarg_" and "iargc_". These are wrappers around the > calls to the c routines. These are part of 0.13 Pro64 release. They were > not part of earlier releases which is why you would have had problems > prior to yesterday's release. done, In the last 5 mins, I have downloaded and installed 0.13, and it all now works fine... [zzcgudk@fourier4 f90]$ ./a.out Thanks Ross The 2 arguments you gave were... 1 Thanks 2 Ross By the way 0.13 is not pointed to from http://oss.sgi.com/projects/Pro64/download.html - I had to guess the URL. :-) Thanks, Daniel ----------------------------------------------------------------------- Dr. Daniel Kidger | E: d.kidger@man.ac.uk High Performance Computing Group | W: www.csar.cfs.ac.uk Manchester Computing, University of Manchester, | T: +44 161 275 7038 Oxford Road, Manchester, M13 9PL, UK | F: +44 161 275 6800 --------------------Q: what's up ? A: X cross Z --------------------- From owner-pro64-support@oss.sgi.com Wed Mar 14 10:55:38 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 10:55:28 -0800 Received: from deliverator.sgi.com ([204.94.214.10]:31551 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 10:55:20 -0800 Received: from cthulhu.engr.sgi.com (gate3-relay.engr.sgi.com [130.62.1.234]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id KAA20223 for ; Wed, 14 Mar 2001 10:54:09 -0800 (PST) mail_from (jbaron@sgi.com) Received: from coolguy.engr.sgi.com (coolguy.engr.sgi.com [130.62.176.135]) by cthulhu.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id KAA94943; Wed, 14 Mar 2001 10:54:02 -0800 (PST) Received: from sgi.com (coolguy.engr.sgi.com [130.62.176.135]) by coolguy.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id KAA09292; Wed, 14 Mar 2001 10:54:01 -0800 (PST) Message-ID: <3AAFBE49.94E1BAD3@sgi.com> Date: Wed, 14 Mar 2001 10:54:01 -0800 From: John Baron Reply-To: jbaron@sgi.com Organization: SGI Compute Intensive Applications X-Mailer: Mozilla 4.75C-SGI [en] (X11; I; IRIX64 6.5 IP30) X-Accept-Language: en MIME-Version: 1.0 To: Arnaud Desitter CC: pro64-support@oss.sgi.com Subject: Re: IA 64 fortran compiler and getargs References: <200103141810.SAA15829@egham.nag.co.uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing Arnaud Desitter wrote: > I would be nice to have Pro64 support in the F2k proposal > implemented in: > http://www.winteracter.com/f2kcli/ There is also another set of Fortran interfaces to command line arguments, environment variables, and other system services in the Pro64 compilers. They're not documented yet, but you can obtain a list of these routines by searching for INTRO_PXF at techpubs.sgi.com; you can find further documentation (the appropriate calling sequence, at least) by searching TechPubs for specific routines listed in the INTRO_PXF page, e.g., PXFGETARG. John -- John Baron jbaron@sgi.com SGI Performance Engineering and Math Libraries From owner-pro64-support@oss.sgi.com Wed Mar 14 11:05:39 2001 Received: by oss.sgi.com id ; Wed, 14 Mar 2001 11:05:30 -0800 Received: from sgi04-e.std.com ([199.172.62.134]:59463 "EHLO sgi04-e.std.com") by oss.sgi.com with ESMTP id ; Wed, 14 Mar 2001 11:05:16 -0800 Received: from world.std.com (world-f.std.com [199.172.62.5]) by sgi04-e.std.com (8.9.3/8.9.3) with ESMTP id OAA9787921 for ; Wed, 14 Mar 2001 14:05:09 -0500 (EST) From: craig@jcb-sc.com Received: from world.std.com (world.std.com [192.74.137.5]) by world.std.com (8.9.3/8.9.3) with ESMTP id OAA08889 for ; Wed, 14 Mar 2001 14:05:06 -0500 (EST) Received: (qmail 13642 invoked by uid 500); 14 Mar 2001 18:43:12 -0000 Date: 14 Mar 2001 18:43:12 -0000 Message-ID: <20010314184312.13641.qmail@deer> To: rat@cchkms.engr.sgi.com CC: pro64-support@oss.sgi.com In-reply-to: <10103140944.ZM90327@cchkms.engr.sgi.com> (rat@cchkms.engr.sgi.com) Subject: Re: IA 64 fortran compiler and getargs References: <3AAF24B5.225034B3@scali.no> <10103140944.ZM90327@cchkms.engr.sgi.com> Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing >A snippet from the man page for them (man pages are being written) FYI, some useful "source" material might be snarfed from the g77 documentation, assuming the licensing issues work out (which seems reasonable, since both *programs* are GPL'ed, suggesting the docs would also have compatible licenses). While the g77 doc sources are written in texinfo (actually the library/intrinsic docs are written in C macro-defined strings, IIRC, that are pretty mundanely turned into texinfo, but the strings themselves are texinfo), the markup should be fairly easy to discard, or process into man-page source format, focusing on the source material itself. tq vm, (burley) From owner-pro64-support@oss.sgi.com Thu Mar 15 00:36:46 2001 Received: by oss.sgi.com id ; Thu, 15 Mar 2001 00:36:37 -0800 Received: from [162.105.169.99] ([162.105.169.99]:43017 "EHLO webmail.ccer.edu.cn") by oss.sgi.com with ESMTP id ; Thu, 15 Mar 2001 00:36:14 -0800 Received: from www.ccermail.net (IDENT:sysop@webmail.ccer.edu.cn [162.105.169.99]) by webmail.ccer.edu.cn (8.9.3/8.9.3) with SMTP id QAA27839; Thu, 15 Mar 2001 16:36:05 -0600 Message-Id: <20010315.AAA98469550027555@ccermail.net> From: "Hu Yanjun" To: pro64-support@oss.sgi.com Cc: huyanjun@ccermail.net Subject: =?gb2312?B?U2Vla2luZyBGb3J0cmFuOTAncyBmcm9udC1lbmQgYXJjaGl2ZXMh?= Date: Thu, 15 Mar 2001 16:36:05 -0600 MIME-Version: 1.0 X-Priority: 3 X-Mailer: @MESSAGE Multidomain-5.3.0104.1.0105 (Linux 2.2.14-5.0smp) Content-Type: multipart/mixed; boundary="----=_NextPart_5CEJXOT32XYPE7J2T9LK5YES" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;pro64-support-outgoing This is a multi-part message in MIME format. ------=_NextPart_5CEJXOT32XYPE7J2T9LK5YES Content-Type: multipart/alternative; boundary="----=_NextPart_K7QJRU7LCBBF6U9WYCR391U8" ------=_NextPart_K7QJRU7LCBBF6U9WYCR391U8 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: base64 RGVhciBTaXIvTWFkYW0sDQoNCiAgICAgSG93IGFyZSB5b3UhDQogICAgIA0KICAgICBXaGVuIEkg YW0gZGV2ZWxvcGluZyBhIHByb2plY3QgSSB3YW50IHRvIHVzZSANCnRoZSBmb3J0cmFuOTAob3Ig Zm9ydHJhbjc3KSBmcm9udC1lbmQgb2YgUHJvNjQuDQogDQogICAgIENhbiB5b3UgdGVsbCB3aGVy ZSBjYW4gSSBnZXQgdGhlIHJlbGV2YW50IA0KYXJjaGl2ZXMuIFRoYW5rIHlvdSENCg0KWW91cnMg U2luY2VyZWx5LA0KSHUgeWFuanVuDQoNCg0KLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0NCkhpZ2ggUXVhbGl0eSBGUkVFIE1haWwgU3lz dGVtIEZvciBFY29ub21pc3QgDQpodHRwOi8vd3d3LmNjZXJtYWlsLm5ldC8gDQoNCkNoaW5hIENl bnRlciBmb3IgRWNvbm9taWMgUmVzZWFyY2ggDQpodHRwOi8vd3d3LmNjZXIuZWR1LmNuLyANCg0K Q2hpbmEgRWNvbm9taWMgTmV0d29yayANCmh0dHA6Ly93d3cuY2VuZXQub3JnLmNuLyANCg0KSGln aCBRdWFsaXR5IE1haWwgU3lzdGVtIEBDQ0VSTUFJTD8gUG93ZXJlZCBieSAzUiBTb2Z0LCBJbmMu DQotLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLQ0KDQoNCg== ------=_NextPart_K7QJRU7LCBBF6U9WYCR391U8 Content-Type: text/html; charset="gb2312" Content-Transfer-Encoding: base64 PEhUTUw+DQo8SEVBRD48L0hFQUQ+DQo8Qk9EWSBCR0NPTE9SPSJ3aGl0ZSI+DQo8Rk9OVCBTSVpF PTI+DQpEZWFyIFNpci9NYWRhbSw8QlI+DQo8QlI+DQogICAgIEhvdyBhcmUgeW91ITxCUj4NCiAg ICAgPEJSPg0KICAgICBXaGVuIEkgYW0gZGV2ZWxvcGluZyBhIHByb2plY3QgSSB3YW50IHRvIHVz ZSA8QlI+DQp0aGUgZm9ydHJhbjkwKG9yIGZvcnRyYW43NykgZnJvbnQtZW5kIG9mIFBybzY0LjxC Uj4NCiA8QlI+DQogICAgIENhbiB5b3UgdGVsbCB3aGVyZSBjYW4gSSBnZXQgdGhlIHJlbGV2YW50 IDxCUj4NCmFyY2hpdmVzLiBUaGFuayB5b3UhPEJSPg0KPEJSPg0KWW91cnMgU2luY2VyZWx5LDxC Uj4NCkh1IHlhbmp1bjxCUj4NCjxCUj4NCjxCUj4NCi0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tPEJSPg0KPEZPTlQgU0laRT0yPkhpZ2gg UXVhbGl0eSBGUkVFIE1haWwgU3lzdGVtIEZvciBFY29ub21pc3QgPC9GT05UPjxCUj4NCjxGT05U IFNJWkU9Mj5odHRwOi8vd3d3LmNjZXJtYWlsLm5ldC8gPC9GT05UPjxCUj4NCjxCUj4NCjxGT05U IFNJWkU9Mj5DaGluYSBDZW50ZXIgZm9yIEVjb25vbWljIFJlc2VhcmNoIDwvRk9OVD48QlI+DQo8 Rk9OVCBTSVpFPTI+aHR0cDovL3d3dy5jY2VyLmVkdS5jbi8gPC9GT05UPjxicj48QlI+DQo8Rk9O VCBTSVpFPTI+Q2hpbmEgRWNvbm9taWMgTmV0d29yayA8L0ZPTlQ+PEJSPg0KPEZPTlQgU0laRT0y Pmh0dHA6Ly93d3cuY2VuZXQub3JnLmNuLyA8L0ZPTlQ+PEJSPg0KPEJSPg0KPEZPTlQgU0laRT0y PkhpZ2ggUXVhbGl0eSBNYWlsIFN5c3RlbSA8ST5AQ0NFUk1BSUw8L0k+PyBQb3dlcmVkIGJ5IDxB IEhSRUY9Imh0dHA6Ly93d3cuM3Jzb2Z0LmNvbS8iPjNSIFNvZnQ8L0E+LCBJbmMuPC9GT05UPjxC Uj4NCi0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tPEJSPg0KPEJSPjxCUj4NCjxJTUcgU1JDPSJodHRwOi8vd3d3LmNjZXJtYWlsLm5ldDo4 MC9jZ2ktYXV0aC9tYWlscmNwdC5jZ2k/dXNlcmlkPWh1eWFuanVuJm1haWxkb21haW5uYW1lPWNj ZXJtYWlsLm5ldCZtc2dpZD0yMDAxMDMxNS5BQUE5ODQ2OTU1MDAyNzU1NUBjY2VybWFpbC5uZXQi Pg0KPC9GT05UPg0KPC9CT0RZPjwvSFRNTD4NCg0K ------=_NextPart_K7QJRU7LCBBF6U9WYCR391U8-- ------=_NextPart_5CEJXOT32XYPE7J2T9LK5YES-- From owner-pro64-support@oss.sgi.com Thu Mar 15 18:13:13 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2G2DD424233 for pro64-support-outgoing; Thu, 15 Mar 2001 18:13:13 -0800 Received: from hotmail.com (f125.law14.hotmail.com [64.4.21.125]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2G2DAM24230; Thu, 15 Mar 2001 18:13:10 -0800 Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Thu, 15 Mar 2001 18:13:05 -0800 Received: from 159.226.40.178 by lw14fd.law14.hotmail.msn.com with HTTP; Fri, 16 Mar 2001 02:13:05 GMT X-Originating-IP: [159.226.40.178] From: "Yanjun HU" To: owner-pro64-support@oss.sgi.com Cc: pro64-support@oss.sgi.com Subject: Can I tansplant Fortran90 frontend for my use? Date: Fri, 16 Mar 2001 10:13:05 +0800 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 16 Mar 2001 02:13:05.0413 (UTC) FILETIME=[A2DC9350:01C0ADBE] Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Dear pro64-support team, Now we are developing a project called ParaVT which visualize the behavior of a PVM/MPI program. Instrumentation the Fortran source program is the key technology of this project. So the fortran frontend of pro64 is useful for us. Can you tell me whether it is practical about my idea? And If it is practical, can you tell me how long we will finish the transplante and where I can get some supporting materials about fortran frontend of pro64 from internet or somewhere else? Thank you in advance! Yours faithfully, Hu Yanjun _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. From owner-pro64-support@oss.sgi.com Mon Mar 19 00:59:38 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2J8xcA26903 for pro64-support-outgoing; Mon, 19 Mar 2001 00:59:38 -0800 Received: from narkis.wisdom.weizmann.ac.il (narkis.wisdom.weizmann.ac.il [132.76.80.32]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2J8xZM26900 for ; Mon, 19 Mar 2001 00:59:36 -0800 Received: from wisdom.weizmann.ac.il (amir8-pc.wisdom.weizmann.ac.il [132.76.81.32]) by narkis.wisdom.weizmann.ac.il (8.9.3/8.9.3) with ESMTP id KAA02283 for ; Mon, 19 Mar 2001 10:58:27 +0200 (IST) Message-ID: <3AB5CA81.42C09D3F@wisdom.weizmann.ac.il> Date: Mon, 19 Mar 2001 10:59:45 +0200 From: raya Organization: Weizmann Institute of Science X-Mailer: Mozilla 4.75 [en] (WinNT; U) X-Accept-Language: il,en MIME-Version: 1.0 To: pro64 Subject: Software pipelining Content-Type: multipart/mixed; boundary="------------C007729AF831DA6A1CCD8947" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk This is a multi-part message in MIME format. --------------C007729AF831DA6A1CCD8947 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello, I do not understand the code produced by PRO64: For the following C program: int a[100]; int b[100]; main() { int i; for (i =0;i<100;i++) { a[i] = b[i] + 5; } } Compiling with -O2 produced the following pipelined loop: { .mii (p19) st4 [r3]=r35,4 // [3*II+0] id:12 a+0x0 (p18) adds r34=5,r37 // [2*II+0] (p18) nop.i 0 // [2*II+0] }; { .mfb (p16) ld4 r35=[r2],4 // [0*II+0] id:11 b+0x0 (p16) nop.f 0 // [0*II+0] br.ctop.dptk.few .Lt_0_4 ;; // [3*II+0] }; Is this a legal code? there is a store from r35 and load to r35 within the same instruction group. (The full assembly code, as well as the C source are attached). -- Raya Leviathan Tel. 972-8-9344208 (office) Tel. 972-3-6358481 (home) Email: raya@wisdom.weizmann.ac.il --------------C007729AF831DA6A1CCD8947 Content-Type: application/x-unknown-content-type-S_auto_file; name="simpleO2.s" Content-Transfer-Encoding: base64 Content-Disposition: inline; filename="simpleO2.s" CS8vICAvdXNyL2xpYi9nY2MtbGliL2lhNjQtc2dpLWxpbnV4L3NnaWNjLTEuMC9iZTo6MC4w MS4wLTkKCgkvLy0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tCgkvLyBDb21waWxpbmcgc2ltcGxlLmMgKC90bXAvY2NCLm81QXZX UCkKCS8vLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0KCgkvLy0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tCgkvLyBPcHRpb25zOgoJLy8tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLQoJLy8gIFRhcmdl dDpJdGFuaXVtLCBJU0E6SVNBXzEsIFBvaW50ZXIgU2l6ZTo2NAoJLy8gIC1PMgkoT3B0aW1p emF0aW9uIGxldmVsKQoJLy8gIC1nMAkoRGVidWcgbGV2ZWwpCgkvLyAgLW0xCShSZXBvcnQg d2FybmluZ3MpCgkvLy0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tCgoJLmNvbW1vbglhIywgNDAwLCA0CgkuY29tbW9uCWIjLCA0 MDAsIDQKCgkuc2VjdGlvbiAudGV4dCwgImF4IiwgInByb2diaXRzIgoJLmFsaWduCTE2Cgku c2VjdGlvbiAudGV4dAoKCS8vIFByb2dyYW0gVW5pdDogbWFpbgoJLnByb2MJbWFpbiMKCS5n bG9iYWwJbWFpbiMKbWFpbiM6CS8vIDB4MAoJLmZpbGUJMCAiL2hvbWUvcmF5YS9iZW5jaG1h cmtzL3NpbXBsZV9sb29wL3NpbXBsZS5jIgoJLmxvYwkwCTQJMAovLyAgIDEgIGludCBhWzEw MF07Ci8vICAgMiAgaW50IGJbMTAwXTsKLy8gICAzICBtYWluKCkKLy8gICA0ICB7Ci5CQjFf bWFpbjoJLy8gMHgwCi8vPGZyZXE+Ci8vPGZyZXE+IEJCOjEgZnJlcXVlbmN5ID0gMS4wMDAw MCAoaGV1cmlzdGljKQovLzxmcmVxPgogewkgICAgICAubWlpCi8vCS5zYXZlIGFyLnBmcwoJ ICAgICAgYWxsb2Mgcjg9YXIucGZzLDAsOCwwLDggCS8vIFswXSAgCgkgICAgICBhZGRsIHIz PUBsdG9mZihhIyksZ3AgICAJLy8gWzBdICBhCgkgICAgICBhZGRsIHIyPUBsdG9mZihiIyks Z3AgICAJLy8gWzBdICBiCiB9OyB7CSAgICAgIC5taWkKCSAgICAgIG5vcC5tIDAgICAgICAg ICAgICAgICAgIAkvLyAgCgkgICAgICBtb3YgcjE0PXByICAgICAgICAgICAgICAJLy8gWzBd ICAKCSAgICAgIG5vcC5pIDAgOzsgICAgICAgICAgICAgIAkvLyAgCiB9OyB7CSAgICAgIC5t bWkKCSAgICAgIGxkOCByMj1bcjJdICAgICAgICAgICAgIAkvLyBbMl0gIAoJICAgICAgbGQ4 IHIzPVtyM10gICAgICAgICAgICAgCS8vIFsyXSAgCgkgICAgICBub3AuaSAwIDs7ICAgICAg ICAgICAgICAJLy8gIAogfTsKLkJCNF9tYWluOgkvLyAweDMwCi8vPGZyZXE+Ci8vPGZyZXE+ IEJCOjQgZnJlcXVlbmN5ID0gMS4wMDAwMCAoaGV1cmlzdGljKQovLzxmcmVxPgogewkgICAg ICAubWliCgkgICAgICBtb3Ygcjk9OTkgICAgICAgICAgICAgICAJLy8gWzBdICAKCSAgICAg IG1vdiByMTA9NCAgICAgICAgICAgICAgIAkvLyBbMF0gIAoJICAgICAgY2xycnJiLnByIDs7 ICAgICAgICAgICAgCS8vIFswXSAgCiB9OyB7CSAgICAgIC5taWkKCSAgICAgIG5vcC5tIDAg ICAgICAgICAgICAgICAgIAkvLyAgCgkgICAgICBtb3YgcHIucm90PTY1NTM2ICAgICAgICAJ Ly8gWzBdICAKCSAgICAgIG1vdi5pIGFyLmVjPXIxMCAgICAgICAgIAkvLyBbMV0gIAogfTsg ewkgICAgICAubWlpCgkgICAgICBub3AubSAwICAgICAgICAgICAgICAgICAJLy8gIAoJICAg ICAgbW92LmkgYXIubGM9cjkgICAgICAgICAgCS8vIFsxXSAgCgkgICAgICBub3AuaSAwIDs7 ICAgICAgICAgICAgICAJLy8gIAogfTsKLkx0XzBfNDoJLy8gMHg2MAovLzxsb29wPiBMb29w IGJvZHkgbGluZSA0LCBuZXN0aW5nIGRlcHRoOiAxLCBpdGVyYXRpb25zOiAxMDAKLy88c3dw cz4gCi8vPHN3cHM+ICAgMSBjeWNsZXMgcGVyIDEgaXRlcmF0aW9uIGluIHN0ZWFkeSBzdGF0 ZQovLzxzd3BzPiAgIDQgcGlwZWxpbmUgc3RhZ2VzCi8vPHN3cHM+IAovLzxzd3BzPiAgICAg IG1pbiAxIGN5Y2xlcyByZXF1aXJlZCBieSByZXNvdXJjZXMKLy88c3dwcz4gICAgICBtaW4g MSBjeWNsZXMgcmVxdWlyZWQgYnkgcmVzb3VyY2VzL3JlY3VycmVuY2UKLy88c3dwcz4gICAg ICBtaW4gNCBjeWNsZXMgKGFjdHVhbCA0IGN5Y2xlcykgcmVxdWlyZWQgdG8gc2NoZWR1bGUg b25lIGl0ZXJhdGlvbgovLzxzd3BzPiAgICAgIAovLzxzd3BzPiAgICAgIDQgaXNzdWUgdW5p dHMgKCA2NiUgb2YgcGVhayApCi8vPHN3cHM+ICAgICAgMyBJMC1JMS1NTy1NMSB1bml0cyAo IDc1JSBvZiBwZWFrICkKLy88c3dwcz4gICAgICAyIG1lbW9yeSB1bml0cyAoIDEwMCUgb2Yg cGVhayApCi8vPHN3cHM+ICAgICAgMSBicmFuY2ggdW5pdHMgKCAzMyUgb2YgcGVhayApCi8v PHN3cHM+ICAgICAgCi8vPGZyZXE+Ci8vPGZyZXE+IEJCOjIgZnJlcXVlbmN5ID0gMTAwLjAw MDAwIChoZXVyaXN0aWMpCi8vPGZyZXE+IEJCOjIgPT4gQkI6NSBwcm9iYWJpbGl0eSA9IDAu MDEwMDAKLy88ZnJlcT4gQkI6MiA9PiBCQjoyIHByb2JhYmlsaXR5ID0gMC45OTAwMAovLzxm cmVxPgogewkgICAgICAubWlpCgkocDE5KSBzdDQgW3IzXT1yMzUsNCAgICAgICAgICAJLy8g WzMqSUkrMF0gIGlkOjEyIGErMHgwCgkocDE4KSBhZGRzIHIzND01LHIzNyAgICAgICAgICAJ Ly8gWzIqSUkrMF0gIAoJKHAxOCkgbm9wLmkgMCAgICAgICAgICAgICAgICAgCS8vIFsyKklJ KzBdICAKIH07IHsJICAgICAgLm1mYgoJKHAxNikgbGQ0IHIzNT1bcjJdLDQgICAgICAgICAg CS8vIFswKklJKzBdICBpZDoxMSBiKzB4MAoJKHAxNikgbm9wLmYgMCAgICAgICAgICAgICAg ICAgCS8vIFswKklJKzBdICAKCSAgICAgIGJyLmN0b3AuZHB0ay5mZXcgLkx0XzBfNCA7Owkv LyBbMypJSSswXSAgCiB9OwouQkI1X21haW46CS8vIDB4ODAKLy88ZnJlcT4KLy88ZnJlcT4g QkI6NSBmcmVxdWVuY3kgPSAxLjAwMDAwIChoZXVyaXN0aWMpCi8vPGZyZXE+CiB7CSAgICAg IC5taWIKCSAgICAgIG5vcC5tIDAgICAgICAgICAgICAgICAgIAkvLyAgCgkgICAgICBub3Au aSAwICAgICAgICAgICAgICAgICAJLy8gIAoJICAgICAgY2xycnJiIDs7ICAgICAgICAgICAg ICAgCS8vIFswXSAgCiB9OwouQkIzX21haW46CS8vIDB4OTAKLy88ZnJlcT4KLy88ZnJlcT4g QkI6MyBmcmVxdWVuY3kgPSAxLjAwMDAwIChoZXVyaXN0aWMpCi8vPGZyZXE+CiB7CSAgICAg IC5taWkKCSAgICAgIG5vcC5tIDAgICAgICAgICAgICAgICAgIAkvLyAgCgkgICAgICBtb3Yg cHI9cjE0LC0xICAgICAgICAgICAJLy8gWzBdICAKCSAgICAgIG5vcC5pIDAgOzsgICAgICAg ICAgICAgIAkvLyAgCiB9OyB7CSAgICAgIC5taWIKCSAgICAgIG5vcC5tIDAgICAgICAgICAg ICAgICAgIAkvLyAgCgkgICAgICBub3AuaSAwICAgICAgICAgICAgICAgICAJLy8gIAoJICAg ICAgYnIucmV0LnNwdGsubWFueSBiMCA7OyAgCS8vIFswXSAgCiB9OwoJLmVuZHAJbWFpbiMK Ly8gZW1pdCB1bndpbmQgaW5mbwoJLnNlY3Rpb24gLklBXzY0LnVud2luZF9pbmZvCi5MdW53 aW5kX2luZm9fMDoKCWRhdGE4IDB4MTAwMDAwMDAwMDAwMQoJZGF0YTggMHg4YjEwMGU2MjA2 MAoJLnNlY3Rpb24gLklBXzY0LnVud2luZAoJZGF0YTggQHNlZ3JlbChtYWluIykKCWRhdGE4 IEBzZWdyZWwobWFpbiMrMHhiMCkKCWRhdGE4IEBzZWdyZWwoLkx1bndpbmRfaW5mb18wKQoJ LnNlY3Rpb24gLnRleHQKCS5hbGlnbiAxNgovLwkuZ3B2YWx1ZSAwCgoJLnNlY3Rpb24gLmRl YnVnX2luZm8sICIiLCAicHJvZ2JpdHMiCgkuYWxpZ24JMAoJZGF0YTEJMHgzYSwgMHgwMCwg MHgwMCwgMHgwMCwgMHgwMiwgMHgwMAoJZGF0YTQudWEJLmRlYnVnX2FiYnJldgoJZGF0YTEJ MHgwOCwgMHgwMSwgMHg3MywgMHg2OSwgMHg2ZCwgMHg3MCwgMHg2YywgMHg2NQoJZGF0YTEJ MHgyZSwgMHg2MywgMHgwMCwgMHgyZCwgMHg0ZiwgMHgzMiwgMHgwMCwgMHgwMQoJZGF0YTEJ MHgwMAoJZGF0YTQudWEJLmRlYnVnX2xpbmUKCWRhdGExCTB4MDIsIDB4MDEsIDB4MDQsIDB4 NmQsIDB4NjEsIDB4NjksIDB4NmUsIDB4MDAKCWRhdGExCTB4MDEsIDB4MDMsIDB4OTIsIDB4 MGMsIDB4MDAKCWRhdGE4LnVhCS5CQjFfbWFpbgoJZGF0YTgudWEJLkJCMV9tYWluICsgMHhi MAoJZGF0YTEJMHgwMCwgMHgwMAoKCS5zZWN0aW9uIC5kZWJ1Z19hcmFuZ2VzLCAiIiwgInBy b2diaXRzIgoJLmFsaWduCTAKCWRhdGExCTB4MmMsIDB4MDAsIDB4MDAsIDB4MDAsIDB4MDIs IDB4MDAKCWRhdGE0LnVhCS5kZWJ1Z19pbmZvCglkYXRhMQkweDA4LCAweDAwLCAweDAwLCAw eDAwLCAweDAwLCAweDAwCglkYXRhOC51YQkuQkIxX21haW4KCWRhdGE4LnVhCS5CQjFfbWFp biAtIC5CQjFfbWFpbiArIDB4YjAKCWRhdGExCTB4MDAsIDB4MDAsIDB4MDAsIDB4MDAsIDB4 MDAsIDB4MDAsIDB4MDAsIDB4MDAKCWRhdGExCTB4MDAsIDB4MDAsIDB4MDAsIDB4MDAsIDB4 MDAsIDB4MDAsIDB4MDAsIDB4MDAKCgkuc2VjdGlvbiAuZGVidWdfcHVibmFtZXMsICIiLCAi cHJvZ2JpdHMiCgkuYWxpZ24JMAoJZGF0YTEJMHgxNywgMHgwMCwgMHgwMCwgMHgwMCwgMHgw MiwgMHgwMAoJZGF0YTQudWEJLmRlYnVnX2luZm8KCWRhdGExCTB4MDAsIDB4MDAsIDB4MDAs IDB4MDAsIDB4MWYsIDB4MDAsIDB4MDAsIDB4MDAKCWRhdGExCTB4NmQsIDB4NjEsIDB4Njks IDB4NmUsIDB4MDAsIDB4MDAsIDB4MDAsIDB4MDAKCWRhdGExCTB4MDAKCgkuc2VjdGlvbiAu ZGVidWdfYWJicmV2LCAiIiwgInByb2diaXRzIgoJLmFsaWduCTAKCWRhdGExCTB4MDEsIDB4 MTEsIDB4MDEsIDB4MDMsIDB4MDgsIDB4MjUsIDB4MDgsIDB4MTMKCWRhdGExCTB4MGIsIDB4 NDIsIDB4MGIsIDB4MTAsIDB4MDYsIDB4MDAsIDB4MDAsIDB4MDIKCWRhdGExCTB4MmUsIDB4 MDAsIDB4M2EsIDB4MGIsIDB4M2IsIDB4MGIsIDB4MDMsIDB4MDgKCWRhdGExCTB4M2YsIDB4 MGMsIDB4NDAsIDB4MGEsIDB4MTEsIDB4MDEsIDB4MTIsIDB4MDEKCWRhdGExCTB4MDAsIDB4 MDAK --------------C007729AF831DA6A1CCD8947 Content-Type: application/x-unknown-content-type-c_auto_file; name="simple.c" Content-Transfer-Encoding: base64 Content-Disposition: inline; filename="simple.c" aW50IGFbMTAwXTsKaW50IGJbMTAwXTsKbWFpbigpCnsKICBpbnQgaTsKICBmb3IgKGkgPTA7 aTwxMDA7aSsrKQogICAgewogICAgICBhW2ldID0gYltpXSArIDU7CiAgICB9Cn0K --------------C007729AF831DA6A1CCD8947-- From owner-pro64-support@oss.sgi.com Mon Mar 19 09:44:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2JHi0U04077 for pro64-support-outgoing; Mon, 19 Mar 2001 09:44:00 -0800 Received: from therouter.routefree.com ([209.21.47.226]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2JHhxM04074 for ; Mon, 19 Mar 2001 09:43:59 -0800 Received: from routefree.com (IDENT:lo@osprey.routefree.com [10.0.0.74]) by therouter.routefree.com (8.9.3/8.9.3) with ESMTP id JAA14884; Mon, 19 Mar 2001 09:43:49 -0800 Message-ID: <3AB64554.531E06E5@routefree.com> Date: Mon, 19 Mar 2001 09:43:48 -0800 From: Raymond Lo Organization: RouteFree Inc. X-Mailer: Mozilla 4.73 [en] (X11; U; Linux 2.2.14-5.0 i686) X-Accept-Language: en MIME-Version: 1.0 To: raya CC: pro64 Subject: Re: Software pipelining References: <3AB5CA81.42C09D3F@wisdom.weizmann.ac.il> Content-Type: text/plain; charset=big5 Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk If my memory is correct, Itanium specifies that register write-after-read dependence at the same cycle is legal. Please check the Itanium manuals for confirmation. -Raymond raya wrote: > Hello, > I do not understand the code produced by PRO64: > For the following C program: > > int a[100]; > int b[100]; > main() > { > int i; > for (i =0;i<100;i++) > { > a[i] = b[i] + 5; > } > } > > Compiling with -O2 produced the following pipelined loop: > { .mii > (p19) st4 [r3]=r35,4 // [3*II+0] id:12 a+0x0 > (p18) adds r34=5,r37 // [2*II+0] > (p18) nop.i 0 // [2*II+0] > }; { .mfb > (p16) ld4 r35=[r2],4 // [0*II+0] id:11 b+0x0 > (p16) nop.f 0 // [0*II+0] > br.ctop.dptk.few .Lt_0_4 ;; // [3*II+0] > }; > > Is this a legal code? there is a store from r35 and load to r35 within > the same instruction group. > > (The full assembly code, as well as the C source are attached). > -- > Raya Leviathan > Tel. 972-8-9344208 (office) > Tel. 972-3-6358481 (home) > Email: raya@wisdom.weizmann.ac.il > > ------------------------------------------------------------------------------ > Name: simpleO2.s > simpleO2.s Type: application/x-unknown-content-type-S_auto_file > Encoding: base64 > > Name: simple.c > simple.c Type: application/x-unknown-content-type-c_auto_file > Encoding: base64 From owner-pro64-support@oss.sgi.com Mon Mar 19 10:02:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2JI23Y04839 for pro64-support-outgoing; Mon, 19 Mar 2001 10:02:03 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2JI23M04836 for ; Mon, 19 Mar 2001 10:02:03 -0800 Received: (from localhost user: 'mcnaught' uid#11891 fake: STDIN (mcnaught@tees.cs.ualberta.ca)) by scapa.cs.ualberta.ca id ; Mon, 19 Mar 2001 11:01:52 -0700 Date: Mon, 19 Mar 2001 11:01:51 -0700 (MST) From: Matthew McNaughton To: raya cc: pro64 Subject: Re: Software pipelining In-Reply-To: <3AB5CA81.42C09D3F@wisdom.weizmann.ac.il> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk On Mon, 19 Mar 2001, raya wrote: > for (i =0;i<100;i++) > { > a[i] = b[i] + 5; > } > } > > Compiling with -O2 produced the following pipelined loop: > { .mii > (p19) st4 [r3]=r35,4 // [3*II+0] id:12 a+0x0 > (p18) adds r34=5,r37 // [2*II+0] > (p18) nop.i 0 // [2*II+0] > }; { .mfb > (p16) ld4 r35=[r2],4 // [0*II+0] id:11 b+0x0 > (p16) nop.f 0 // [0*II+0] > br.ctop.dptk.few .Lt_0_4 ;; // [3*II+0] > }; > > > Is this a legal code? there is a store from r35 and load to r35 within > the same instruction group. Notice also how r37 is used and never defined. And where did the predicates (p19), (p18), (p16) come from? The ia64's "br.ctop" instruction renames (some) of the registers and sets these predicates for software pipelining. Here is a fine presentation on the topic: http://www.cs.ualberta.ca/~amaral/courses/680/webslides/TA-HWSupSoftPipeline/sld001.htm -- Matthew McNaughton From owner-pro64-support@oss.sgi.com Mon Mar 19 13:17:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2JLH2t09230 for pro64-support-outgoing; Mon, 19 Mar 2001 13:17:02 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2JLGvM09227; Mon, 19 Mar 2001 13:16:58 -0800 Received: from dsl-proof.corp.sgi.com (dsl-proof.corp.sgi.com [192.102.142.250]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id NAA02884; Mon, 19 Mar 2001 13:26:46 -0800 (PST) mail_from (dlstephe@sgi.com) Received: from sgi.com (localhost [127.0.0.1]) by dsl-proof.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id NAA42257; Mon, 19 Mar 2001 13:15:58 -0800 (PST) Message-ID: <3AB6770D.36C36721@sgi.com> Date: Mon, 19 Mar 2001 13:15:57 -0800 From: David Stephenson Organization: SGI -- Compilers X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: Yanjun HU CC: owner-pro64-support@oss.sgi.com, pro64-support@oss.sgi.com Subject: Re: Can I tansplant Fortran90 frontend for my use? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Yanjun HU wrote: > Now we are developing a project called ParaVT which visualize the > behavior of a PVM/MPI program. Instrumentation the Fortran source program is > the key technology of this project. So the fortran frontend of pro64 is > useful for us. Feedback instrumentation is performed early in the back end (just before lowering from VH whirl to H whirl), but not in the front end. So the whirl code is instrumented, instead of the fortran. It might be possible to match the data back to the original fortran code. - David -- David Stephenson http://reality.sgi.com/dlstephe_engr/ From owner-pro64-support@oss.sgi.com Tue Mar 20 12:02:22 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2KK2Mk01625 for pro64-support-outgoing; Tue, 20 Mar 2001 12:02:22 -0800 Received: from paco.llnl.gov (paco.llnl.gov [134.9.17.119]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2KK2MM01622 for ; Tue, 20 Mar 2001 12:02:22 -0800 Received: from paco.llnl.gov (IDENT:waynev@localhost [127.0.0.1]) by paco.llnl.gov (SGI-8.9.3/8.9.3) with ESMTP id MAA05879 for ; Tue, 20 Mar 2001 12:01:48 -0800 (PST) Message-ID: <3AB7B72C.C652B37D@paco.llnl.gov> Date: Tue, 20 Mar 2001 12:01:48 -0800 From: Wayne Vieira Reply-To: waynev@sgi.com Organization: SGI X-Mailer: Mozilla 4.75C-SGI [en] (X11; U; IRIX64 6.5 IP30) X-Accept-Language: en MIME-Version: 1.0 To: pro64-support@oss.sgi.com Subject: ld: could not read symbols when using -Ofast Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk I just installed pro64-0.01.0-13 on a Big Sur system running TL1123, and things seem to be going well. While -O3 works fine on one of my codes, -Ofast gives me the following error on linking: sgif90 -r8 main.o sppm.o runhyd3.o bdrys.o cpthreads_sppm_IA64.o c_io.o timers.o -o run/sppm main.o: could not read symbols: Bad value collect2: ld returned 1 exit status make: *** [run/sppm] Error 2 If I add the -Ofast to the line I get an _extremely_ long link time, with the final error looking something like this: ./run/sppm.ipaqdO6rs//1.o: In function `hydxx_': ./run/sppm.ipaqdO6rs//1.o(.text+0xef90): relocation truncated to fit: GPREL22 volit_ ./run/sppm.ipaqdO6rs//1.o(.text+0xf602): relocation truncated to fit: GPREL22 vectr1_ Wayne From owner-pro64-support@oss.sgi.com Tue Mar 20 13:45:02 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2KLj2Q03606 for pro64-support-outgoing; Tue, 20 Mar 2001 13:45:02 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2KLj2M03603 for ; Tue, 20 Mar 2001 13:45:02 -0800 Received: (from localhost user: 'cam' uid#302 fake: STDIN (cam@sunset.cs.ualberta.ca)) by scapa.cs.ualberta.ca id ; Tue, 20 Mar 2001 14:44:38 -0700 Date: Tue, 20 Mar 2001 14:44:37 -0700 (MST) From: "A. Cameron Macdonell" To: pro64-support@oss.sgi.com Subject: A question on feedback in the Pro64 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk My question is: do you know if SGI introduced automatic profiling in the PRo64. Or else any facilities to enable profile feedback optimization. For instance if I have an aplication, say matrix.c, and a typical input file, say matrix.inp. What would be the compilation steps that I would have to do to make sure that the Pro64 hyperblock formation uses the edge frequency in the CFG from my input file. I can find lots of places in the code where it says it will feedback data if exists but I can't determine how or where its being outputted or how to trace it. A pointer to the right place is as good as a direct answer. Thanks, Cam From owner-pro64-support@oss.sgi.com Tue Mar 20 14:21:21 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2KMLLA04597 for pro64-support-outgoing; Tue, 20 Mar 2001 14:21:21 -0800 Received: from mail-in.hq.tensilica.com ([38.170.141.29]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2KMLKM04594 for ; Tue, 20 Mar 2001 14:21:20 -0800 Received: from sys33 (sys33.hq.tensilica.com [192.168.10.2]) by mail-in.hq.tensilica.com (8.9.3/8.9.3) with SMTP id OAA09193 for ; Tue, 20 Mar 2001 14:21:20 -0800 From: "Darin Petkov" To: Subject: invalid use of void expression Date: Tue, 20 Mar 2001 14:21:19 -0800 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hello, I am getting an error from the sgicc's front-end on this program: #include typedef void my_void; my_void blah (int i) { printf("%d\n",i); return; } main () { blah(3); } > sgicc void.c void.c: In function `main': void.c:11: invalid use of void expression > uname -a Linux heart 2.2.16-3 #1 Mon Jun 19 19:11:44 EDT 2000 ia64 unknown > sgicc -version SGIcc Compilers: Version 0.01.0-12 Anybody have a fix? Thanks, Darin From owner-pro64-support@oss.sgi.com Fri Mar 23 12:46:16 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2NKkGx09345 for pro64-support-outgoing; Fri, 23 Mar 2001 12:46:16 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2NKkFM09342 for ; Fri, 23 Mar 2001 12:46:15 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Fri, 23 Mar 2001 13:45:35 -0700 Date: Fri, 23 Mar 2001 13:45:25 -0700 (MST) From: Peng Zhao To: sgi Subject: question on -tr(s)17 ~ -tr(s)21 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hi, I have some problem in generating the IR or ST trace file using -tr17~-tr21. My command is like following: sgicc -show -Wb,-tr17 -INLINE:must=callit test.c The file test.t is generated but the size is 0. Why? Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Fri Mar 23 14:47:29 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2NMlTI11613 for pro64-support-outgoing; Fri, 23 Mar 2001 14:47:29 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2NMlSM11610 for ; Fri, 23 Mar 2001 14:47:28 -0800 Received: from rohi.engr.sgi.com (rohi.engr.sgi.com [130.62.180.74]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id OAA00676 for ; Fri, 23 Mar 2001 14:57:29 -0800 (PST) mail_from (mpm@rohi.engr.sgi.com) Received: (from mpm@localhost) by rohi.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id OAA64328; Fri, 23 Mar 2001 14:45:50 -0800 (PST) Date: Fri, 23 Mar 2001 14:45:50 -0800 (PST) From: mpm@rohi.engr.sgi.com (Michael Murphy) Message-Id: <200103232245.OAA64328@rohi.engr.sgi.com> To: sgi , Peng Zhao Subject: Re: question on -tr(s)17 ~ -tr(s)21 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk As it says in the howto-debug-compiler file (newly released in v13): The -tr and -ts dump the IR and symbol table respectively after the specified phase. The IR is WHIRL until the time of code generation, at which point the OPS are dumped. These options only work if the phase supports them; some phases do not support the dumping of IR and/or symtab. In this case, there is no support for dumping IR after inlining. Typically, what you do instead is ir_b2a of the .I file, since the inliner is its own phase. -- Mike Murphy -- mpm@sgi.com -- quote of the day: -- "Everybody is a victim, everybody can prove that they were oppressed -- somewhere, everybody has a righteous grudge." (Richard Rohr) From owner-pro64-support@oss.sgi.com Fri Mar 23 18:31:57 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2O2Vvt17233 for pro64-support-outgoing; Fri, 23 Mar 2001 18:31:57 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2O2VvM17230 for ; Fri, 23 Mar 2001 18:31:57 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Fri, 23 Mar 2001 19:31:27 -0700 Date: Fri, 23 Mar 2001 19:31:26 -0700 (MST) From: Peng Zhao To: sgi Subject: a question about the whirl node instruction. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk HI, I have some question about the meaning of the instructions of the whirl IR, e.g. following is one line taken from the result of "ir_b2a xxx.I", it simplements "i++" ... LOC 1 19 I4I4LDID 0 <2,2,i> T<4, .predef_I4, 4> I4INTCONST 1 (0x1) I4ADD I4STID 0 <2,2,i> T<4, .predef_I4, 4> ... what is the meaning of <2,2,i>, T<4, .predef_I4, 4>, I4I4 , "LOC 1 19"? Although I can guess something, it is still not very clear to me. Can somebody tell me the meaning of these fields explicitly and the rule to diciper them? I cannot find the description of them in the Whirl specificatoin. Thank you very much. BTW: It seems to me that there is a little bug in the LOC(source position information) related things in ir_b2a. Sometimes, the LOC appears before the other codes of the same whirl node, but sometimes it appears behind them. It is somewhat confusing. Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Sat Mar 24 18:27:06 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2P2R6507365 for pro64-support-outgoing; Sat, 24 Mar 2001 18:27:06 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2P2R5M07362 for ; Sat, 24 Mar 2001 18:27:05 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@sunnybrook.cs.ualberta.ca)) by scapa.cs.ualberta.ca id ; Sat, 24 Mar 2001 19:26:47 -0700 Date: Sat, 24 Mar 2001 19:26:47 -0700 (MST) From: Peng Zhao To: sgi Subject: error on -fb_create in Pro64 " Message-ID: MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="-2138968809-1880008973-985487207=:9039" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. Send mail to mime@docserver.cac.washington.edu for more info. ---2138968809-1880008973-985487207=:9039 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, Enclosed file is the output of the command "sgicc -show -fb_create . -O0 main.c" The error is like this: main.o(.text+0xa2): undefined reference to `__profile_pu_init' Actually, when I run the above command on another MIPS 10000 2.7 machine (IRIX 6.5), I found that cc generate nothing. It seems the WN_Instrument can never be executed, because I cannot find the place the varialbe "Instrumentation_Enabled" is set to TRUE. Thank you very much. Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 ---2138968809-1880008973-985487207=:9039 Content-Type: TEXT/PLAIN; charset=US-ASCII; name=error Content-Transfer-Encoding: BASE64 Content-ID: Content-Description: Content-Disposition: attachment; filename=error L3Vzci9iaW4vZ2NjIC1EX0xBTkdVQUdFX0MgLURfU0dJX0NPTVBJTEVSX1ZF UlNJT049MDAxIC1EX19ob3N0X2lhNjQgLURfX0lOTElORV9JTlRSSU5TSUNT IC1EX0xQNjQgLURfX2lhNjQ9MSBtYWluLmMgLUUgPiAvdG1wL2NjaS5FMHhL bnkgDQovdXNyL2xpYi9nY2MtbGliL2lhNjQtc2dpLWxpbnV4L3NnaWNjLTEu MC9nZmVjIC1PUFQ6aW5zdHI9MDppbnN0cl91bmlxdWVfb3V0cHV0PW9uIC1P MCAtZHggLXF1aWV0IC1kdW1wYmFzZSBtYWluLmMgL3RtcC9jY2kuRTB4S255 IC1vIC90bXAvY2NCLmU4Q2Q4WCANCi91c3IvbGliL2djYy1saWIvaWE2NC1z Z2ktbGludXgvc2dpY2MtMS4wL2JlIC1MSVNUOnNvdXJjZT1vZmY6bm90ZXM9 b2ZmIC1QSEFTRTpjIC1HOCAtVEVOVjpQSUMgLW0xIC1JTlRFUk5BTDpyZXR1 cm5fdmFsPW9uIC1JTlRFUk5BTDptbGRpZF9tc3RpZD1vbiAtSU5URVJOQUw6 cmV0dXJuX2luZm89b24gLXNob3cgLU9QVDppbnN0cj0wOmluc3RyX3VuaXF1 ZV9vdXRwdXQ9b24gLU8wIC1UQVJHOmFiaT1pNjQgLUxBTkc6PWFuc2lfYyAt ZkIsL3RtcC9jY0IuZThDZDhYIC1maSwuLmluc3RyIC1zIC1mcywvdG1wL2Nj cy5FT0NMRnEgbWFpbi5jIA0KQ29tcGlsaW5nIG1haW4uYyAoL3RtcC9jY0Iu ZThDZDhYKSAtLSBCYWNrIEVuZA0KQ29tcGlsaW5nIG1haW4oMCkNCkNvbXBp bGluZyBJbm5lclByb2R1Y3QoMSkNCg0KL3Vzci9saWIvLi4vYmluL2FzIC90 bXAvY2NzLkVPQ0xGcSAtbyBtYWluLm8gDQovdXNyL2Jpbi9nY2MgLUwvdXNy L2xpYi9nY2MtbGliL2lhNjQtc2dpLWxpbnV4L3NnaWNjLTEuMCBtYWluLm8g DQptYWluLm86L3Vzci9zYWt3YXRhbWF1L2JydWxlMTIvZ3JhZC9wZW5nemhh by90ZW1wL21haW4vbWFpbi5jOjUyOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRv IGBfX3Byb2ZpbGVfaW5pdCcNCm1haW4ubzogSW4gZnVuY3Rpb24gYG1haW4n Og0KbWFpbi5vKC50ZXh0KzB4YTIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRv IGBfX3Byb2ZpbGVfcHVfaW5pdCcNCm1haW4ubygudGV4dCsweGMyKTogdW5k ZWZpbmVkIHJlZmVyZW5jZSB0byBgX19wcm9maWxlX2ludm9rZV9pbml0Jw0K bWFpbi5vKC50ZXh0KzB4ZTIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBf X3Byb2ZpbGVfYnJhbmNoX2luaXQnDQptYWluLm8oLnRleHQrMHgxMDIpOiB1 bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2ZpbGVfbG9vcF9pbml0Jw0K bWFpbi5vKC50ZXh0KzB4MTIyKTogdW5kZWZpbmVkIHJlZmVyZW5jZSB0byBg X19wcm9maWxlX2NhbGxfaW5pdCcNCm1haW4ubygudGV4dCsweDE0Mik6IHVu ZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9pbnZva2UnDQptYWlu Lm8oLnRleHQrMHgxYTIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3By b2ZpbGVfYnJhbmNoJw0KbWFpbi5vKC50ZXh0KzB4MWYyKTogdW5kZWZpbmVk IHJlZmVyZW5jZSB0byBgX19wcm9maWxlX2NhbGxfZW50cnknDQptYWluLm8o LnRleHQrMHgyNjIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2Zp bGVfY2FsbF9leGl0Jw0KbWFpbi5vKC50ZXh0KzB4MjkyKTogdW5kZWZpbmVk IHJlZmVyZW5jZSB0byBgX19wcm9maWxlX2NhbGxfZW50cnknDQptYWluLm8o LnRleHQrMHgyZDIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2Zp bGVfY2FsbF9leGl0Jw0KbWFpbi5vKC50ZXh0KzB4MzEyKTogdW5kZWZpbmVk IHJlZmVyZW5jZSB0byBgX19wcm9maWxlX2NhbGxfZW50cnknDQptYWluLm8o LnRleHQrMHgzODIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2Zp bGVfY2FsbF9leGl0Jw0KbWFpbi5vKC50ZXh0KzB4M2YyKTogdW5kZWZpbmVk IHJlZmVyZW5jZSB0byBgX19wcm9maWxlX2JyYW5jaCcNCm1haW4ubygudGV4 dCsweDQ0Mik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9j YWxsX2VudHJ5Jw0KbWFpbi5vKC50ZXh0KzB4NGIyKTogdW5kZWZpbmVkIHJl ZmVyZW5jZSB0byBgX19wcm9maWxlX2NhbGxfZXhpdCcNCm1haW4ubygudGV4 dCsweDRlMik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9j YWxsX2VudHJ5Jw0KbWFpbi5vKC50ZXh0KzB4NTIyKTogdW5kZWZpbmVkIHJl ZmVyZW5jZSB0byBgX19wcm9maWxlX2NhbGxfZXhpdCcNCm1haW4ubygudGV4 dCsweDU1Mik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9s b29wJw0KbWFpbi5vKC50ZXh0KzB4NWEyKTogdW5kZWZpbmVkIHJlZmVyZW5j ZSB0byBgX19wcm9maWxlX2xvb3BfaXRlcicNCm1haW4ubygudGV4dCsweDg0 Mik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9jYWxsX2Vu dHJ5Jw0KbWFpbi5vKC50ZXh0KzB4OGMyKTogdW5kZWZpbmVkIHJlZmVyZW5j ZSB0byBgX19wcm9maWxlX2NhbGxfZXhpdCcNCm1haW4ubygudGV4dCsweDkx Mik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9jYWxsX2Vu dHJ5Jw0KbWFpbi5vKC50ZXh0KzB4OTkyKTogdW5kZWZpbmVkIHJlZmVyZW5j ZSB0byBgX19wcm9maWxlX2NhbGxfZXhpdCcNCm1haW4ubzogSW4gZnVuY3Rp b24gYElubmVyUHJvZHVjdCc6DQptYWluLm8oLnRleHQrMHhhNTIpOiB1bmRl ZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2ZpbGVfaW5pdCcNCm1haW4ubygu dGV4dCsweGE5Mik6IHVuZGVmaW5lZCByZWZlcmVuY2UgdG8gYF9fcHJvZmls ZV9wdV9pbml0Jw0KbWFpbi5vKC50ZXh0KzB4YWIyKTogdW5kZWZpbmVkIHJl ZmVyZW5jZSB0byBgX19wcm9maWxlX2ludm9rZV9pbml0Jw0KbWFpbi5vKC50 ZXh0KzB4YWQyKTogdW5kZWZpbmVkIHJlZmVyZW5jZSB0byBgX19wcm9maWxl X2xvb3BfaW5pdCcNCm1haW4ubygudGV4dCsweGFmMik6IHVuZGVmaW5lZCBy ZWZlcmVuY2UgdG8gYF9fcHJvZmlsZV9pbnZva2UnDQptYWluLm8oLnRleHQr MHhiMzIpOiB1bmRlZmluZWQgcmVmZXJlbmNlIHRvIGBfX3Byb2ZpbGVfbG9v cCcNCm1haW4ubygudGV4dCsweGI4Mik6IHVuZGVmaW5lZCByZWZlcmVuY2Ug dG8gYF9fcHJvZmlsZV9sb29wX2l0ZXInDQpjb2xsZWN0MjogbGQgcmV0dXJu ZWQgMSBleGl0IHN0YXR1cw0Kc2dpY2MgRVJST1I6ICAvdXNyL2Jpbi9nY2Mg cmV0dXJuZWQgbm9uLXplcm8gc3RhdHVzIDENCi0tLS0tLS0tLS0tLS0tLS0t LS0tLQ0K ---2138968809-1880008973-985487207=:9039-- From owner-pro64-support@oss.sgi.com Sat Mar 24 21:51:46 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2P5pkG09818 for pro64-support-outgoing; Sat, 24 Mar 2001 21:51:46 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2P5pjM09815 for ; Sat, 24 Mar 2001 21:51:45 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Sat, 24 Mar 2001 22:51:32 -0700 Date: Sat, 24 Mar 2001 22:51:32 -0700 (MST) From: Peng Zhao To: Peng Zhao cc: sgi Subject: Re: error on -fb_create in Pro64 " In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk On Sat, 24 Mar 2001, Peng Zhao wrote: > Hi, > > Enclosed file is the output of the command "sgicc -show -fb_create > . -O0 main.c" The error is like this: > main.o(.text+0xa2): undefined reference to `__profile_pu_init' > > Actually, when I run the above command on another MIPS 10000 2.7 > machine (IRIX 6.5), I found that cc generate nothing. The above description is wrong, I have found the problem. I made a mistake when use "-fb_create .", I didnot include the name of the ammplication. Thus the profile information is stored in a file named with " ..instr.", so I cannot see it with simple "ls" command. > It seems the WN_Instrument can never be executed, > because I cannot find the place the varialbe "Instrumentation_Enabled" is > set to TRUE. > > Thank you very much. > Peng > -- > Peng Zhao pengzhao@cs.ualberta.ca > http://www.cs.ualberta.ca/~pengzhao > TEL (Lab): (780)492-3725 Lab: CSC251 > > Peng -- Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Mon Mar 26 07:49:37 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2QFnbD24723 for pro64-support-outgoing; Mon, 26 Mar 2001 07:49:37 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2QFnaM24720 for ; Mon, 26 Mar 2001 07:49:36 -0800 Received: from gaea.engr.sgi.com (gaea.engr.sgi.com [130.62.180.97]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id HAA03386 for ; Mon, 26 Mar 2001 07:59:40 -0800 (PST) mail_from (murthy@sgi.com) Received: from sgi.com (localhost [127.0.0.1]) by gaea.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id HAA85037; Mon, 26 Mar 2001 07:48:04 -0800 (PST) Message-ID: <3ABF64B3.EBB47BAC@sgi.com> Date: Mon, 26 Mar 2001 07:48:03 -0800 From: Chandrasekhar Murthy X-Mailer: Mozilla 4.51C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: Peng Zhao CC: sgi Subject: Re: a question about the whirl node instruction. References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Peng Zhao wrote: > > HI, > > I have some question about the meaning of the instructions of the > whirl IR, e.g. following is one line taken from the result of > "ir_b2a xxx.I", it simplements "i++" > I usually use the -st2 option on ir_b2a to also get the symbol table dump. For <2,2,i>, the first 2 gives the nesting level of the symbol table for the function, the second 2 gives the index of i within this symbol table. All types are stored in a single table. For T<4, .predef_I4, 4> The first 4 represents the idex for this type in the type table .predef_I4 represents the name (we use predefined type names for some frequently used types such as char, short, int) The last 4 specifies the alignment for using usage of the type. LOC gives information regarding the source line > ... > LOC 1 19 > I4I4LDID 0 <2,2,i> T<4, .predef_I4, 4> > I4INTCONST 1 (0x1) > I4ADD > I4STID 0 <2,2,i> T<4, .predef_I4, 4> > ... > > what is the meaning of <2,2,i>, T<4, .predef_I4, 4>, I4I4 , "LOC 1 > 19"? > Although I can guess something, it is still not very clear to me. > > Can somebody tell me the meaning of these fields explicitly and > the rule to diciper them? I cannot find the description of them in the > Whirl specificatoin. > > > Thank you very much. > > BTW: It seems to me that there is a little bug in the LOC(source > position information) related > things in ir_b2a. Sometimes, the LOC appears before the other codes of the > same whirl node, but sometimes it appears behind them. It is somewhat > confusing. > That is possible. In WHIRL, line number information is associated with WHIRL statement nodes and they might not match in the source language (for example, CALL or post increment). It might alo not be correct at higher optimization levels. It is also possible that the frontend is not generating them correctly. Please send a sample example with the commandline line used. > Peng > -- > Peng Zhao pengzhao@cs.ualberta.ca > http://www.cs.ualberta.ca/~pengzhao > TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Mon Mar 26 09:55:31 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2QHtVV29185 for pro64-support-outgoing; Mon, 26 Mar 2001 09:55:31 -0800 Received: from yog-sothoth.sgi.com (eugate.sgi.com [192.48.160.10]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2QHtSM29181 for ; Mon, 26 Mar 2001 09:55:28 -0800 Received: from gaea.engr.sgi.com (gaea.engr.sgi.com [130.62.180.97]) by yog-sothoth.sgi.com (980305.SGI.8.8.8-aspam-6.2/980304.SGI-aspam-europe) via ESMTP id TAA6405439 for ; Mon, 26 Mar 2001 19:55:25 +0200 (CEST) mail_from (murthy@sgi.com) Received: from sgi.com (localhost [127.0.0.1]) by gaea.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id JAA81073; Mon, 26 Mar 2001 09:53:53 -0800 (PST) Message-ID: <3ABF8230.2FDC5931@sgi.com> Date: Mon, 26 Mar 2001 09:53:52 -0800 From: Chandrasekhar Murthy X-Mailer: Mozilla 4.51C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: Peng Zhao , sgi Subject: Re: a question about the whirl node instruction. References: <3ABF64B3.EBB47BAC@sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk > The last 4 specifies the alignment for using usage of the type. I meant The last 4 specifies the alignment for this particular usage of the type. Murthy From owner-pro64-support@oss.sgi.com Mon Mar 26 14:39:43 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2QMdhQ07301 for pro64-support-outgoing; Mon, 26 Mar 2001 14:39:43 -0800 Received: from mta6.snfc21.pbi.net (mta6.snfc21.pbi.net [206.13.28.240]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2QMdfM07298 for ; Mon, 26 Mar 2001 14:39:41 -0800 Received: from ted ([63.206.179.222]) by mta6.snfc21.pbi.net (Sun Internet Mail Server sims.3.5.2000.01.05.12.18.p9) with SMTP id <0GAT003Q4SXV8S@mta6.snfc21.pbi.net> for pro64-support@oss.sgi.com; Mon, 26 Mar 2001 14:25:08 -0800 (PST) Date: Mon, 26 Mar 2001 14:37:27 -0800 From: Wayne Vieira Subject: Using OpenMP on IA64 In-reply-to: <3ABF64B3.EBB47BAC@sgi.com> To: sgi Cc: chan1@llnl.gov Message-id: MIME-version: 1.0 X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) Content-type: text/plain; charset="us-ascii" Content-transfer-encoding: 7bit Importance: Normal X-MSMail-Priority: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 X-Priority: 3 (Normal) Sender: owner-pro64-support@oss.sgi.com Precedence: bulk We have an application that used OpenMP and has been crashing on IA64. The code looks something like this: printf(" befor OMP \n"); fflush(); #pragma omp parallel for ....etc for (....) { printf("inside OMP \n"); fflush(); } The code will crash AFTER it print "before OMP" and never goes to the "inside OMP". My experience for other platform is that it usually indicate the stack size is too small because it does not even do any work at all and the OMP lib crash. The MP_SLAVE_STACKSIZE is an env variable. All the other unix platforms that we know of use env variable to set the stack size for THREAD. We've tried to set MP_SLAVE_STACKSIZE and it does NOT help. We also tried to lower my opt level from -O3 to nothing(i.e. default) and it does not help either. Is there something obvious to using OMP on IA64 that we're missing? Thanks, Wayne From owner-pro64-support@oss.sgi.com Tue Mar 27 06:37:10 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2REbAQ30347 for pro64-support-outgoing; Tue, 27 Mar 2001 06:37:10 -0800 Received: from mail.eecis.udel.edu (louie.udel.edu [128.175.2.33]) by oss.sgi.com (8.11.3/8.11.3) with SMTP id f2REbAM30341 for ; Tue, 27 Mar 2001 06:37:10 -0800 Received: from louie.udel.edu by mail.eecis.udel.edu id aa04972; 27 Mar 2001 09:36 EST Date: Tue, 27 Mar 2001 09:36:39 -0500 (EST) From: Ziang Hu To: pro64-support@oss.sgi.com MMDF-Warning: Parse error in original version of preceding line at mail.eecis.udel.edu Subject: can Pro64 work with gprof or other profiling tools? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Thanks Ziang Hu From owner-pro64-support@oss.sgi.com Tue Mar 27 07:24:20 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2RFOKF32463 for pro64-support-outgoing; Tue, 27 Mar 2001 07:24:20 -0800 Received: from ii.uib.no (eik.ii.uib.no [129.177.16.3]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2RFOJM32460 for ; Tue, 27 Mar 2001 07:24:19 -0800 Received: from apal-192.ii.uib.no (apal.ii.uib.no) [129.177.192.27] by ii.uib.no with esmtp (Exim 3.03) id 14hvKh-0000qe-00 for ; Tue, 27 Mar 2001 17:24:31 +0200 Received: (from janfrode@localhost) by apal.ii.uib.no (8.9.3+Sun/8.9.3) id RAA13143 for pro64-support@oss.sgi.com; Tue, 27 Mar 2001 17:24:18 +0200 (MEST) Date: Tue, 27 Mar 2001 17:24:16 +0200 From: Jan-Frode Myklebust To: pro64-support@oss.sgi.com Subject: A bug? Message-ID: <20010327172415.A12770@ii.uib.no> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Sender: owner-pro64-support@oss.sgi.com Precedence: bulk I'm running the pro64 compilers under NUE, and am trying to build one of our benchmarks. Unfortunately I believe I'm hitting a bug in f90. Can someone confirm if this message is a compiler crash? bash$ sgif90 -fullwarn -c fieldinit2.f90 Signal: Illegal instruction in Front End Parse/Semantic phase. Error: Signal Illegal instruction in phase Front End Parse/Semantic -- processing aborted bash$ sgif90 -version SGIcc Compilers: Version 0.01.0-13 Here's a short strace of it: [snip] stat("fieldinit2.f90", {st_mode=S_IFREG|0600, st_size=6481, ...}) = 0 stat("/usr/lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/be.so", {st_mode=S_IFREG|0755, st_size=5538433, ...}) = 0 stat("/usr/lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/be", {st_mode=S_IFREG|0755, st_size=148432, ...}) = 0 stat("/tmp", {st_mode=S_IFDIR|S_ISVTX|0777, st_size=4096, ...}) = 0 gettimeofday({985706439, 814035}, NULL) = 0 getpid() = 2390 stat("/tmp/ccL.6iWlb2", 0xbfffdec4) = -1 ENOENT (No such file or directory) open("/tmp/ccL.6iWlb2", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=0, ...}) = 0 mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40014000 brk(0x80a4000) = 0x80a4000 getcwd("/home/janfrode/bomb/sourcecode", 4095) = 31 write(3, "sgif90 -fullwarn -c fieldinit2.f"..., 66) = 66 close(3) = 0 munmap(0x40014000, 4096) = 0 stat("/tmp", {st_mode=S_IFDIR|S_ISVTX|0777, st_size=4096, ...}) = 0 gettimeofday({985706439, 818440}, NULL) = 0 getpid() = 2390 stat("/tmp/ccB.ps8PKX", 0xbfffdc80) = -1 ENOENT (No such file or directory) fork() = 2391 wait4(-1, Signal: Illegal instructionError: Signal Illegal instruction in phase Front End Parse/Semantic -- processing aborted in Front End Parse/Semantic phase. [WIFEXITED(s) && WEXITSTATUS(s) == 4], 0, NULL) = 2391 --- SIGCHLD (Barnet avsluttet) --- unlink("/tmp/ccL.6iWlb2") = 0 unlink("/tmp/ccB.ps8PKX") = 0 _exit(2) = ? This builds cleanly on IRIX. Are there any lists of bugs or missing features somewhere? Or maybe an archive of this mailinglist? -jf From owner-pro64-support@oss.sgi.com Tue Mar 27 08:59:28 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2RGxS102132 for pro64-support-outgoing; Tue, 27 Mar 2001 08:59:28 -0800 Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2RGxSM02129 for ; Tue, 27 Mar 2001 08:59:28 -0800 Received: from SMTP (orsmsxvs02-1.jf.intel.com [192.168.65.201]) by ganymede.or.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id QAA29335; Tue, 27 Mar 2001 16:59:10 GMT Received: from orsmsx29.jf.intel.com ([192.168.70.29]) by 192.168.70.201 (Norton AntiVirus for Internet Email Gateways 1.0) ; Tue, 27 Mar 2001 16:59:09 0000 (GMT) Received: by orsmsx29.jf.intel.com with Internet Mail Service (5.5.2653.19) id ; Tue, 27 Mar 2001 08:59:08 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9A5@FMSMSX41> From: "Chan, Sun C" To: "'Ziang Hu'" , pro64-support@oss.sgi.com Subject: RE: can Pro64 work with gprof or other profiling tools? Date: Tue, 27 Mar 2001 08:59:05 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk gprof on IA64 is useless. I heard that .13 release includes the instrumentation library. If you back port the library to .12 the .12 compiler does fully support profiling where instrumentation is done right after the frontend. Sun > -----Original Message----- > From: Ziang Hu [mailto:hu@mail.eecis.udel.edu] > Sent: Tuesday, March 27, 2001 6:37 AM > To: pro64-support@oss.sgi.com > Subject: can Pro64 work with gprof or other profiling tools? > > > > Thanks > > Ziang Hu > > From owner-pro64-support@oss.sgi.com Tue Mar 27 11:37:04 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2RJb4c07913 for pro64-support-outgoing; Tue, 27 Mar 2001 11:37:04 -0800 Received: from ii.uib.no (eik.ii.uib.no [129.177.16.3]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2RJb3M07910 for ; Tue, 27 Mar 2001 11:37:03 -0800 Received: from apal-192.ii.uib.no (apal.ii.uib.no) [129.177.192.27] by ii.uib.no with esmtp (Exim 3.03) id 14hzHH-0005q8-00 ; Tue, 27 Mar 2001 21:37:15 +0200 Received: (from janfrode@localhost) by apal.ii.uib.no (8.9.3+Sun/8.9.3) id VAA19467; Tue, 27 Mar 2001 21:37:01 +0200 (MEST) Date: Tue, 27 Mar 2001 21:37:01 +0200 From: Jan-Frode Myklebust To: Michael Murphy Cc: pro64-support@oss.sgi.com, Helge Avlesen Subject: Re: A bug? Message-ID: <20010327213700.A19207@ii.uib.no> References: <200103271755.JAA82663@rohi.engr.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200103271755.JAA82663@rohi.engr.sgi.com>; from mpm@rohi.engr.sgi.com on Tue, Mar 27, 2001 at 09:55:28AM -0800 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk On Tue, Mar 27, 2001 at 09:55:28AM -0800, Michael Murphy wrote: > When it dies with something like: > Signal: Illegal instruction in Front End Parse/Semantic phase. > That generally means a compiler error. > Please post a test case and the command line used, > and someone will look into it. #!/bin/sh cat >bug.f90 << EOF SUBROUTINE FIELDINIT REAL DELTAR DELTAR = SQRT(3.) RETURN END SUBROUTINE FIELDINIT EOF sgif90 -c bug.f90 It doesn't seem to like the '.'. -jf From owner-pro64-support@oss.sgi.com Tue Mar 27 11:43:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2RJhRE08066 for pro64-support-outgoing; Tue, 27 Mar 2001 11:43:27 -0800 Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2RJhRM08061 for ; Tue, 27 Mar 2001 11:43:27 -0800 Received: from cchkms.engr.sgi.com (cchkms.engr.sgi.com [130.62.180.48]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id LAA21755 for ; Tue, 27 Mar 2001 11:42:13 -0800 (PST) mail_from (rat@cchkms.engr.sgi.com) Received: (from rat@localhost) by cchkms.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id LAA34957; Tue, 27 Mar 2001 11:42:06 -0800 (PST) From: "Ross A. Towle" Message-Id: <10103271142.ZM35513@cchkms.engr.sgi.com> Date: Tue, 27 Mar 2001 11:42:06 -0800 In-Reply-To: Jan-Frode Myklebust "A bug?" (Mar 27, 5:24pm) References: <20010327172415.A12770@ii.uib.no> X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: Jan-Frode Myklebust , pro64-support@oss.sgi.com Subject: Re: A bug? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk As we state in our webpages, Pro64 release 0.13 will have problems working in NUE until there are upgrades made by HP. Note you are failing in fstat. I will guess that you are running RedHat 7.0 or later. The compiler itself is built assuming glibc2.1(for IA32). The IA-64 runtime libraries are built assuming glibc2.2 (for IA64). Solution: go back to RedHat6.2 -Ross From owner-pro64-support@oss.sgi.com Thu Mar 29 14:17:34 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2TMHYr12559 for pro64-support-outgoing; Thu, 29 Mar 2001 14:17:34 -0800 Received: from hypnos.cps.intel.com (hypnos.cps.intel.com [192.198.165.17]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2TMHTM12548 for ; Thu, 29 Mar 2001 14:17:34 -0800 Received: from SMTP (fmsmsxvs03-1.fm.intel.com [132.233.42.203]) by hypnos.cps.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id WAA13509 for ; Thu, 29 Mar 2001 22:17:05 GMT Received: from fmsmsx19.fm.intel.com ([132.233.48.19]) by 132.233.48.203 (Norton AntiVirus for Internet Email Gateways 1.0) ; Thu, 29 Mar 2001 22:16:54 0000 (GMT) Received: by fmsmsx19.fm.intel.com with Internet Mail Service (5.5.2653.19) id ; Thu, 29 Mar 2001 14:16:53 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9BB@FMSMSX41> From: "Chan, Sun C" To: "'pro64-support@oss.sgi.com'" Cc: "Chan, Sun C" , "Shankar, Bhanu" Subject: sbss overflow, linker problem? Date: Thu, 29 Mar 2001 14:16:51 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk All, for the following two files: ! cat bm2.f SUBROUTINE SETFM(IPAR) IMPLICIT DOUBLE PRECISION(A-H,O-Z) ! IMPLICIT INTEGER(A-H,O-Z) C PARAMETER (MEMSIZ= 1750 000) ! PARAMETER (MEMSIZ= 3) COMMON /FMCOM / X(MEMSIZ) END !cat bm1.f PROGRAM bm ! IMPLICIT DOUBLE PRECISION(A-H,O-Z) IMPLICIT INTEGER(A-H,O-Z) COMMON /FMCOM / X(1) END on the nue native environment, for the a.out produced, readelf shows: [25] .sbss NOBITS 6000000000015a00 000f5a00 0000000000d5a040 0000000000000000 WA 0 0 8 clearly the sbss overflowed. But the linker DID NOT report error. The version of linker is 2.9 (000216) With a newer linker: version 2.9 (000717) I got: ld: a.out: short data segment overflowed (0x5bafb8 >= 0x400000) depending on the link order, i.e the order "bm1.f bm2.f" give the above error. the order "bm2.f bm1.f" does not. the .s file generated has the following declaration for fmcom: .common fmcom_#, 14000000, 8 I believe Pro64 is not the culprit here. SGI folks, is this a known problem? Also, appreciate any workaround or hints as to when Cygnus is fixing this? Or if you already have a local fix, care to share that? Thx, Sun From owner-pro64-support@oss.sgi.com Thu Mar 29 14:24:30 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2TMOUW12808 for pro64-support-outgoing; Thu, 29 Mar 2001 14:24:30 -0800 Received: from scapa.cs.ualberta.ca (scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2TMOUM12805 for ; Thu, 29 Mar 2001 14:24:30 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Thu, 29 Mar 2001 15:24:10 -0700 Date: Thu, 29 Mar 2001 15:24:11 -0700 (MST) From: Peng Zhao To: sgi cc: Jose Nelson Amaral Subject: debug the inliner. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hi, In pro64 0.12. I have a question of how to debug the "inline" program alone. I set teh LD_LIBRARY_PATH, LD_LIBRARYN32_PATH variables. Following is my process to use gdb: gdb ../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/inline > set args -LIST:source=off:notes=off -PHASE:c -O0 -TENV:PIC -m1 -INTERNAL:return_val=on -INTERNAL:mldid_mstid=on -INTERNAL:return_info=on -INLINE:must=callit:list -TARG:abi=i64 -fB,../ccB.k5XNKH.pengzhao -fI,../ccI.nxMoS6.pengzhao test.c >(gdb) l 359 364 main (INT argc, char **argv) 365 { 366 INT32 ecount, local_ecount, local_wcount; 367 368 ecount = 0; >(gdb) b 368 Breakpoint 1 at 0x807b62e: file ../../ipa/inline/inline_driver.cxx, line 368. >r Starting program: /nue/usr/sakwatamau/brule12/grad/pengzhao/temp/../lib/gcc-lib/ia64-sgi-linux/sgicc-1.0/inline -LIST:source=off:notes=off -PHASE:c -O0 -TENV:PIC -m1 -INTERNAL:return_val=on -INTERNAL:mldid_mstid=on -INTERNAL:return_info=on -INLINE:must=callit:list -TARG:abi=i64 -fB,../ccB.k5XNKH.pengzhao -fI,../ccI.nxMoS6.pengzhao test.c Then the gdb will halt there still. Nothing happens. I noticed that in 0.13 document, there is some texts decribing how to debug BE DSO's (using be_debug). Is this problem also dso-related? thank you. When I use gdb to trace from gcc to inlining, the problem also appears. That is, I cannot trace into the inliner. I forgot what is the error message. Can sb else explain sth to me? Thank you. From owner-pro64-support@oss.sgi.com Thu Mar 29 14:53:36 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2TMraZ13492 for pro64-support-outgoing; Thu, 29 Mar 2001 14:53:36 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2TMraM13489 for ; Thu, 29 Mar 2001 14:53:36 -0800 Received: from rohi.engr.sgi.com (rohi.engr.sgi.com [130.62.180.74]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id PAA02097 for ; Thu, 29 Mar 2001 15:03:43 -0800 (PST) mail_from (mpm@rohi.engr.sgi.com) Received: (from mpm@localhost) by rohi.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id OAA92457; Thu, 29 Mar 2001 14:52:01 -0800 (PST) Date: Thu, 29 Mar 2001 14:52:01 -0800 (PST) From: mpm@rohi.engr.sgi.com (Michael Murphy) Message-Id: <200103292252.OAA92457@rohi.engr.sgi.com> To: sgi , Peng Zhao Subject: Re: debug the inliner. Cc: Jose Nelson Amaral Sender: owner-pro64-support@oss.sgi.com Precedence: bulk I notice that you are referring to tmp files. It may be that your tmp files have disappeared. Try compiling with -keep and then running the debugger. -- Mike Murphy -- mpm@sgi.com -- quote of the day: -- "Compassion is sometimes the fatal capacity for feeling what it is like -- inside somebody else's skin. It is the knowledge that there can never -- really be any peace and joy for me until there is peace and joy finally -- for you too." (Frederick Buechner) From owner-pro64-support@oss.sgi.com Thu Mar 29 15:56:59 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2TNuxG14448 for pro64-support-outgoing; Thu, 29 Mar 2001 15:56:59 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2TNuwM14445 for ; Thu, 29 Mar 2001 15:56:58 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Thu, 29 Mar 2001 16:56:49 -0700 Date: Thu, 29 Mar 2001 16:56:50 -0700 (MST) From: Peng Zhao To: Michael Murphy cc: sgi , Jose Nelson Amaral Subject: Re: debug the inliner. In-Reply-To: <200103292252.OAA92457@rohi.engr.sgi.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk The file is there. I saved them intentionally. On Thu, 29 Mar 2001, Michael Murphy wrote: > I notice that you are referring to tmp files. It may be that your > tmp files have disappeared. Try compiling with -keep and then > running the debugger. > -- Mike Murphy > -- mpm@sgi.com > -- quote of the day: > -- "Compassion is sometimes the fatal capacity for feeling what it is like > -- inside somebody else's skin. It is the knowledge that there can never > -- really be any peace and joy for me until there is peace and joy finally > -- for you too." (Frederick Buechner) > -- Regards Peng Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Thu Mar 29 16:08:42 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2U08go14751 for pro64-support-outgoing; Thu, 29 Mar 2001 16:08:42 -0800 Received: from thalia.fm.intel.com (fmfdns02.fm.intel.com [132.233.247.11]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2U08aM14745 for ; Thu, 29 Mar 2001 16:08:41 -0800 Received: from SMTP (fmsmsxvs01-1.fm.intel.com [132.233.42.201]) by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id AAA04327; Fri, 30 Mar 2001 00:07:46 GMT Received: from fmsmsx17.intel.com ([132.233.48.17]) by 132.233.48.201 (Norton AntiVirus for Internet Email Gateways 1.0) ; Fri, 30 Mar 2001 00:07:45 0000 (GMT) Received: by fmsmsx17.fm.intel.com with Internet Mail Service (5.5.2653.19) id ; Thu, 29 Mar 2001 16:07:43 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9BF@FMSMSX41> From: "Chan, Sun C" To: "'Peng Zhao'" , Michael Murphy Cc: sgi , Jose Nelson Amaral Subject: RE: debug the inliner. Date: Thu, 29 Mar 2001 16:07:41 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="ISO-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Each compile creates a new tmp file name. You need to do -keep first, and then, at debug time, do -keep again. This way, the intermediate files uses the same name (not tmp). Sun > -----Original Message----- > From: Peng Zhao [mailto:pengzhao@cs.ualberta.ca] > Sent: Thursday, March 29, 2001 3:57 PM > To: Michael Murphy > Cc: sgi; Jose Nelson Amaral > Subject: Re: debug the inliner. > > > > The file is there. I saved them intentionally. > > > On Thu, 29 Mar 2001, Michael Murphy wrote: > > > I notice that you are referring to tmp files. It may be that your > > tmp files have disappeared. Try compiling with -keep and then > > running the debugger. > > -- Mike Murphy > > -- mpm@sgi.com > > -- quote of the day: > > -- "Compassion is sometimes the fatal capacity for feeling > what it is like > > -- inside somebody else's skin. It is the knowledge that > there can never > > -- really be any peace and joy for me until there is > peace and joy finally > > -- for you too." (Frederick Buechner) > > > > -- > Regards > > Peng > Peng Zhao pengzhao@cs.ualberta.ca > http://www.cs.ualberta.ca/~pengzhao > TEL (Lab): (780)492-3725 Lab: CSC251 > > > From owner-pro64-support@oss.sgi.com Thu Mar 29 16:22:39 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2U0MdJ15054 for pro64-support-outgoing; Thu, 29 Mar 2001 16:22:39 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2U0MXM15051 for ; Thu, 29 Mar 2001 16:22:33 -0800 Received: from rohi.engr.sgi.com (rohi.engr.sgi.com [130.62.180.74]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id QAA06094 for ; Thu, 29 Mar 2001 16:32:41 -0800 (PST) mail_from (mpm@rohi.engr.sgi.com) Received: (from mpm@localhost) by rohi.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id QAA92646; Thu, 29 Mar 2001 16:21:10 -0800 (PST) Date: Thu, 29 Mar 2001 16:21:10 -0800 (PST) From: mpm@rohi.engr.sgi.com (Michael Murphy) Message-Id: <200103300021.QAA92646@rohi.engr.sgi.com> To: "'pro64-support@oss.sgi.com'" , "Chan, Sun C" Subject: Re: sbss overflow, linker problem? Cc: "Shankar, Bhanu" , "Chan, Sun C" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk From: "Chan, Sun C" for the following two files: ! cat bm2.f SUBROUTINE SETFM(IPAR) IMPLICIT DOUBLE PRECISION(A-H,O-Z) ! IMPLICIT INTEGER(A-H,O-Z) C PARAMETER (MEMSIZ= 1750 000) ! PARAMETER (MEMSIZ= 3) COMMON /FMCOM / X(MEMSIZ) END !cat bm1.f PROGRAM bm ! IMPLICIT DOUBLE PRECISION(A-H,O-Z) IMPLICIT INTEGER(A-H,O-Z) COMMON /FMCOM / X(1) END on the nue native environment, for the a.out produced, readelf shows: [25] .sbss NOBITS 6000000000015a00 000f5a00 0000000000d5a040 0000000000000000 WA 0 0 8 clearly the sbss overflowed. But the linker DID NOT report error. The version of linker is 2.9 (000216) With a newer linker: version 2.9 (000717) I got: ld: a.out: short data segment overflowed (0x5bafb8 >= 0x400000) depending on the link order, i.e the order "bm1.f bm2.f" give the above error. the order "bm2.f bm1.f" does not. the .s file generated has the following declaration for fmcom: .common fmcom_#, 14000000, 8 I believe Pro64 is not the culprit here. This definitely looks like a linker bug, as it is the linker that is putting the resulting common in .sbss here. I notice that it doesn't use sbss if both files have the large common size. SGI folks, is this a known problem? Not to me, though it may explain some similar weird problems we've seen with gp overflow. Also, appreciate any workaround or hints as to when Cygnus is fixing this? I suggest you contact cygnus/gcc directly, as they may be most attentive to problems from Intel. -- Mike Murphy -- mpm@sgi.com -- quote of the day: -- "Compassion is sometimes the fatal capacity for feeling what it is like -- inside somebody else's skin. It is the knowledge that there can never -- really be any peace and joy for me until there is peace and joy finally -- for you too." (Frederick Buechner) From owner-pro64-support@oss.sgi.com Thu Mar 29 17:10:55 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2U1Atd16079 for pro64-support-outgoing; Thu, 29 Mar 2001 17:10:55 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2U1AnM16075 for ; Thu, 29 Mar 2001 17:10:49 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Thu, 29 Mar 2001 18:10:35 -0700 Date: Thu, 29 Mar 2001 18:10:35 -0700 (MST) From: Peng Zhao To: "Chan, Sun C" cc: Michael Murphy , sgi Subject: RE: debug the inliner. In-Reply-To: <9287DC1579B0D411AA2F009027F44C3F042DF9BF@FMSMSX41> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk I think the problem is not the temp file. I save the tmp file and use -fB to tell the inline to find the file. Actually it runs, after more than 5-10 minutes, I found gdb finally came to the breakpoint. But so slow unexpectedly. later "next" will go smoothly, but ocassionaly, "step" still consume much time when the source file switch. On Thu, 29 Mar 2001, Chan, Sun C wrote: > Each compile creates a new tmp file name. You need to do -keep > first, and then, at debug time, do -keep again. This way, the > intermediate files uses the same name (not tmp). > Sun > > > -----Original Message----- > > From: Peng Zhao [mailto:pengzhao@cs.ualberta.ca] > > Sent: Thursday, March 29, 2001 3:57 PM > > To: Michael Murphy > > Cc: sgi; Jose Nelson Amaral > > Subject: Re: debug the inliner. > > > > > > > > The file is there. I saved them intentionally. > > > > > > On Thu, 29 Mar 2001, Michael Murphy wrote: > > > > > I notice that you are referring to tmp files. It may be that your > > > tmp files have disappeared. Try compiling with -keep and then > > > running the debugger. > > > -- Mike Murphy > > > -- mpm@sgi.com > > > -- quote of the day: > > > -- "Compassion is sometimes the fatal capacity for feeling > > what it is like > > > -- inside somebody else's skin. It is the knowledge that > > there can never > > > -- really be any peace and joy for me until there is > > peace and joy finally > > > -- for you too." (Frederick Buechner) > > > > > > > -- > > Regards > > > > Peng > > Peng Zhao pengzhao@cs.ualberta.ca > > http://www.cs.ualberta.ca/~pengzhao > > TEL (Lab): (780)492-3725 Lab: CSC251 > > > > > > > -- Regards Peng Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251 From owner-pro64-support@oss.sgi.com Fri Mar 30 00:30:00 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2U8U0525152 for pro64-support-outgoing; Fri, 30 Mar 2001 00:30:00 -0800 Received: from roura.ac.upc.es (roura.ac.upc.es [147.83.33.10]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2U8TwM25149 for ; Fri, 30 Mar 2001 00:29:59 -0800 Received: from ac.upc.es (florensa.ac.upc.es [147.83.33.3]) by roura.ac.upc.es (8.11.0/8.11.0) with ESMTP id f2U8Tmb24258 for ; Fri, 30 Mar 2001 10:29:49 +0200 (MET DST) Message-ID: <3AC443FC.8BE2CF51@ac.upc.es> Date: Fri, 30 Mar 2001 10:29:48 +0200 From: Eduard Santamaria Organization: DAC-UPC X-Mailer: Mozilla 4.76 [en] (X11; U; SunOS 5.7 sun4m) X-Accept-Language: en MIME-Version: 1.0 To: pro64 support Subject: problem with gfec Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hi, I'm running the Pro64 on an IRIX/MIPS machine and have a problem with the c front-end. gfec translates this simple program int main( void ) { unsigned int a; a = 1; return a; } to the following _INT32 main() { _UINT32 a; a = 0U; return * (_INT32 *)(&a); } /* main */ Note tha a 0 and not a 1 is being assigned to a. The .B file looks like this: LOC 1 3 unsigned int a; LOC 1 4 LOC 1 5 a = 1; U4INTCONST 0 (0x0) U4STID 0 <2,1,a> T<8,.predef_U4,4> If I declare a as a signed int or as a long (signed or unsigned, it doesn't matter) it works fine. My binaries of the Pro64 are N32 mips-4. Does anyone out there have any idea of what might be happenning? I've been doing some debugging and I think that perhaps the problem is due to an ill performed conversion of the constant to the type of the variable, but I'm not sure. Thank you, Eduard _____________________________________________________________________ o o o Eduard Santamaria Barnadas o o o Department of Computer Architecture o o o Universitat Politecnica de Catalunya Phone: +34 934 011 649 C/ Jordi Girona, 1-3 Campus Nord, Modul C6 - S103 U P C 08034 - BARCELONA (SPAIN) mailto:esantama@ac.upc.es _____________________________________________________________________ From owner-pro64-support@oss.sgi.com Fri Mar 30 02:56:09 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UAu9x30319 for pro64-support-outgoing; Fri, 30 Mar 2001 02:56:09 -0800 Received: from roura.ac.upc.es (roura.ac.upc.es [147.83.33.10]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UAu7M30316 for ; Fri, 30 Mar 2001 02:56:07 -0800 Received: from ac.upc.es (florensa.ac.upc.es [147.83.33.3]) by roura.ac.upc.es (8.11.0/8.11.0) with ESMTP id f2UAu0b28634 for ; Fri, 30 Mar 2001 12:56:00 +0200 (MET DST) Message-ID: <3AC4663F.6A44356A@ac.upc.es> Date: Fri, 30 Mar 2001 12:55:59 +0200 From: Eduard Santamaria Organization: DAC-UPC X-Mailer: Mozilla 4.76 [en] (X11; U; SunOS 5.7 sun4m) X-Accept-Language: en MIME-Version: 1.0 To: pro64 support Subject: problem with gfec Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Don't mind about my previous message, I got it. It was related to include files and type definitions. Sorry and thanks anyway, Eduard. _____________________________________________________________________ o o o Eduard Santamaria Barnadas o o o Department of Computer Architecture o o o Universitat Politecnica de Catalunya Phone: +34 934 011 649 C/ Jordi Girona, 1-3 Campus Nord, Modul C6 - S103 U P C 08034 - BARCELONA (SPAIN) mailto:esantama@ac.upc.es _____________________________________________________________________ From owner-pro64-support@oss.sgi.com Fri Mar 30 06:18:01 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UEI1n01454 for pro64-support-outgoing; Fri, 30 Mar 2001 06:18:01 -0800 Received: from energy.pdb.sbs.de (energy.pdb.sbs.de [192.109.2.19]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UEHvM01451 for ; Fri, 30 Mar 2001 06:17:58 -0800 Received: from trulli.pdb.fsc.net ([172.25.96.20]) by energy.pdb.sbs.de (8.9.3/8.9.3) with ESMTP id QAA21480 for ; Fri, 30 Mar 2001 16:17:48 +0200 Received: from abg0972e.abg.fsc.net (abg0972e.abg.fsc.net [141.29.77.162]) by trulli.pdb.fsc.net (8.9.3/8.9.3) with ESMTP id QAA06741 for ; Fri, 30 Mar 2001 16:17:48 +0200 Received: by abg0972e.abg.fsc.net with Internet Mail Service (5.5.2653.19) id ; Fri, 30 Mar 2001 16:17:50 +0200 Message-ID: <09BE2D952F35D411ABAF009027B6B1D30171EB01@abg0971e.abg.fsc.net> From: "Kratzer, Willi" To: "'pro64-support@oss.sgi.com'" Subject: auto-pic or non-auto-pic Date: Fri, 30 Mar 2001 16:17:49 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id f2UEHxM01452 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hi, when I try to link objects from the Pro64 compilers with Intel`s MKL (Mathematical Kernel Library) on an Itanium system, I get the messages /usr/lib/gcc-lib/ia64-redhat-linux/2.96-ia64-000717/../../../../ ia64-redhat-linux/bin/ld: smpsrv64.o: linking auto-pic files with non-auto-pic files Bad value: failed to merge target specific data of file /usr/local/intel/mkl/LIB/libmkl64_itp.a(smpsrv64.o) collect2: ld returned 1 exit status First question: What do the Pro64 compilers generate? auto-pic files or non-auto-pic files? Second question: Can I control it? Mit freundlichem Gruß / Best Regards Willi Kratzer --- Willi Kratzer, Fujitsu Siemens Computers PO PC Workstations Buergermeister-Ulrich-Str. 100 D-86199 Augsburg, Germany Tel: (..821)-804-4810 Fax: (..821)-804-2568 Email: willi.kratzer@fujitsu-siemens.com From owner-pro64-support@oss.sgi.com Fri Mar 30 08:58:25 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UGwPe04446 for pro64-support-outgoing; Fri, 30 Mar 2001 08:58:25 -0800 Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UGwOM04443 for ; Fri, 30 Mar 2001 08:58:24 -0800 Received: from SMTP (orsmsxvs02-1.jf.intel.com [192.168.65.201]) by ganymede.or.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id QAA10554; Fri, 30 Mar 2001 16:58:21 GMT Received: from orsmsx29.jf.intel.com ([192.168.70.29]) by 192.168.70.201 (Norton AntiVirus for Internet Email Gateways 1.0) ; Fri, 30 Mar 2001 16:58:21 0000 (GMT) Received: by orsmsx29.jf.intel.com with Internet Mail Service (5.5.2653.19) id ; Fri, 30 Mar 2001 08:58:18 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9C4@FMSMSX41> From: "Chan, Sun C" To: "'Kratzer, Willi'" , "'pro64-support@oss.sgi.com'" Subject: RE: auto-pic or non-auto-pic Date: Fri, 30 Mar 2001 08:58:07 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id f2UGwOM04444 Sender: owner-pro64-support@oss.sgi.com Precedence: bulk do "file smpsrv64.o" to see what type of object it is. I believe Intel defaults to non-shared (non-pic) while Pro64 defaults to shared. If you have source to libmkl64_itp, rebuild that with Pro64 compiler and put it in the same directory with the .a. It should fix your problem. Another way is to try adding -non_shared flag. The potential problem there is that there might not be a matching pair of non_shared fortran libraries. Sun > -----Original Message----- > From: Kratzer, Willi [mailto:Willi.Kratzer@fujitsu-siemens.com] > Sent: Friday, March 30, 2001 6:18 AM > To: 'pro64-support@oss.sgi.com' > Subject: auto-pic or non-auto-pic > > > Hi, > > when I try to link objects from the Pro64 compilers > with Intel`s MKL (Mathematical Kernel Library) on an > Itanium system, I get the messages > > /usr/lib/gcc-lib/ia64-redhat-linux/2.96-ia64-000717/../../../../ > ia64-redhat-linux/bin/ld: smpsrv64.o: linking auto-pic files with > non-auto-pic files > Bad value: failed to merge target specific data of file > /usr/local/intel/mkl/LIB/libmkl64_itp.a(smpsrv64.o) > collect2: ld returned 1 exit status > > First question: What do the Pro64 compilers generate? > auto-pic files or non-auto-pic files? > Second question: Can I control it? > > Mit freundlichem Gruß / Best Regards > Willi Kratzer > > --- > Willi Kratzer, Fujitsu Siemens Computers > PO PC Workstations > Buergermeister-Ulrich-Str. 100 > D-86199 Augsburg, Germany > Tel: (..821)-804-4810 > Fax: (..821)-804-2568 > Email: willi.kratzer@fujitsu-siemens.com > > From owner-pro64-support@oss.sgi.com Fri Mar 30 09:44:03 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UHi3705587 for pro64-support-outgoing; Fri, 30 Mar 2001 09:44:03 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UHi3M05584 for ; Fri, 30 Mar 2001 09:44:03 -0800 Received: (from localhost user: 'cam' uid#302 fake: STDIN (cam@sunset.cs.ualberta.ca)) by scapa.cs.ualberta.ca id ; Fri, 30 Mar 2001 10:43:45 -0700 Date: Fri, 30 Mar 2001 10:43:45 -0700 (MST) From: "A. Cameron Macdonell" To: pro64-support@oss.sgi.com Subject: creating a feedback mechanism Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Hello, I am currently trying to generate some profiling feedback information with the Pro64. I am trying to collect very limited profiling information about pointer accesses in the code and so I want to implement my own feedback mechanism seeing as there does not seem to be a current foundation from which to work within the Pro64 (please correct me if I am wrong about this). Does anyone with more experience with WHIRL know how feasible it is to add data structures to collect such data into the WHIRL representation and whether it should be done at the procedural level or at the PU level? Also, at which WHIRL level should I try to insert code to modify the added data structures? Pointers to answers are as good as direct answers. Thanks, Cam From owner-pro64-support@oss.sgi.com Fri Mar 30 09:51:16 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UHpGu05732 for pro64-support-outgoing; Fri, 30 Mar 2001 09:51:16 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UHpFM05729 for ; Fri, 30 Mar 2001 09:51:15 -0800 Received: from johmar.engr.sgi.com (johmar.engr.sgi.com [130.62.180.56]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id KAA05812 for ; Fri, 30 Mar 2001 10:01:23 -0800 (PST) mail_from (jcarter@johmar.engr.sgi.com) Received: (from jcarter@localhost) by johmar.engr.sgi.com (SGI-8.9.3/8.9.3) id JAA23991; Fri, 30 Mar 2001 09:50:02 -0800 (PST) Date: Fri, 30 Mar 2001 09:50:02 -0800 (PST) From: Jack Carter Message-Id: <10103300950.ZM23719@johmar.engr.sgi.com> In-Reply-To: "Chan, Sun C" "RE: auto-pic or non-auto-pic" (Mar 30, 8:58am) References: <9287DC1579B0D411AA2F009027F44C3F042DF9C4@FMSMSX41> X-Mailer: Z-Mail-SGI (3.2S.3 08feb96 MediaMail) To: "Chan, Sun C" , "'Kratzer, Willi'" , "'pro64-support@oss.sgi.com'" Subject: Re: auto-pic or non-auto-pic Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-pro64-support@oss.sgi.com Precedence: bulk I thought the golden age of ia64 ELF ABI was to have only PIC? The non-shared options were there only to get things moving until dso's came on the scene. We have dso support now so why are we still allowing non-shared objects? If the first assumption I stated is correct then the longer we tolerate the non-shared objects to proliferate then the messier things will get in the future because archives tend to linger around forever. Jack From owner-pro64-support@oss.sgi.com Fri Mar 30 09:52:27 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UHqRr05833 for pro64-support-outgoing; Fri, 30 Mar 2001 09:52:27 -0800 Received: from hypnos.cps.intel.com (hypnos.cps.intel.com [192.198.165.17]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UHqRM05830 for ; Fri, 30 Mar 2001 09:52:27 -0800 Received: from SMTP (fmsmsxvs03-1.fm.intel.com [132.233.42.203]) by hypnos.cps.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id RAA04775; Fri, 30 Mar 2001 17:52:01 GMT Received: from fmsmsx19.fm.intel.com ([132.233.48.19]) by 132.233.48.203 (Norton AntiVirus for Internet Email Gateways 1.0) ; Fri, 30 Mar 2001 17:52:01 0000 (GMT) Received: by fmsmsx19.fm.intel.com with Internet Mail Service (5.5.2653.19) id ; Fri, 30 Mar 2001 09:51:59 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9C6@FMSMSX41> From: "Chan, Sun C" To: "'Jack Carter'" , "Chan, Sun C" , "'Kratzer, Willi'" , "'pro64-support@oss.sgi.com'" Subject: RE: auto-pic or non-auto-pic Date: Fri, 30 Mar 2001 09:51:56 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Jack, the world is not just SGI anymore. I believe most compilers defaults to non-shared/static. Sun > -----Original Message----- > From: Jack Carter [mailto:jcarter@johmar.engr.sgi.com] > Sent: Friday, March 30, 2001 9:50 AM > To: Chan, Sun C; 'Kratzer, Willi'; 'pro64-support@oss.sgi.com' > Subject: Re: auto-pic or non-auto-pic > > > I thought the golden age of ia64 ELF ABI was to have only > PIC? The non-shared options were there only to get things > moving until dso's came on the scene. > > We have dso support now so why are we still allowing non-shared > objects? If the first assumption I stated is correct then the > longer we tolerate the non-shared objects to proliferate then > the messier things will get in the future because archives > tend to linger around forever. > > Jack > From owner-pro64-support@oss.sgi.com Fri Mar 30 14:46:51 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UMkpW11021 for pro64-support-outgoing; Fri, 30 Mar 2001 14:46:51 -0800 Received: from osc.edu (osc.edu [192.148.249.4]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UMkoM11018 for ; Fri, 30 Mar 2001 14:46:50 -0800 Received: from neptune.osc.edu (IDENT:djohnson@neptune.osc.edu [192.148.249.73]) by osc.edu (8.9.3/8.9.3/OSC 2.0) with ESMTP id RAA26945; Fri, 30 Mar 2001 17:46:14 -0500 (EST) Date: Fri, 30 Mar 2001 17:46:13 -0500 (EST) From: Douglas Johnson X-Sender: djohnson@localhost.localdomain To: "Chan, Sun C" cc: "'Kratzer, Willi'" , "'pro64-support@oss.sgi.com'" Subject: RE: auto-pic or non-auto-pic In-Reply-To: <9287DC1579B0D411AA2F009027F44C3F042DF9C4@FMSMSX41> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN Content-Transfer-Encoding: 8BIT Sender: owner-pro64-support@oss.sgi.com Precedence: bulk This object is the only one in the blas library with, private flags = LE, NOFUNCDESC_CONS_GP, ABI64 The rest are, private flags = LE, ABI64 I can't figure out what flags were used to get the NOFUNCDESC_CONS_GP, it's not just using PIC. I'd report this bug to the web site that is referenced in the release notes, but I'm still waitng on my login id and password. If every object is compiled and linked with -non_shared, the following errors still occur, /usr/bin/ld: smpsrv64.o: linking auto-pic files with non-auto-pic files Bad value: failed to merge target specific data of file /usr/local/mkl/LIB/libmkl64_itp.a(smpsrv64.o) /usr/bin/ld: pberror_.o: @gprel relocation against dynamic symbol stderr@@GLIBC_2.2 /usr/bin/ld: pberror_.o: @gprel relocation against dynamic symbol stderr@@GLIBC_2.2 /usr/bin/ld: pberror_.o: @gprel relocation against dynamic symbol stderr@@GLIBC_2.2 /usr/bin/ld: pberror_.o: @gprel relocation against dynamic symbol stderr@@GLIBC_2.2 Doug On Fri, 30 Mar 2001, Chan, Sun C wrote: > do "file smpsrv64.o" to see what type of object it is. > I believe Intel defaults to non-shared (non-pic) while Pro64 defaults > to shared. If you have source to libmkl64_itp, rebuild that with > Pro64 compiler and put it in the same directory with the .a. It should > fix your problem. Another way is to try adding -non_shared flag. > The potential problem there is that there might not be a matching > pair of non_shared fortran libraries. > Sun > > > -----Original Message----- > > From: Kratzer, Willi [mailto:Willi.Kratzer@fujitsu-siemens.com] > > Sent: Friday, March 30, 2001 6:18 AM > > To: 'pro64-support@oss.sgi.com' > > Subject: auto-pic or non-auto-pic > > > > > > Hi, > > > > when I try to link objects from the Pro64 compilers > > with Intel`s MKL (Mathematical Kernel Library) on an > > Itanium system, I get the messages > > > > /usr/lib/gcc-lib/ia64-redhat-linux/2.96-ia64-000717/../../../../ > > ia64-redhat-linux/bin/ld: smpsrv64.o: linking auto-pic files with > > non-auto-pic files > > Bad value: failed to merge target specific data of file > > /usr/local/intel/mkl/LIB/libmkl64_itp.a(smpsrv64.o) > > collect2: ld returned 1 exit status > > > > First question: What do the Pro64 compilers generate? > > auto-pic files or non-auto-pic files? > > Second question: Can I control it? > > > > Mit freundlichem Gruß / Best Regards > > Willi Kratzer > > > > --- > > Willi Kratzer, Fujitsu Siemens Computers > > PO PC Workstations > > Buergermeister-Ulrich-Str. 100 > > D-86199 Augsburg, Germany > > Tel: (..821)-804-4810 > > Fax: (..821)-804-2568 > > Email: willi.kratzer@fujitsu-siemens.com > > > > > -- Doug Johnson Systems Developer/Engineer The Ohio Supercomputer Center djohnson@osc.edu From owner-pro64-support@oss.sgi.com Fri Mar 30 15:13:13 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2UNDDN11420 for pro64-support-outgoing; Fri, 30 Mar 2001 15:13:13 -0800 Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2UNDDM11417 for ; Fri, 30 Mar 2001 15:13:13 -0800 Received: from SMTP (orsmsxvs02-1.jf.intel.com [192.168.65.201]) by ganymede.or.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.35 2001/02/12 09:03:45 smothers Exp $) with SMTP id XAA07667; Fri, 30 Mar 2001 23:12:37 GMT Received: from orsmsx29.jf.intel.com ([192.168.70.29]) by 192.168.70.201 (Norton AntiVirus for Internet Email Gateways 1.0) ; Fri, 30 Mar 2001 23:12:37 0000 (GMT) Received: by orsmsx29.jf.intel.com with Internet Mail Service (5.5.2653.19) id ; Fri, 30 Mar 2001 15:12:35 -0800 Message-ID: <9287DC1579B0D411AA2F009027F44C3F042DF9CA@FMSMSX41> From: "Chan, Sun C" To: "'Douglas Johnson'" , "Chan, Sun C" Cc: "'Kratzer, Willi'" , "'pro64-support@oss.sgi.com'" Subject: RE: auto-pic or non-auto-pic Date: Fri, 30 Mar 2001 15:12:28 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Looks like I've spoken too hastily. It is not simply pic vs non-pic. I apologize. Apparently, it has to do with the function descriptor portion of the ABI. Sun > -----Original Message----- > From: Douglas Johnson [mailto:djohnson@osc.edu] > Sent: Friday, March 30, 2001 2:46 PM > To: Chan, Sun C > Cc: 'Kratzer, Willi'; 'pro64-support@oss.sgi.com' > Subject: RE: auto-pic or non-auto-pic > > > This message uses a character set that is not supported by > the Internet Service. To view the original message content, > open the attached message. If the text doesn't display > correctly, save the attachment to disk, and then open it > using a viewer that can display the original character set. > From owner-pro64-support@oss.sgi.com Fri Mar 30 20:11:13 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2V4BD416875 for pro64-support-outgoing; Fri, 30 Mar 2001 20:11:13 -0800 Received: from pneumatic-tube.sgi.com (pneumatic-tube.sgi.com [204.94.214.22]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2V4BCM16872 for ; Fri, 30 Mar 2001 20:11:12 -0800 Received: from dsl-proof.corp.sgi.com (dsl-proof.corp.sgi.com [192.102.142.250]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id UAA01354 for ; Fri, 30 Mar 2001 20:21:06 -0800 (PST) mail_from (dlstephe@sgi.com) Received: from sgi.com (localhost [127.0.0.1]) by dsl-proof.corp.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id UAA49572; Fri, 30 Mar 2001 20:13:39 -0800 (PST) Message-ID: <3AC55972.BBF6CA5D@sgi.com> Date: Fri, 30 Mar 2001 20:13:38 -0800 From: David Stephenson Organization: SGI -- Compilers X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP32) X-Accept-Language: en MIME-Version: 1.0 To: pro64-support@oss.sgi.com Subject: Feedback in Pro64 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-pro64-support@oss.sgi.com Precedence: bulk Below are some notes on feedback in the SGI Pro64 compiler. How To Use Feedback ------------------- Using feedback is a three step process: (1) Compile the source with intrumentation turned on, producing a binary; (2) Execute the binary on sample input, generating a file of feedback data; (3) Recompile the source with annotation turned on. Feedback instrumentation uses the libinstr.so library. Suppose you want to run frequency for the program test.c, and you want to store the data in a file fbdata. (Usually, I use the same name for the feedback data file and the program, so fbdata would be test.) First, compile test.c with instrumentation on. "" is the path to libinstr.so sgicc -fb_create fbdata -L -linstr test.c Second, execute the compiled program on the sample input data. The instrumented program will run significantly slower than usual. ./a.out This will create a file of frequency data, named "fbpath.instr0.######". The 0 in "instr0" indicates that instrumentation occurred before VHO lowering. (Eventually, it should be possible to instrument at a variety of stages.) You can perform multiple runs using different input data to generate multiple frequency data files. Third, recompile test.c using the feedback data. Feedback data will be read and combined from all files with names "fbpath.instr0.######". (So, you need to delete obsolete feedback data files or use a different feedback file name.) $TOOLROOT/usr/bin/cc -fb_opt fbdata test.c Any flags that affect compilation should be the same as during the first, instumentation compilation. What Feedback Does ------------------ Feedback collects frequency counts on control flow related whirl instructions, such as branches, loops, and procedure calls. Memory accesses are not counted. During the instrumentation step, the compiler inserts extra code into the whirl representation. The insterted code consists of procedure calls to libinstr.so routines which gather control flow data, such has the number of times a branch is taken or not taken, how many times a loop iterates, and how many times procedures are invoked. When the instrumented binary is invoked on the test data, the program performs as normal, except that the libinstr.so routines collect control flow frequencies and store the counts into a file when the program completes. During the annotation step, the compiler reads the frequency data from any available feedback data files and "attaches" the data to the whirl nodes. Annotation occurs at the same point of compilation that instrumentation occured (warly in the backend, just before VHO lowering). In fact, the wrirl representation look the same as it did during the instrumentation step, so the same compiler flags should be used. Later phases of the compiler (particular IPA and CG) then can use the data to guide optimization decisions. Compiler optimizations that transform the whirl control flow structure also invoke procedures that update the frequency counts, sometimes making educated guesses when the exact counts cannot be determined. Trace Flags ----------- If you want to see the frequency counts during the second compilation, use the flags: -Wb,-tt16:0x1111 This will cause frequency data to be dumped to the trace file test.t. For debugging purposes, inconsistancies in the feedback data are also identified in the trace file. Other tracing flags are defined in be/com/fb_whirl.h Internally, frequency values are stored as 64-bit FB_FREQ data types, which can have one of the following values: --> an EXACT nonnegative 32-bit float value --> a GUESS nonnegative 32-bit float value --> UNKNOWN --> UNINIT --> ERROR EXACT values give a precise count of the number of times that a particular cfg edge was traveled during the program's instrumentated executions. GUESS values are educated estimates of the count. GUESS and UNKNOWN values can be introduced when the compiler transforms the controlflow of the whirl representation. The symbols "!" and "?" are used to distinguish between EXACT and GUESS frequencies in trace files. One Warning ----------- Feedback is significantly less finished and tested than the rest of the compiler. From owner-pro64-support@oss.sgi.com Fri Mar 30 20:36:47 2001 Received: (from majordomo@localhost) by oss.sgi.com (8.11.3/8.11.3) id f2V4ale17210 for pro64-support-outgoing; Fri, 30 Mar 2001 20:36:47 -0800 Received: from scapa.cs.ualberta.ca (root@scapa.cs.ualberta.ca [129.128.4.44]) by oss.sgi.com (8.11.3/8.11.3) with ESMTP id f2V4akM17207 for ; Fri, 30 Mar 2001 20:36:46 -0800 Received: (from localhost user: 'pengzhao' uid#483 fake: STDIN (pengzhao@peers)) by scapa.cs.ualberta.ca id ; Fri, 30 Mar 2001 21:36:36 -0700 Date: Fri, 30 Mar 2001 21:36:38 -0700 (MST) From: Peng Zhao To: sgi Subject: failed link of pro64 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-pro64-support@oss.sgi.com Precedence: bulk http://oss.sgi.com/projects/Pro64/support.html | | | ----->http://oss.sgi.com/projects/Pro64/mail/pro64-support/thread.html Not Found -- Regards Peng Peng Zhao pengzhao@cs.ualberta.ca http://www.cs.ualberta.ca/~pengzhao TEL (Lab): (780)492-3725 Lab: CSC251