From: Dick Rous (dick++at++sgi.com)
Date: 07/10/2003 09:47:03
Before pfConfig(), there is a call to the function pfMultiprocess.
You need to OR PFMP_FORK_CULL_SIDEKICK with the tokens already present, eg:
pfMultiprocess(PFMP_APP_CULL_DRAW | PFMP_FORK_ISECT | PFMP_FORK_LPOINT |
PFMP_FORK_DBASE | PFMP_FORK_COMPUTE | PFMP_FORK_CULL_SIDEKICK)
Hope this helps,
Dick.
> -----Original Message-----
> From: M. Kragtwijk [mailto:M.Kragtwijk++at++rc.rug.nl]
> Sent: Thursday, July 10, 2003 14:11
> To: info-performer++at++sgi.com
> Subject: [info-performer] how to get cull_sidekick to work?
>
>
> I'm trying to get the CULL helper processes to work on our
> 16-CPU Onyx (IR3, IRIX
> 6.5.20, Performer 3.0.1). However, it seems as if the
> processes are not started at all.
> Here's a code fragment that shows what I've added to the
> C++-perfly source:
>
> from perfly.C:
>
> void
> initConfig(void)
> {
> int numCalli;
> if (ViewState->startCallig)
> {
> /* lets start the calligraphic boards if any */
> numCalli = StartCalligraphic(NumPipes);
> pfNotify(PFNFY_NOTICE, PFNFY_RESOURCE,
> "InitConfig() - found %d calligraphic board(s)", numCalli);
> if (numCalli)
> ViewState->calligBoard = 1;
> else
> ViewState->calligBoard = 0;
> }
>
> // our new additions:
> pfMultithread(0, PFPROC_CULL_SIDEKICK, 2); // pipenr,
> stage, nrofprocs
> pfMultithread(1, PFPROC_CULL_SIDEKICK, 2); // pipenr,
> stage, nrofprocs
> pfMultithread(2, PFPROC_CULL_SIDEKICK, 2); // pipenr,
> stage, nrofprocs
> }
>
> When I run it, I get the following messages:
>
> PF Info(2): All 16 processors available on
> this machine.
> PF Info: DBASE Process 207124 running on CPU 0.
> PF Info: ----------- Performer Process
> State --------
> PF Proc: APP pid:206860
> PF Proc: ISECT pid:207105
> PF Proc: DBASE pid:207124
> PF Proc: CLOCK pid:207028
> PF Proc: COMPUTE pid:206860
> PF Proc: SYNC pid:0
> PF Pipe Proc: CULL Pipes:3
> PF Proc: CULL Pipe:0 pid:207032
> PF Proc: CULL Pipe:1 pid:207020
> PF Proc: CULL Pipe:2 pid:207039
> PF Pipe Proc: DRAW Pipes:3
> PF Proc: DRAW Pipe:0 pid:207047
> PF Proc: DRAW Pipe:1 pid:207063
> PF Proc: DRAW Pipe:2 pid:207070
> PF Pipe Proc: LPOINT Pipes:3
> PF Proc: LPOINT Pipe:0 pid:0
> PF Proc: LPOINT Pipe:1 pid:0
> PF Proc: LPOINT Pipe:2 pid:0
> PF Pipe Proc: CULL_SIDEKICK Pipes:3
> PF Thread Proc: CULL_SIDEKICK
> Pipe:0 Threads:2
> PF Proc: CULL_SIDEKICK Pipe:0 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:0 Thread:0 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:0 Thread:1 pid:0
> PF Thread Proc: CULL_SIDEKICK
> Pipe:1 Threads:2
> PF Proc: CULL_SIDEKICK Pipe:1 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:1 Thread:0 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:1 Thread:1 pid:0
> PF Thread Proc: CULL_SIDEKICK
> Pipe:2 Threads:2
> PF Proc: CULL_SIDEKICK Pipe:2 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:2 Thread:0 pid:0
> PF Proc: CULL_SIDEKICK
> Pipe:2 Thread:1 pid:0
> PF
> -------------------------------------------
>
> The 'pid=0' clearly indicates that something is wrong here...
> By the way, the same thing happens when I run the 'standard' perfly.
> Do I have to do something to explicitely 'enable' the forking
> of the sidekick processes?
> Or is something else wrong?
>
> Thanks for any help,
>
> Martijn Kragtwijk
>
> --------------------------------------------------------------
> ---------
> List Archives, Info, FAQ: http://www.sgi.com/software/performer/
> Open Development Project: http://oss.sgi.com/projects/performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com
> --------------------------------------------------------------
> ---------
> SIGGRAPH 2003 Friends of Performer Meeting:
> Wednesday July 30 6:00pm San Diego Hilton Hotel Santa Rosa Room
> --------------------------------------------------------------
> ---------
>
This archive was generated by hypermail 2b29 : Thu Jul 10 2003 - 09:46:05 PDT