Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*Csum\s+and\s+csum\s+copyroutines\s+benchmark\s*$/: 12 ]

Total 12 documents matching your query.

1. ion] SMP for Linux (score: 1)
Author: "Larry Sendlosky" <Larry.Sendlosky@xxxxxxxxxxxx>
Date: Fri, 25 Oct 2002 09:36:22 -0200
ort TX checksumming. The disk DMAs to ram, then the net card
/archives/netdev/2002-10/msg00512.html (16,189 bytes)

2. ralito !!!!! (score: 1)
Author: xxxxxxxxx>
Date: 25 Oct 2002 10:48:10 +0300
his also lives in ICMP(v4)). --yoshfuji
/archives/netdev/2002-10/msg00515.html (13,754 bytes)

3. ort, v2 (is Re: [PATCH] IPv6: Allow Both (score: 1)
Author: on Fraser" <J_Fraser@xxxxxxxxxxx>
Date: Fri, 25 Oct 2002 11:59:06 -0200
ity. -- Robert Jordan: A Crown of Swords
/archives/netdev/2002-10/msg00516.html (12,740 bytes)

4. Re: Csum and csum copyroutines benchmark (score: 1)
Author: aser@xxxxxxxxxxx>
Date: 25 Oct 2002 12:47:05 +0300
nts in the later kernel over the former.
/archives/netdev/2002-10/msg00517.html (13,965 bytes)

5. Re: Csum and csum copyroutines benchmark (score: 1)
Author: i / 吉藤英明 <yoshfuji@xxxxxxxxxxxxxx>
Date: 25 Oct 2002 11:19:51 +0100
S file: /cvsroot/usagi/usagi-backport/lin
/archives/netdev/2002-10/msg00518.html (11,218 bytes)

6. Re: Csum and csum copyroutines benchmark (score: 1)
Author: xxxxx>
Date: Fri, 25 Oct 2002 14:00:37 -0200
" So, how does one use sendfile() for lar
/archives/netdev/2002-10/msg00519.html (11,938 bytes)

7. Csum and csum copyroutines benchmark (score: 1)
Author: Denis Vlasenko <vda@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Fri, 25 Oct 2002 09:36:22 -0200
Short conclusion: 1. It is possible to speed up csum routines for AMD processors by 30%. 2. It is possible to speed up csum_copy routines for both AMD and Intel three times or more. Roy, do you like
/archives/netdev/2002-10/msg01123.html (16,313 bytes)

8. Re: Csum and csum copyroutines benchmark (score: 1)
Author: Momchil Velikov <velco@xxxxxxxxx>
Date: 25 Oct 2002 10:48:10 +0300
Additional data point: Short summary: 1. Checksum - kernelpii_csum is ~19% faster 2. Copy - lernelpii_csum is ~6% faster Dual Pentium III, 1266Mhz, 512K cache, 2G SDRAM (133Mhz, ECC) The only change
/archives/netdev/2002-10/msg01126.html (13,911 bytes)

9. Re: Csum and csum copyroutines benchmark (score: 1)
Author: Denis Vlasenko <vda@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Fri, 25 Oct 2002 11:59:06 -0200
[please drop libc from CC:] Well, that makes it run entirely in L0 cache. This is unrealistic for actual use. movntq is x3 faster when you hit RAM instead of L0. You need to be more clever than that
/archives/netdev/2002-10/msg01127.html (12,847 bytes)

10. Re: Csum and csum copyroutines benchmark (score: 1)
Author: Momchil Velikov <velco@xxxxxxxxx>
Date: 25 Oct 2002 12:47:05 +0300
Oops ... Here it is: Csum benchmark program buffer size: 1 K Each test tried 1024 times, max and min CPU cycles are reported. Please disregard max values. They are due to system interference only. c
/archives/netdev/2002-10/msg01128.html (14,116 bytes)

11. Re: Csum and csum copyroutines benchmark (score: 1)
Author: Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>
Date: 25 Oct 2002 11:19:51 +0100
In a lot of cases its extremely realistic to assume the network buffers are in cache. The copy/csum path is often touching just generated data, or data we just accessed via read(). The csum RX path f
/archives/netdev/2002-10/msg01129.html (11,369 bytes)

12. Re: Csum and csum copyroutines benchmark (score: 1)
Author: Denis Vlasenko <vda@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Fri, 25 Oct 2002 14:00:37 -0200
'Touching' is not interesting since it will pump data into cache, no matter how you 'touch' it. Running benchmarks against 1K static buffer makes cache red hot and causes _all writes_ to hit it. It m
/archives/netdev/2002-10/msg01130.html (12,067 bytes)


This search system is powered by Namazu