By: Paradox (paradox1000.delete@this.gmail.com), February 9, 2010 8:33 am
Room: Moderated Discussions
Michael S (already5chosen@yahoo.com) on 2/9/10 wrote:
---------------------------
>Linus Torvalds (torvalds@linux-foundation.org) on 2/9/10 wrote:
>---------------------------
>>
>>At 3.33Ghz, the four-core Nehalem will likely beat
>>POWER7, though, so Intel probably retains the SpecInt
>>crown. Anybody want to guess at actual numbers?
>>
>>(Of course, Nehalem has that Intel "libquantum" advantage.
>>That thing should just be removed from specint)
>>
>>Linus
>
>Unfortunately, due to autopar, Nehalem (and SPARC64) SpecInt2k6 scores have very
>remote relationship to single-threaded performance.
>SpecInt as a measure of single-threaded performance is now defunct and we should look for something else.
>Now what's the chance that we'll find the same "something else" for both Nehalem and POWER7? I'd say, close to zero.
Well, we could compile SPEC ourselves using GCC with no autopar on both machines...
I think it would be interesting to compare the performance of a number of programs on both platforms running Linux, maybe simply installed with the standard configure, make, make install.
---------------------------
>Linus Torvalds (torvalds@linux-foundation.org) on 2/9/10 wrote:
>---------------------------
>>
>>At 3.33Ghz, the four-core Nehalem will likely beat
>>POWER7, though, so Intel probably retains the SpecInt
>>crown. Anybody want to guess at actual numbers?
>>
>>(Of course, Nehalem has that Intel "libquantum" advantage.
>>That thing should just be removed from specint)
>>
>>Linus
>
>Unfortunately, due to autopar, Nehalem (and SPARC64) SpecInt2k6 scores have very
>remote relationship to single-threaded performance.
>SpecInt as a measure of single-threaded performance is now defunct and we should look for something else.
>Now what's the chance that we'll find the same "something else" for both Nehalem and POWER7? I'd say, close to zero.
Well, we could compile SPEC ourselves using GCC with no autopar on both machines...
I think it would be interesting to compare the performance of a number of programs on both platforms running Linux, maybe simply installed with the standard configure, make, make install.
Topic | Posted By | Date |
---|---|---|
POWER7 Spec | Paradox | 2010/02/08 11:05 AM |
POWER7 Spec | Thu Nguyen | 2010/02/08 12:58 PM |
POWER7 Spec | Ian Ameline | 2010/02/08 09:22 PM |
POWER7 Spec | Thu Nguyen | 2010/02/08 11:54 PM |
POWER7 Spec | Ian Ameline | 2010/02/09 06:46 AM |
POWER7 Spec | Linus Torvalds | 2010/02/09 07:57 AM |
industry-standard single-threaded performance benchmarks absent | Michael S | 2010/02/09 08:09 AM |
industry-standard single-threaded performance benchmarks absent | Paradox | 2010/02/09 08:33 AM |
industry-standard single-threaded performance benchmarks absent | Michael S | 2010/02/09 09:30 AM |
industry-standard single-threaded performance benchmarks absent | Paradox | 2010/02/09 10:52 AM |
industry-standard single-threaded performance benchmarks absent | Jesper Frimann | 2010/02/09 11:33 AM |
industry-standard single-threaded performance benchmarks absent | Linus Torvalds | 2010/02/09 08:48 AM |
industry-standard single-threaded performance benchmarks absent | Michael S | 2010/02/09 09:26 AM |
industry-standard single-threaded performance benchmarks absent | Linus Torvalds | 2010/02/09 09:58 AM |
industry-standard single-threaded performance benchmarks absent | Jesper Frimann | 2010/02/09 12:17 PM |
industry-standard single-threaded performance benchmarks absent | Linus Torvalds | 2010/02/09 12:54 PM |
industry-standard single-threaded performance benchmarks absent | Jesper Frimann | 2010/02/09 02:10 PM |
industry-standard single-threaded performance benchmarks absent | Paradox | 2010/02/09 01:22 PM |
industry-standard single-threaded performance benchmarks absent | anon | 2010/02/09 10:21 AM |
POWER7 Spec | Jesper Frimann | 2010/02/09 12:30 PM |
POWER7 Spec | Doug Siebert | 2010/02/09 05:38 PM |
POWER7 Spec | Linus Torvalds | 2010/02/09 07:28 PM |
POWER7 Spec | Linus Torvalds | 2010/02/09 07:28 PM |
POWER7 Spec | Ian Ameline | 2010/02/09 08:02 PM |
POWER7 Spec | Doug Siebert | 2010/02/09 10:18 PM |
POWER7 Spec | someone | 2010/02/09 08:20 PM |
POWER7 Spec | Linus Torvalds | 2010/02/10 09:17 AM |
POWER7 Spec | Linus Torvalds | 2010/02/10 09:46 AM |
POWER7 Spec | Emil | 2010/02/10 11:06 AM |
POWER7 Spec | Ian Ameline | 2010/02/10 10:13 AM |
POWER7 Spec | someone | 2010/02/10 11:01 AM |
POWER7 Spec | Linus Torvalds | 2010/02/10 11:32 AM |
POWER7 Spec | Michael S | 2010/02/10 12:30 PM |
POWER7 Spec | Linus Torvalds | 2010/02/10 01:25 PM |
POWER7 Spec | mpx | 2010/02/10 02:58 PM |
POWER7 Spec | nemlis | 2010/02/11 12:24 AM |
POWER7 Spec | none | 2010/02/11 12:52 AM |
POWER7 Spec | nemlis | 2010/02/11 01:52 PM |
POWER7 Spec | mpx | 2010/02/09 08:18 AM |
POWER7 Spec | Michael S | 2010/02/09 09:08 AM |
POWER7 Spec | Jesper Frimann | 2010/02/16 02:29 AM |
POWER7 Spec | Jamie Lucier | 2010/02/16 06:35 AM |
POWER7 Spec | anon | 2010/02/16 07:16 AM |
POWER7 Spec | Jamie Lucier | 2010/02/16 10:46 AM |
POWER7 Spec | anon | 2010/02/16 03:42 PM |
POWER7 Spec | Jamie Lucier | 2010/02/18 06:07 AM |
POWER7 Spec | Thu Nguyen | 2010/02/16 01:23 PM |
POWER7 Spec | Paradox | 2010/02/16 09:57 AM |
POWER7 Spec | Jamie Lucier | 2010/02/16 11:22 AM |
POWER7 Spec | Paradox | 2010/02/16 02:00 PM |
POWER7 Spec | Jesper Frimann | 2010/02/16 02:14 PM |
POWER7 Spec | Jamie Lucier | 2010/02/17 10:41 AM |
POWER7 Spec | Jesper Frimann | 2010/02/09 04:43 AM |