Schwarzkaeppchen

アクセスカウンタ

zoom RSS GeForce GTX 460 のdistributed.net CUDAクライアントベンチ結果

<<   作成日時 : 2010/12/11 16:35   >>

ブログ気持玉 0 / トラックバック 0 / コメント 0

GeForce GTX460 (1024MBytes Memory, Overclock model)を導入したのでベンチマーク結果を貼るよ。
(Palit GeForce GTX 460 1GB Sonic Platinum (NE5X460HF1102))

CUDA 2.2とCUDA 3.1の両方で試してみました。
dnetc516-win32-x86-cuda22 : v2.9107.516
dnetc517b-win32-x86-cuda31 : v2.9108.517b

[結果]
全体にCUDA 3.1の方がいいなぁ。Coreは#9が速いけど僅差で#0。#9はCPU時間も全部持っていくので#0がいいのかなぁ。#10はCPU時間をほとんど食わないのだけども#0との差が50 Mkeys/sあるので#0かねぇ (9800GTなどは差が少なかったので#10が有効だった)。

採用 : CUDA 3.1でコアは#0 (default coreが#0なので設定はいらないです)。

[感想]
で、実際に回してみるとCPUパワーが足りていないかも?(タスクマネージャを見ながらレートを見ている印象で)。運用環境はAthlon 64 X2 の 2GHzです。あとmicroATXケースに詰め込んだので冷却が足りてない感が…#10コアで85℃、#0コアだと90℃まで達する。隣のPCIスロットに刺さってる奴をどけるとか対策したい→どけたら5℃下がった。PCIeスロットとPCIスロットの位置関係が現状と違うマザーボードが欲しいですねぇ…そうすればPCIスロットのカードも使える(できれば使いたいので)。

[CUDA 2.2]
C:\Program Files (x86)\dnetc516-win32-x86-cuda22.v2.9107.516>dnetc -bench > bench2.2.txt

distributed.net client for CUDA 2.2 on Win32 Copyright 1997-2009, distribut ...
Please visit http://www.distributed.net/ for up-to-date contest information.
Start the client with '-help' for a list of valid command line options.


dnetc v2.9107-516-CTR-09122713 for CUDA 2.2 on Win32 (WindowsNT 6.1).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at http://bugs.distributed.net/
Using email address (distributed.net ID) '.@.'

[Dec 11 07:05:14 UTC] nvcuda.dll Version: 8.17.12.6099
[Dec 11 07:05:14 UTC] RC5-72: using core #0 (CUDA 1-pipe 64-thd).
[Dec 11 07:05:31 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 64-thd)
0.00:00:14.18 [304,498,531 keys/sec]
[Dec 11 07:05:31 UTC] RC5-72: using core #1 (CUDA 1-pipe 128-thd).
[Dec 11 07:05:49 UTC] RC5-72: Benchmark for core #1 (CUDA 1-pipe 128-thd)
0.00:00:16.08 [268,249,107 keys/sec]
[Dec 11 07:05:49 UTC] RC5-72: using core #2 (CUDA 1-pipe 256-thd).
[Dec 11 07:06:09 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 256-thd)
0.00:00:17.11 [184,558,071 keys/sec]
[Dec 11 07:06:09 UTC] RC5-72: using core #3 (CUDA 2-pipe 64-thd).
[Dec 11 07:06:28 UTC] RC5-72: Benchmark for core #3 (CUDA 2-pipe 64-thd)
0.00:00:16.44 [268,458,673 keys/sec]
[Dec 11 07:06:28 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Dec 11 07:06:48 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:17.03 [187,116,935 keys/sec]
[Dec 11 07:06:48 UTC] RC5-72: using core #5 (CUDA 2-pipe 256-thd).
[Dec 11 07:07:08 UTC] RC5-72: Benchmark for core #5 (CUDA 2-pipe 256-thd)
0.00:00:16.45 [189,408,270 keys/sec]
[Dec 11 07:07:08 UTC] RC5-72: using core #6 (CUDA 4-pipe 64-thd).
[Dec 11 07:07:27 UTC] RC5-72: Benchmark for core #6 (CUDA 4-pipe 64-thd)
0.00:00:17.03 [184,558,071 keys/sec]
[Dec 11 07:07:27 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Dec 11 07:07:47 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:16.50 [189,541,750 keys/sec]
[Dec 11 07:07:47 UTC] RC5-72: using core #8 (CUDA 4-pipe 256-thd).
[Dec 11 07:08:07 UTC] RC5-72: Benchmark for core #8 (CUDA 4-pipe 256-thd)
0.00:00:17.39 [190,870,551 keys/sec]
[Dec 11 07:08:07 UTC] RC5-72: using core #9 (CUDA 1-pipe 64-thd busy wait).
[Dec 11 07:08:24 UTC] RC5-72: Benchmark for core #9 (CUDA 1-pipe 64-thd bus ...
0.00:00:14.13 [306,526,721 keys/sec]
[Dec 11 07:08:24 UTC] RC5-72: using core #10 (CUDA 1-pipe 64-thd sleep 100us).
[Dec 11 07:08:43 UTC] RC5-72: Benchmark for core #10 (CUDA 1-pipe 64-thd sl ...
0.00:00:16.78 [257,015,053 keys/sec]
[Dec 11 07:08:43 UTC] RC5-72: using core #11 (CUDA 1-pipe 64-thd sleep dyna ...
[Dec 11 07:09:02 UTC] RC5-72: Benchmark for core #11 (CUDA 1-pipe 64-thd sl ...
0.00:00:16.08 [268,070,041 keys/sec]
[Dec 11 07:09:02 UTC] RC5-72 benchmark summary :
Default core : #0 (CUDA 1-pipe 64-thd)
Fastest core : #9 (CUDA 1-pipe 64-thd busy wait)
[Dec 11 07:09:02 UTC] Core #9 is marginally faster than the default core.
Testing variability might lead to pick one or the other.



[CUDA 3.1]
C:\Program Files (x86)\dnetc517b-win32-x86-cuda31.v2.9108.517b>dnetc -bench > bench3.1.txt

distributed.net client for CUDA 3.1 on Win32 Copyright 1997-2009, distribut ...
Please visit http://www.distributed.net/ for up-to-date contest information.
Start the client with '-help' for a list of valid command line options.


dnetc v2.9108-517-CTR-10062905 for CUDA 3.1 on Win32 (WindowsNT 6.1).
Please provide the *entire* version descriptor when submitting bug reports.
The distributed.net bug report pages are at http://bugs.distributed.net/
Using email address (distributed.net ID) '.@.'

[Dec 11 07:16:39 UTC] nvcuda.dll Version: 8.17.12.6099
[Dec 11 07:16:40 UTC] RC5-72: using core #0 (CUDA 1-pipe 64-thd).
[Dec 11 07:16:57 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 64-thd)
0.00:00:13.96 [313,092,494 keys/sec]
[Dec 11 07:16:57 UTC] RC5-72: using core #1 (CUDA 1-pipe 128-thd).
[Dec 11 07:17:16 UTC] RC5-72: Benchmark for core #1 (CUDA 1-pipe 128-thd)
0.00:00:16.11 [269,518,510 keys/sec]
[Dec 11 07:17:16 UTC] RC5-72: using core #2 (CUDA 1-pipe 256-thd).
[Dec 11 07:17:35 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 256-thd)
0.00:00:17.06 [186,954,929 keys/sec]
[Dec 11 07:17:35 UTC] RC5-72: using core #3 (CUDA 2-pipe 64-thd).
[Dec 11 07:17:54 UTC] RC5-72: Benchmark for core #3 (CUDA 2-pipe 64-thd)
0.00:00:16.11 [268,804,883 keys/sec]
[Dec 11 07:17:54 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Dec 11 07:18:14 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:17.03 [186,954,929 keys/sec]
[Dec 11 07:18:14 UTC] RC5-72: using core #5 (CUDA 2-pipe 256-thd).
[Dec 11 07:18:33 UTC] RC5-72: Benchmark for core #5 (CUDA 2-pipe 256-thd)
0.00:00:16.50 [189,408,270 keys/sec]
[Dec 11 07:18:34 UTC] RC5-72: using core #6 (CUDA 4-pipe 64-thd).
[Dec 11 07:18:53 UTC] RC5-72: Benchmark for core #6 (CUDA 4-pipe 64-thd)
0.00:00:17.02 [187,116,935 keys/sec]
[Dec 11 07:18:53 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Dec 11 07:19:13 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:16.45 [189,408,270 keys/sec]
[Dec 11 07:19:13 UTC] RC5-72: using core #8 (CUDA 4-pipe 256-thd).
[Dec 11 07:19:32 UTC] RC5-72: Benchmark for core #8 (CUDA 4-pipe 256-thd)
0.00:00:16.59 [199,571,253 keys/sec]
[Dec 11 07:19:32 UTC] RC5-72: using core #9 (CUDA 1-pipe 64-thd busy wait).
[Dec 11 07:19:49 UTC] RC5-72: Benchmark for core #9 (CUDA 1-pipe 64-thd bus ...
0.00:00:13.77 [315,248,528 keys/sec]
[Dec 11 07:19:49 UTC] RC5-72: using core #10 (CUDA 1-pipe 64-thd sleep 100us).
[Dec 11 07:20:08 UTC] RC5-72: Benchmark for core #10 (CUDA 1-pipe 64-thd sl ...
0.00:00:16.56 [261,619,092 keys/sec]
[Dec 11 07:20:08 UTC] RC5-72: using core #11 (CUDA 1-pipe 64-thd sleep dyna ...
[Dec 11 07:20:26 UTC] RC5-72: Benchmark for core #11 (CUDA 1-pipe 64-thd sl ...
0.00:00:16.08 [268,468,054 keys/sec]
[Dec 11 07:20:26 UTC] RC5-72 benchmark summary :
Default core : #0 (CUDA 1-pipe 64-thd)
Fastest core : #9 (CUDA 1-pipe 64-thd busy wait)
[Dec 11 07:20:26 UTC] Core #9 is marginally faster than the default core.
Testing variability might lead to pick one or the other.

テーマ

関連テーマ 一覧


月別リンク

ブログ気持玉

クリックして気持ちを伝えよう!
ログインしてクリックすれば、自分のブログへのリンクが付きます。
→ログインへ

トラックバック(0件)

タイトル (本文) ブログ名/日時

トラックバック用URL help


自分のブログにトラックバック記事作成(会員用) help

タイトル
本 文

コメント(0件)

内 容 ニックネーム/日時

コメントする help

ニックネーム
本 文
GeForce GTX 460 のdistributed.net CUDAクライアントベンチ結果 Schwarzkaeppchen/BIGLOBEウェブリブログ
文字サイズ:       閉じる