FAH 硬體及客戶端綜合討論 (part 2)

minihi001

F@H Taiwan Team
已加入
10/21/07
訊息
3,290
互動分數
0
點數
36
網站
sites.google.com
再幾天各家的新核心將會釋出來改善 lambda 封包的對應效能

期待吧 ;)

真期待4850可以大展雄風:PPP:
 

JCornell

5w337
已加入
5/16/04
訊息
2,448
互動分數
0
點數
0
今天SMP CPU怪怪的 ... ;ranger;
 

minihi001

F@H Taiwan Team
已加入
10/21/07
訊息
3,290
互動分數
0
點數
36
網站
sites.google.com
今天SMP CPU怪怪的 ... ;ranger;

我是GPU怪怪…
都不會跑;smash;
讓他閒了一整晚

[15:52:31] + Attempting to send results [September 12 15:52:31 UTC]
[15:52:31] - Reading file work/wuresults_09.dat from core
[15:52:31] (Read 597 bytes from disk)
[15:52:31] Connecting to http://171.64.65.103:8080/
[15:52:34] Posted data.
[15:52:34] Initial: 0000; - Uploaded at ~0 kB/s
[15:52:34] - Averaged speed for that direction ~0 kB/s
[15:52:34] + Results successfully sent
[15:52:34] Thank you for your contribution to Folding@Home.
[15:52:38] Trying to send all finished work units
[15:52:38] + No unsent completed units remaining.
[15:52:38] - Preparing to get new work unit...
[15:52:38] + Attempting to get work packet
[15:52:38] - Will indicate memory of 4094 MB
[15:52:38] - Connecting to assignment server
[15:52:38] Connecting to http://assign-GPU.stanford.edu:8080/
[15:52:42] Posted data.
[15:52:42] Initial: 40AB; - Successful: assigned to (171.64.65.103).
[15:52:42] + News From Folding@Home: GPU folding beta
[15:52:42] Loaded queue successfully.
[15:52:42] Connecting to http://171.64.65.103:8080/
[15:52:46] Posted data.
[15:52:46] Initial: 0000; - Receiving payload (expected size: 88375)
[15:53:14] - Downloaded at ~3 kB/s
[15:53:14] - Averaged speed for that direction ~2 kB/s
[15:53:14] + Received work.
[15:53:14] Trying to send all finished work units
[15:53:14] + No unsent completed units remaining.
[15:53:14] + Closed connections
[15:53:19]
[15:53:19] + Processing work unit
[15:53:19] Core required: FahCore_11.exe
[15:53:19] Core found.
[15:53:19] Working on queue slot 00 [September 12 15:53:19 UTC]
[15:53:19] + Working ...
[15:53:19] - Calling '.\FahCore_11.exe -dir work/ -suffix 00 -priority 96 -noassembly -nocpulock -checkpoint 3 -verbose -lifeline 5456 -version 620'

[15:53:19]
[15:53:19] *------------------------------*
[15:53:19] Folding@Home GPU Core - Beta
[15:53:19] Version 1.10 (Tue Aug 12 10:03:11 PDT 2008)
[15:53:19]
[15:53:19] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[15:53:19] Build host: amoeba
[15:53:19] Board Type: AMD
[15:53:19] Core :
[15:53:19] Preparing to commence simulation
[15:53:19] - User disabled assembly optimizations.
[15:53:19] - Created dyn
[15:53:19] - Files status OK
[15:53:19] - Expanded 87863 -> 447304 (decompressed 509.0 percent)
[15:53:19] Called DecompressByteArray: compressed_data_size=87863 data_size=447304, decompressed_data_size=447304 diff=0
[15:53:19] - Digital signature verified
[15:53:19]
[15:53:19] Project: 4742 (Run 0, Clone 118, Gen 4)
[15:53:19]
[15:53:19] Entering M.D.
[15:53:25] Working on p4742_lam5w_300K
[15:53:25] Client config found, loading data.
[15:53:25] Starting GUI Server
[15:53:28] mdrun_gpu returned -1
[15:53:28] Going to send back what have done.
[15:53:29] logfile size: 0 info=0 bed=0 hdr=25
[15:53:29] - Writing 558 bytes of core data to disk...
[15:53:29] Done: 46 -> 85 (compressed to 184.7 percent)
[15:53:29] ... Done.
[15:53:29]
[15:53:29] Folding@home Core Shutdown: EARLY_UNIT_END
[15:53:33] CoreStatus = 72 (114)
[15:53:33] Sending work to server
[15:53:33] Project: 4742 (Run 0, Clone 118, Gen 4)
[15:53:33] - Read packet limit of 540015616... Set to 524286976.


[15:53:33] + Attempting to send results [September 12 15:53:33 UTC]
[15:53:33] - Reading file work/wuresults_00.dat from core
[15:53:33] (Read 597 bytes from disk)
[15:53:33] Connecting to http://171.64.65.103:8080/
[15:53:36] Posted data.
[15:53:36] Initial: 0000; - Uploaded at ~0 kB/s
[15:53:36] - Averaged speed for that direction ~0 kB/s
[15:53:36] + Results successfully sent
[15:53:36] Thank you for your contribution to Folding@Home.
[15:53:40] Trying to send all finished work units
[15:53:40] + No unsent completed units remaining.
[15:53:40] - Preparing to get new work unit...
[15:53:40] + Attempting to get work packet
[15:53:40] - Will indicate memory of 4094 MB
[15:53:40] - Connecting to assignment server
[15:53:40] Connecting to http://assign-GPU.stanford.edu:8080/
[15:53:43] Posted data.
[15:53:43] Initial: 40AB; - Successful: assigned to (171.64.65.103).
[15:53:43] + News From Folding@Home: GPU folding beta
[15:53:43] Loaded queue successfully.
[15:53:43] Connecting to http://171.64.65.103:8080/
[15:53:50] Posted data.
[15:53:50] Initial: 0000; - Receiving payload (expected size: 88600)
[15:54:32] - Downloaded at ~2 kB/s
[15:54:32] - Averaged speed for that direction ~2 kB/s
[15:54:32] + Received work.
[15:54:32] Trying to send all finished work units
[15:54:32] + No unsent completed units remaining.
[15:54:32] + Closed connections
[15:54:37]
[15:54:37] + Processing work unit
[15:54:37] Core required: FahCore_11.exe
[15:54:37] Core found.
[15:54:37] Working on queue slot 01 [September 12 15:54:37 UTC]
[15:54:37] + Working ...
[15:54:37] - Calling '.\FahCore_11.exe -dir work/ -suffix 01 -priority 96 -noassembly -nocpulock -checkpoint 3 -verbose -lifeline 5456 -version 620'

[15:54:37]
[15:54:37] *------------------------------*
[15:54:37] Folding@Home GPU Core - Beta
[15:54:37] Version 1.10 (Tue Aug 12 10:03:11 PDT 2008)
[15:54:37]
[15:54:37] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[15:54:37] Build host: amoeba
[15:54:37] Board Type: AMD
[15:54:37] Core :
[15:54:37] Preparing to commence simulation
[15:54:37] - User disabled assembly optimizations.
[15:54:37] - Created dyn
[15:54:37] - Files status OK
[15:54:37] - Expanded 88088 -> 447304 (decompressed 507.7 percent)
[15:54:37] Called DecompressByteArray: compressed_data_size=88088 data_size=447304, decompressed_data_size=447304 diff=0
[15:54:38] - Digital signature verified
[15:54:38]
[15:54:38] Project: 4742 (Run 4, Clone 40, Gen 21)
[15:54:38]
[15:54:38] Entering M.D.
[15:54:44] Working on p4742_lam5w_300K
[15:54:44] Client config found, loading data.
[15:54:44] Starting GUI Server
[15:54:47] mdrun_gpu returned -1
[15:54:47] Going to send back what have done.
[15:54:48] logfile size: 0 info=0 bed=0 hdr=25
[15:54:48] - Writing 558 bytes of core data to disk...
[15:54:48] Done: 46 -> 85 (compressed to 184.7 percent)
[15:54:48] ... Done.
[15:54:48]
[15:54:48] Folding@home Core Shutdown: EARLY_UNIT_END
[15:54:52] CoreStatus = 72 (114)
[15:54:52] Sending work to server
[15:54:52] Project: 4742 (Run 4, Clone 40, Gen 21)
[15:54:52] - Read packet limit of 540015616... Set to 524286976.

一直出現,快被煩死了;ng;
 

monkp

... loading ...
已加入
5/1/07
訊息
1,037
互動分數
0
點數
0
ATI 新消息

剛得知一個關於 ATI 運算優化的好消息

AMD Stream SDK 近日剛推出版本 1.2 beta

似乎也稱做 ATI 版本的 "CUDA"

其實是 CAL 開發軟體的升級, 提高現有 Brook+ 語法的效能

也為未來的 OpenCL 和 DirectX 11 先行鋪路

官網已有回報 CPU 佔有率降至 60%, 換句話說, 不需要高時脈的 CPU 也能保持一定的效能, 像目前 NV 顯卡一樣.

但是 ...

要完全發揮 800SP 效能仍需要運算核心的優化, mhouston 有強調

看來 ...

使用 ATI 顯卡運算的隊友, 終於可以走出顯卡不濟的陰影了.

相關報導和討論

ATIs savior? Here's ATIs CUDA. [new CAL version]

AMD brings Vista into Stream SDK

AMD Stream SDK 1.2 下載
 

minihi001

F@H Taiwan Team
已加入
10/21/07
訊息
3,290
互動分數
0
點數
36
網站
sites.google.com
中午有看到囉

裝了
SDK v1.2-beta Windows Vista SP1 64-bit AMD Stream SDK built for 64-bit Windows Vista
CPU還是一樣吃100%;smash;
 

ssss8kimo

榮譽會員
已加入
11/10/07
訊息
1,156
互動分數
0
點數
36
年齡
37
不知到顯示卡的記憶體有沒有影響GPU運算?
256MB 512MB 不知有何差異
 

ssss8kimo

榮譽會員
已加入
11/10/07
訊息
1,156
互動分數
0
點數
36
年齡
37
剛有競標活動,3850 512MB,想說把舊的256MB換掉XD
不過沒標到 囧
 

cooky7620

一般般會員
已加入
5/4/08
訊息
138
互動分數
0
點數
16
年齡
30
網站
www.wretch.cc
剛得知一個關於 ATI 運算優化的好消息

AMD Stream SDK 近日剛推出版本 1.2 beta

似乎也稱做 ATI 版本的 "CUDA"

其實是 CAL 開發軟體的升級, 提高現有 Brook+ 語法的效能

也為未來的 OpenCL 和 DirectX 11 先行鋪路

官網已有回報 CPU 佔有率降至 60%, 換句話說, 不需要高時脈的 CPU 也能保持一定的效能, 像目前 NV 顯卡一樣.

但是 ...

要完全發揮 8...

灌了還是沒啥改變

是要等F@H官方釋出新的CORE才會有效果嗎
 
▌延伸閱讀