那不就跟我之前的 PPD drop 很像？延伸桌面設定正確、顯示接頭沒問題，但就是兩張卡的效能等於一張單卡。我在想如果一開始先插上第一張卡，再灌軟體。確定可以穩跑後在插上第二張是否會正常？



我的是雙卡就出問題了。NV 180.70 驅動程式、Intel Q9550

事實上CPU...挺鳥的 (XP90是好物 )
目前是用3800+X2去推整個牧場(想找個時機換C/P佳的肥龍II)
使用率約40~50%吧...也沒有推不動的情況
況且驅動如果換成180.XX...CPU 負載會更低...

插三張卡都很常...但只要任一槽插上第四張
整個牧場就會歡樂起來... (櫻花妹大戰->戰神妹->PK雙狐妹)
果然太花心就是這樣的下場...








<--(這傢伙到底在講什啊)

目前A卡我沒在用, 但以N卡現今的驅動來說, 已經把CPU使用率降到極低了.其實CPU已不重要.

現在我的Q6600@3.2Ghz 幾乎都是全速在跑WCG. 從換180.70後,近日已經再前進30名了.

WCG世排450後要進30名可是不太容易的事情(至少16cores+),多虧CPU被釋放出來.

他說4P是極限了

一次不能搞太多妹!

剛才無聊,用Furmark顯示卡燒機程式1.5版來跑,在室溫20度C的室內,8800GT 512顯示卡在5分鐘後就上升到最高溫72度.跑一般384點數或480點數的Project,大約在65-66度.
至於511點數的57系列雷U,運作時的顯卡溫度大約70度,己經可以和燒機程式稱兄道弟了.
所 以說N社的卡來跑FAH,幾乎己經將其架構的浮點性能都發揮出來了.不知A卡的情況,
跑GPU Client時的溫度和燒機軟體比較是如何呢?月底消費券到手想入手A卡4850,想知道
其情況如何,有沒有網兄能心得分享?謝謝.

理論上DirectX跟OpenGL一直都可以撰寫一些讓GPU做原本應該是CPU的工作
只是DX跟OpenGL本來就不是這方面的專門，簡單的說是可程式化度不夠
而且也不是專門針對這功能做的設計，在開發上可能也比較困難
雖然速度上還是比CPU快，但是效率無法完全發揮
NV的CUDA可以用C語言開發能夠以低階層的方式存取GPU最大資源的APi
有興趣可以翻翻這篇GPGPU在GT200上的介紹
http://www.realworldtech.com/page.c...RWT090808195242

[09:05:38] Writing local files
[09:05:38] Completed 122500 out of 250000 steps (49 percent)
[09:08:39] Timered checkpoint triggered.
[09:11:40] Timered checkpoint triggered.
[09:12:38] CoreStatus = 63 (99)
[09:12:38] + Error starting Folding@Home core.
[09:12:43]
[09:12:43] + Processing work unit
[09:12:43] Work type a1 not eligible for variable processors
[09:12:43] Core required: FahCore_a1.exe
[09:12:43] Core found.
[09:12:43] Working on queue slot 08 [January 3 09:12:43 UTC]
[09:12:43] + Working ...
[09:12:43] - Calling 'mpiexec -np 4 -channel shm -env MPICH_USE_SMP_OPTIMIZATION
S 1 -host 127.0.0.1 FahCore_a1.exe -dir work/ -suffix 08 -nocpulock -checkpoint
3 -forceasm -verbose -lifeline 2648 -version 622'

[09:12:45]
[09:12:45] *------------------------------*
[09:12:45] Folding@Home Gromacs SMP Core
[09:12:45] Version 1.76 (February 23, 2008)
[09:12:45]
[09:12:45] Preparing to commence simulation
[09:12:45] - Ensuring status. Please wait.
[09:13:02] - Assembly optimizations manually forced on.
[09:13:02] - Not checking prior termination.
[09:13:14] - Expanded 4740020 -> 24426905 (decompressed 515.3 percent)
[09:13:15]
[09:13:15] Project: 2665 (Run 3, Clone 677, Gen 82)
[09:13:15]
[09:13:16] Assembly optimizations on if available.
[09:13:16] Entering M.D.
[09:13:23] Calling FAH init
[09:13:24] ater
[09:13:24] Writing local files
[09:13:24] Completed 123106 out of 250000 steps (49 percent)
[09:13:24] er
[09:13:24] Writing local files
[09:13:24] Completed 123106 out of 250000 steps (49 percent)
[09:13:30] Extra SSE boost OK.
[09:16:25] Timered checkpoint triggered.


各位能幫我解答一下這是什麼原因嗎??
前面幾次發言都沒人回答

有時其他的原因也會造成這個結果.即使上一個WU正常完成,等接到下一個WU要運算時,
就會出現這一行 : Error starting Folding@Home core.
重新啓動,也一樣;刪除core,或者Work也一樣.因為在下同時執行多個SMP,遇到這種情形,
就把其他主機正常跑的移到BT機來執行.我的跑的核心是1.9版,似乎比你跑的新,僅供參考.

07:43:44] - Ask before connecting: No
[07:43:44] - User name: Ruark8263 (Team 31403)
[07:43:44] - User ID: 4C71D40A3FC01129
[07:43:44] - Machine ID: 16
[07:43:44]
[07:43:44] Loaded queue successfully.
[07:43:44]
[07:43:44] + Processing work unit
[07:43:44] Core required: FahCore_78.exe
[07:43:44] Core found.
[07:43:44] Working on queue slot 08 [January 3 07:43:44 UTC]
[07:43:44] + Working ...
[07:43:44]
[07:43:44] *------------------------------*
[07:43:44] Folding@Home Gromacs Core
[07:43:44] Version 1.90 (March 8, 2006)
[07:43:44]
[07:43:44] Preparing to commence simulation
[07:43:44] - Looking at optimizations...
[07:43:44] - Files status OK
[07:43:44] - Expanded 238119 -> 1167461 (decompressed 490.2 percent)
[07:43:44]
[07:43:44] Project: 4445 (Run 669, Clone 0, Gen 4)
[07:43:44]
[07:43:44] Assembly optimizations on if available.
[07:43:44] Entering M.D.
[07:44:05] (Starting from checkpoint)
[07:44:05] Protein: p4445_Seq44_Amber03
[07:44:05]
[07:44:05] Writing local files
[07:45:32] Completed 510000 out of 1500000 steps (34%)
[07:45:32] Extra SSE boost OK.
[07:56:18] Writing local files
[07:56:18] Completed 525000 out of 1500000 steps (35%)
[08:07:05] Writing local files
[08:07:05] Completed 540000 out of 1500000 steps (36%)