|
|
本帖最后由 金鹏 于 2013-7-6 12:10 编辑
由于是测试包，8900/8901稳定性存在一定问题，强烈建议超频的显卡在稳定跑7663的频率基础上降低20MHZ，以避免计算出错
by diwakar » Sat Jul 06, 2013 3:33 am
New Core17 Project 8901 released for beta testing. by diwakar » Sun Jun 23, 2013 5:41 am
Core 17 Project 8900 moved to advanced.
7月6日起，V7客户端加BETA参数可以接到8901内测包了
6月23日起加 -advmethods 参数GPU3_6.14R2客户端也能像V7加advanced参数一样接到CORE17内核8900包(需要填写PASSKEY），
不建议GTX480以下N卡和GCN架构以下A卡跑，PPD坑爹的很
（比如GTS450跑CORE15有14K+，跑CORE17只有10K+；GTX560TI跑CORE15包有29K,跑CORE17包只有25K）
适合GTX480/580/570和开普勒架构的N卡和HD7XXX的GCN架构A卡跑CORE17内核包
New Core17 Project 8901 released for beta testing
by diwakar » Sat Jul 06, 2013 3:37 am
This project will run on both ATI (HD 5000 or higher) and NVIDIA (Fermi or higher) clients.
Please let me know if there are any issues. I would like beta testers to report the GPU temperature as well because these WUs have ~75000 atoms as compared to ~45000 atoms in p8900.
Stats:
Points:7500 ; timeout:8.13 days; deadline:10.46 days;
Project Description:
The Src family protein tyrosine kinases are enzymes that play key roles in transducing cellular signals regulating cell growth, differentiation, proliferation, migration and survival. These enzymes are responsible for diseases such as cancer in which the cells undergo uncontrolled growth and proliferation. Crystallographic x-ray structures of human c-Src in the inactive and active conformation allow clear structural distinctions to be drawn between the inactive and active states. Those x-ray structures, though rich in information about the two end-points of the activation event, do not show how the activation occurs and how it might be regulated. Simulations and computational models, at different levels of approximation, can complement some of the missing information about Src and help address these important questions. Characterizing conformational transitions in large biomolecules such as Src is challenging, however, because the slow processes are not easily observed during simple unbiased molecular dynamics (MD) simulations. To circumvent those difficulties, previous studies of Src by our collaborators have used homology modeling to get the series of structures which show the structural changes involved in the activation process. In this project, we perform simulations of src kinase from the structures obtained from the already reported crystal structures.
by diwakar » Wed May 29, 2013 3:10 am
Project Description:
The Src family protein tyrosine kinases are enzymes that play key roles in transducing cellular signals regulating cell growth, differentiation, proliferation, migration and survival. These enzymes are responsible for diseases such as cancer in which the cells undergo uncontrolled growth and proliferation. Crystallographic x-ray structures of human c-Src in the inactive and active conformation allow clear structural distinctions to be drawn between the inactive and active states. Those x-ray structures, though rich in information about the two end-points of the activation event, do not show how the activation occurs and how it might be regulated. Simulations and computational models, at different levels of approximation, can complement some of the missing information about Src and help address these important questions. Characterizing conformational transitions in large biomolecules such as Src is challenging, however, because the slow processes are not easily observed during simple unbiased molecular dynamics (MD) simulations. To circumvent those difficulties, previous studies of Src by our collaborators have used homology modeling to get the series of structures which show the structural changes involved in the activation process. In this project, we perform simulations of src kinase from the structures obtained from the already reported crystal structures.
Stats:
Points:6000 ; timeout:6.5 days; deadline:8.8 days;
This project will run on both ATI (HD 5000 or higher) and NVIDIA (Fermi or higher) clients.
Please let me know if there are any issues.
by bollix47 » Wed May 29, 2013 1:18 pm
Windows Vista 64-bit running on an i7 940@stock ... dedicated(no other slots running)
HD 7970@1100Mhz
Catalyst 13.5beta
Only one assigned and completed so far ... Project: 8900 (Run 27, Clone 4, Gen 0)
HFM:
P8900:
Avg. Time / Frame : 00:03:53 - 110,068 PPD
For comparison P7663:
Avg. Time / Frame : 00:01:22 - 111,143 PPD - Project ID: 8900
- Core: ZETA
- Credit: 6000
- Frames: 100
- Name: PX-FAH-W Slot 01 -> GTX 660 Ti OC
- Path: PX-FAH-W-36330
- Number of Frames Observed: 300
- Min. Time / Frame : 00:05:06 - 73,132.69 PPD
- Avg. Time / Frame : 00:05:21 - 68,066.89 PPD
- Name: PX-FAH-W Slot 02 -> GTX 650 OC
- Path: PX-FAH-W-36330
- Number of Frames Observed: 300
- Min. Time / Frame : 00:14:08 - 15,852.59 PPD
- Avg. Time / Frame : 00:14:21 - 15,494.92 PPD
- Name: PX-Spaceship Slot 01 -> GTX 675M Stock
- Path: 127.0.0.1-36330
- Number of Frames Observed: 300
- Min. Time / Frame : 00:14:54 - 14,644.92 PPD
- Avg. Time / Frame : 00:15:03 - 14,426.56 PPD
- <P>
- <P>Name: PX-FAH-W Slot 01 -> GTX 660 Ti OC
- Path: PX-FAH-W-36330
- Number of Frames Observed: 300
- For comparison, here is data from P7663 (it used FahCore_17 version 0.0.44)
- </P>
- <P> </P>
- <P>Avg. Time / Frame : 00:02:06 - 58,350.77 PPD
- Cur. Time / Frame : 00:02:09 - 56,404.92 PPD
- R3F. Time / Frame : 00:02:07 - 57,602.98 PPD
- All Time / Frame : 00:02:06 - 58,218.13 PPD
- Eff. Time / Frame : 00:02:07 - 57,602.98 PPD
- Name: PX-FAH-W Slot 02 -> GTX 650 OC
- Path: PX-FAH-W-36330
- Number of Frames Observed: 300
- Avg. Time / Frame : 00:05:12 - 14,975.11 PPD
- Cur. Time / Frame : 00:05:08 - 15,169.01 PPD
- R3F. Time / Frame : 00:05:10 - 15,052.82 PPD
- All Time / Frame : 00:05:11 - 14,995.31 PPD
- Eff. Time / Frame : 00:05:14 - 14,825.10 PPD
- Name: PX-Spaceship Slot 01 -> GTX 675M Stock
- Path: 127.0.0.1-36330
- Number of Frames Observed: 300
- Avg. Time / Frame : 00:05:15 - 14,761.71 PPD
- Cur. Time / Frame : 00:05:13 - 14,886.20 PPD
- R3F. Time / Frame : 00:05:14 - 14,823.20 PPD
- All Time / Frame : 00:05:14 - 14,823.20 PPD
- Eff. Time / Frame : 00:05:15 - 14,760.70 PPD</P></P>
同样660TI跑8900的PPD要比跑7663高 而TITAN跑8900相比7663的PPD居然飚升20%
参见ynwowpop 兄弟的660TI跑8900达到了71.9K http://www.equn.com/forum/forum.php?mod=redirect&goto=findpost&ptid=37375&pid=502264
参见yuukosama 兄弟的TT草频后居然跑出了180K http://www.equn.com/forum/forum.php?mod=redirect&goto=findpost&ptid=37375&pid=502288
|
|
发表于 2013-5-29 16:35:29
发表于 2013-5-29 17:00:07
（需要PASSKEY）