找回密码
 新注册用户
搜索
楼主: refla

[讨论] 680 VS 7970

[复制链接]
发表于 2012-3-31 19:32:01 | 显示全部楼层
回复 14# ocw

【跑11293的PPD才6121
  1. *********************** Log Started 2012-01-18T16:09:36 ************************
  2. 16:09:36:************************* Folding@home Client *************************
  3. 16:09:36:      Website: http://folding.stanford.edu/
  4. 16:09:36:    Copyright: (c) 2009-2012 Stanford University
  5. 16:09:36:       Author: Joseph Coffland <joseph@cauldrondevelopment.com>
  6. 16:09:36:         Args:
  7. 16:09:36:       Config: C:/Users/MARK KOHLER/AppData/Roaming/FAHClient/config.xml
  8. 16:09:36:******************************** Build ********************************
  9. 16:09:36:      Version: 7.1.43
  10. 16:09:36:         Date: Jan 2 2012
  11. 16:09:36:         Time: 12:33:05
  12. 16:09:36:      SVN Rev: 3223
  13. 16:09:36:       Branch: fah/trunk/client
  14. 16:09:36:     Compiler: Intel(R) C++ MSVC 1500 mode 1200
  15. 16:09:36:      Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
  16. 16:09:36:               /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT
  17. 16:09:36:     Platform: win32 XP
  18. 16:09:36:         Bits: 32
  19. 16:09:36:         Mode: Release
  20. 16:09:36:******************************* System ********************************
  21. 16:09:36:          CPU: AMD Phenom(tm) II X4 965 Processor
  22. 16:09:36:       CPU ID: AuthenticAMD Family 16 Model 4 Stepping 3
  23. 16:09:36:         CPUs: 4
  24. 16:09:36:       Memory: 16.00GiB
  25. 16:09:36:  Free Memory: 13.61GiB
  26. 16:09:36:      Threads: WINDOWS_THREADS
  27. 16:09:36:   On Battery: false
  28. 16:09:36:   UTC offset: -6
  29. 16:09:36:          PID: 4636
  30. 16:09:36:          CWD: C:/Users/MARK KOHLER/AppData/Roaming/FAHClient
  31. 16:09:36:           OS: Windows 7 Professional
  32. 16:09:36:      OS Arch: AMD64
  33. 16:09:36:         GPUs: 1
  34. 16:09:36:        GPU 0: ATI:4 Tahiti XT [Radeon HD 7970]
  35. 16:09:36:         CUDA: Not detected
  36. 16:09:36:Win32 Service: false
  37. 16:09:36:***********************************************************************
  38. 16:09:36:<config>
  39. 16:09:36:  <!-- FahCore Control -->
  40. 16:09:36:  <checkpoint v='27'/>
  41. 16:09:36:  <core-priority v='low'/>
  42. 16:09:36:
  43. 16:09:36:  <!-- Network -->
  44. 16:09:36:  <proxy v=':8080'/>
  45. 16:09:36:
  46. 16:09:36:  <!-- User Information -->
  47. 16:09:36:  <passkey v='********************************'/>
  48. 16:09:36:  <team v='36837'/>
  49. 16:09:36:  <user v='mdk777'/>
  50. 16:09:36:
  51. 16:09:36:  <!-- Folding Slots -->
  52. 16:09:36:  <slot id='0' type='SMP'>
  53. 16:09:36:    <cpus v='-1'/>
  54. 16:09:36:  </slot>
  55. 16:09:36:  <slot id='1' type='GPU'>
  56. 16:09:36:    <client-type v='advanced'/>
  57. 16:09:36:  </slot>
  58. 16:09:36:</config>
  59. 16:09:36:Trying to access database...
  60. 16:09:36:Successfully acquired database lock
  61. 16:09:36:Enabled folding slot 00: READY smp:4
  62. 16:09:36:Enabled folding slot 01: READY gpu:0:"Tahiti XT [Radeon HD 7970]"
  63. 16:09:36:WU01:FS01:Starting
  64. 16:09:36:WU01:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" "C:/Users/MARK KOHLER/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/ATI/R600/Core_11.fah/FahCore_11.exe" -dir 01 -suffix 01 -version 701 -checkpoint 27 -gpu 0
  65. 16:09:36:Server connection id=1 on 0.0.0.0:36330 from 127.0.0.1
  66. 16:09:36:WU01:FS01:Started FahCore on PID 1216
  67. 16:09:36:WU01:FS01:Core PID:3516
  68. 16:09:36:WU01:FS01:FahCore 0x11 started
  69. 16:09:36:WU02:FS00:Starting
  70. 16:09:36:WU02:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" "C:/Users/MARK KOHLER/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/Core_a4.fah/FahCore_a4.exe" -dir 02 -suffix 01 -version 701 -checkpoint 27 -np 4
  71. 16:09:36:WU02:FS00:Started FahCore on PID 1080
  72. 16:09:36:WU02:FS00:Core PID:3564
  73. 16:09:36:WU02:FS00:FahCore 0xa4 started
  74. 16:09:36:WU01:FS01:0x11:
  75. 16:09:36:WU01:FS01:0x11:*------------------------------*
  76. 16:09:36:WU01:FS01:0x11:Folding@Home GPU Core - Beta
  77. 16:09:36:WU01:FS01:0x11:Version 1.24 (Mon Feb 9 11:00:12 PST 2009)
  78. 16:09:36:WU01:FS01:0x11:
  79. 16:09:36:WU01:FS01:0x11:Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
  80. 16:09:36:WU01:FS01:0x11:Build host: amoeba
  81. 16:09:36:WU01:FS01:0x11:Board Type: AMD
  82. 16:09:36:WU01:FS01:0x11:Core      :
  83. 16:09:36:WU01:FS01:0x11:Preparing to commence simulation
  84. 16:09:36:WU01:FS01:0x11:- Looking at optimizations...
  85. 16:09:36:WU01:FS01:0x11:- Files status OK
  86. 16:09:36:WU01:FS01:0x11:- Expanded 98711 -> 492188 (decompressed 498.6 percent)
  87. 16:09:36:WU01:FS01:0x11:Called DecompressByteArray: compressed_data_size=98711 data_size=492188, decompressed_data_size=492188 diff=0
  88. 16:09:36:WU01:FS01:0x11:- Digital signature verified
  89. 16:09:36:WU01:FS01:0x11:
  90. 16:09:36:WU01:FS01:0x11:Project: 5732 (Run 4, Clone 581, Gen 1076)
  91. 16:09:36:WU01:FS01:0x11:
  92. 16:09:36:WU01:FS01:0x11:Assembly optimizations on if available.
  93. 16:09:36:WU01:FS01:0x11:Entering M.D.
  94. 16:09:37:WU02:FS00:0xa4:
  95. 16:09:37:WU02:FS00:0xa4:*------------------------------*
  96. 16:09:37:WU02:FS00:0xa4:Folding@Home Gromacs GB Core
  97. 16:09:37:WU02:FS00:0xa4:Version 2.27 (Dec. 15, 2010)
  98. 16:09:37:WU02:FS00:0xa4:
  99. 16:09:37:WU02:FS00:0xa4:Preparing to commence simulation
  100. 16:09:37:WU02:FS00:0xa4:- Looking at optimizations...
  101. 16:09:37:WU02:FS00:0xa4:- Files status OK
  102. 16:09:37:WU02:FS00:0xa4:- Expanded 663812 -> 1297016 (decompressed 195.3 percent)
  103. 16:09:37:WU02:FS00:0xa4:Called DecompressByteArray: compressed_data_size=663812 data_size=1297016, decompressed_data_size=1297016 diff=0
  104. 16:09:37:WU02:FS00:0xa4:- Digital signature verified
  105. 16:09:37:WU02:FS00:0xa4:
  106. 16:09:37:WU02:FS00:0xa4:Project: 7200 (Run 31, Clone 25, Gen 57)
  107. 16:09:37:WU02:FS00:0xa4:
  108. 16:09:37:WU02:FS00:0xa4:Assembly optimizations on if available.
  109. 16:09:37:WU02:FS00:0xa4:Entering M.D.
  110. 16:09:42:WU02:FS00:0xa4:Using Gromacs checkpoints
  111. 16:09:42:WU01:FS01:0x11:Tpr hash 01/wudata_01.tpr:  2080191161 3749261823 3162513416 554572999 1027544223
  112. 16:09:42:WU02:FS00:0xa4:Mapping NT from 4 to 4
  113. 16:09:43:WU01:FS01:0x11:Working on Protein
  114. 16:09:43:WU02:FS00:0xa4:Resuming from checkpoint
  115. 16:09:43:WU02:FS00:0xa4:Verified 02/wudata_01.log
  116. 16:09:43:WU02:FS00:0xa4:Verified 02/wudata_01.trr
  117. 16:09:43:WU02:FS00:0xa4:Verified 02/wudata_01.xtc
  118. 16:09:43:WU02:FS00:0xa4:Verified 02/wudata_01.edr
  119. 16:09:43:WU01:FS01:0x11:Client config unavailable.
  120. 16:09:43:WU02:FS00:0xa4:Completed 1996 out of 750000 steps  (0%)
  121. 16:09:43:WU01:FS01:0x11:Starting GUI Server
  122. 16:10:17:Server connection id=1 ended
  123. 16:11:40:Server connection id=2 on 0.0.0.0:36330 from 127.0.0.1
  124. 16:14:21:WU01:FS01:FahCore returned: FAILED_2 (1 = 0x1)
  125. 16:14:21:WU01:FS01:Sending unit results: id:01 state:SEND error:FAILED project:5732 run:4 clone:581 gen:1076 core:0x11 unit:0x223c9b844f0ee43d0434024500041664
  126. 16:14:21:WARNING:WU01:FS01:Work server too old for fail report, dumping
  127. 16:14:21:WU01:FS01:Cleaning up
  128. 16:14:21:WU00:FS01:Connecting to assign-GPU.stanford.edu:80
  129. 16:14:21:WU00:FS01:News: Welcome to Folding@Home
  130. 16:14:21:WU00:FS01:Assigned to work server 171.67.108.44
  131. 16:14:21:WU00:FS01:Requesting new work unit for slot 01: READY gpu:0:"Tahiti XT [Radeon HD 7970]" from 171.67.108.44
  132. 16:14:21:WU00:FS01:Connecting to 171.67.108.44:8080
  133. 16:14:22:WU00:FS01:Downloading 44.30KiB
  134. 16:14:22:WU00:FS01:Download complete
  135. 16:14:22:WU00:FS01:Received Unit: id:00 state:DOWNLOAD error:OK project:11293 run:25 clone:332 gen:1 core:0x16 unit:0x000000016652edbc4d94ba1657b365a3
  136. 16:14:22:WU00:FS01:Starting
  137. 16:14:22:WU00:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" "C:/Users/MARK KOHLER/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/ATI/R600/Core_16.fah/FahCore_16.exe" -dir 00 -suffix 01 -version 701 -checkpoint 27 -gpu 0
  138. 16:14:22:WU00:FS01:Started FahCore on PID 3588
  139. 16:14:22:WU00:FS01:Core PID:1884
  140. 16:14:22:WU00:FS01:FahCore 0x16 started
  141. 16:14:23:WU00:FS01:0x16:
  142. 16:14:23:WU00:FS01:0x16:*------------------------------*
  143. 16:14:23:WU00:FS01:0x16:Folding@Home GPU Core
  144. 16:14:23:WU00:FS01:0x16:Version 2.11 (Thu Dec 9 15:00:14 PST 2010)
  145. 16:14:23:WU00:FS01:0x16:
  146. 16:14:23:WU00:FS01:0x16:Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.30729.01 for 80x86
  147. 16:14:23:WU00:FS01:0x16:Build host: user-f6d030f24f
  148. 16:14:23:WU00:FS01:0x16:Board Type: AMD/OpenCL
  149. 16:14:23:WU00:FS01:0x16:Core      : x=16
  150. 16:14:23:WU00:FS01:0x16: Window's signal control handler registered.
  151. 16:14:23:WU00:FS01:0x16:Preparing to commence simulation
  152. 16:14:23:WU00:FS01:0x16:- Looking at optimizations...
  153. 16:14:23:WU00:FS01:0x16:DeleteFrameFiles: successfully deleted file=00/wudata_01.ckp
  154. 16:14:23:WU00:FS01:0x16:- Created dyn
  155. 16:14:23:WU00:FS01:0x16:- Files status OK
  156. 16:14:23:WU00:FS01:0x16:sizeof(CORE_PACKET_HDR) = 512 file=<>
  157. 16:14:23:WU00:FS01:0x16:- Expanded 44854 -> 171163 (decompressed 381.6 percent)
  158. 16:14:23:WU00:FS01:0x16:Called DecompressByteArray: compressed_data_size=44854 data_size=171163, decompressed_data_size=171163 diff=0
  159. 16:14:23:WU00:FS01:0x16:- Digital signature verified
  160. 16:14:23:WU00:FS01:0x16:
  161. 16:14:23:WU00:FS01:0x16:Project: 11293 (Run 25, Clone 332, Gen 1)
  162. 16:14:23:WU00:FS01:0x16:
  163. 16:14:23:WU00:FS01:0x16:Assembly optimizations on if available.
  164. 16:14:23:WU00:FS01:0x16:Entering M.D.
  165. 16:14:24:WU00:FS01:0x16:Tpr hash 00/wudata_01.tpr:  3297842355 2472556598 1446642422 3227500222 1197824835
  166. 16:14:24:WU00:FS01:0x16:Working on ALZHEIMER DISEASE AMYLOID
  167. 16:14:24:WU00:FS01:0x16:Client config unavailable.
  168. 16:14:24:WU00:FS01:0x16:Starting GUI Server
  169. 16:14:28:WU00:FS01:0x16:Setting checkpoint frequency: 500000
  170. 16:14:28:WU00:FS01:0x16:Completed         3 out of 50000000 steps (0%).
  171. 16:18:24:WU00:FS01:0x16:Completed    500000 out of 50000000 steps (1%).
  172. 16:22:48:WU00:FS01:0x16:Completed   1000000 out of 50000000 steps (2%).
  173. 16:27:06:WU00:FS01:0x16:Completed   1500000 out of 50000000 steps (3%).
  174. 16:31:25:WU00:FS01:0x16:Completed   2000000 out of 50000000 steps (4%).
  175. 16:35:43:WU00:FS01:0x16:Completed   2500000 out of 50000000 steps (5%).
  176. 16:40:02:WU00:FS01:0x16:Completed   3000000 out of 50000000 steps (6%).
  177. 16:44:20:WU00:FS01:0x16:Completed   3500000 out of 50000000 steps (7%).
  178. 16:45:03:WU02:FS00:0xa4:Completed 7500 out of 750000 steps  (1%)
  179. 16:48:38:WU00:FS01:0x16:Completed   4000000 out of 50000000 steps (8%).
  180. 16:52:57:WU00:FS01:0x16:Completed   4500000 out of 50000000 steps (9%).
  181. 16:57:15:WU00:FS01:0x16:Completed   5000000 out of 50000000 steps (10%).
  182. 17:01:33:WU00:FS01:0x16:Completed   5500000 out of 50000000 steps (11%).
  183. 17:05:52:WU00:FS01:0x16:Completed   6000000 out of 50000000 steps (12%).
  184. 17:10:10:WU00:FS01:0x16:Completed   6500000 out of 50000000 steps (13%).
  185. 17:14:29:WU00:FS01:0x16:Completed   7000000 out of 50000000 steps (14%).
  186. 17:18:47:WU00:FS01:0x16:Completed   7500000 out of 50000000 steps (15%).
复制代码
回复

使用道具 举报

发表于 2012-4-5 20:17:06 | 显示全部楼层
本帖最后由 fevaoctwh 于 2012-4-5 20:19 编辑

还是要等斯坦福更新GPU客户端了,虽然运算绝对性能远不如7970和580,但也不会那么低
毕竟架构整个换了,优化好了580 70%的PPD应该还是有的吧
回复

使用道具 举报

发表于 2012-4-6 07:49:10 | 显示全部楼层
还是要等斯坦福更新GPU客户端了,虽然运算绝对性能远不如7970和580,但也不会那么低
毕竟架构整个换了,优 ...
fevaoctwh 发表于 2012-4-5 20:17


这个真不好说,从目前的所有底层通用计算测试结果来看都比不上580,与GK104大幅删减通用计算模块的架构改动是相符的,

下半年的GK110的通用计算性能才是值得期待的
回复

使用道具 举报

发表于 2012-4-6 12:09:47 | 显示全部楼层
我预计GK110的构架和GK104相比不会有太大变化,只会再增加一些shader数以及增加显存带宽。
双精度方面GK110会有大幅增强,但仅限于tesla版本,游戏卡的双精度性能估计和GK104类似。

回复 18# 金鹏
回复

使用道具 举报

发表于 2012-4-6 13:44:10 | 显示全部楼层
回复 19# cuda

tesla版本也是基于GK110核心吧?
回复

使用道具 举报

发表于 2012-4-6 15:33:26 | 显示全部楼层
估计是的。这里有篇文章提到GK110显存位宽升到512bit, 具有1.3-1.5TFlops的双精度能力,以及首批的GK110多数用于专业卡。是否靠谱未知。
http://www.cnbeta.com/articles/180881.htm

回复 20# 金鹏
回复

使用道具 举报

发表于 2012-4-6 21:06:06 | 显示全部楼层
回复 19# cuda


    我个人认为,GK110的SMX中,Warp Scheduler应该与每一组Block都像GF110那样硬件上做到一一对应,而不是像GK104那样通过驱动程序去分发Warp线程。还有就是缓存可能会暴力扩充,这样一来通用计算会大幅度提高。强大的游戏性能+强大的通用计算性能+大核心才是NVIDIA的旗舰啊,历代都是这样的。
回复

使用道具 举报

发表于 2012-4-11 21:16:31 | 显示全部楼层
从评测结果看,gtx680在双精度性能、显存带宽、整数性能方面都弱于gtx580,但单精度浮点性能很强。
就拿gtx680的gpc测试结果来说,其实也是很不错的,单精度浮点分数比gtx580高了50%以上。其他一些软件的测试结果基本也是一致的。例如nvidia cuda sdk的nbody测试结果:gtx580大约800gflops, gtx680则有1200多gflops。
跑fah只要单精度性能好就可以了,其他指标可以忽略。我觉得gtx680如果优化好了正常的ppd应该在3万左右。

这PPD结果与GPC测试比较基本上吻合,对680跑FAH不报高期望
金鹏 发表于 2012-3-31 18:20
回复

使用道具 举报

发表于 2012-4-17 08:47:31 | 显示全部楼层
支支持楼主发力
回复

使用道具 举报

发表于 2012-4-17 09:16:38 | 显示全部楼层
果断坑爹啊,果断不能等680,看那个CUDA的数量,纳闷现在N是不是也开始走流处理数量很吓人,效果却一般的路线了
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 新注册用户

本版积分规则

论坛官方淘宝店开业啦~

Archiver|手机版|小黑屋|中国分布式计算总站 ( 沪ICP备05042587号 )

GMT+8, 2024-5-6 04:13

Powered by Discuz! X3.5

© 2001-2024 Discuz! Team.

快速回复 返回顶部 返回列表