76xx on gt 430

[skylineking88]

well i am folding on my gt 430 and with wu 7642/43/44 it fails the wu at 1 to 2 hrs and now this is the only wu it gets it has been liike this for 2 day


and can i get help on how to add a img file

[GreyWhiskers]

Welcome to the fold.

Can you post the log from your run? Windows instructions here. If you are running v7, you can select the LOG tab, then in the text part, do a CTRL-A (select all). Then select the "COPY" button.

In the Full Editor in the Forum, choose the "CODE" button on top of the text entry box, then paste the log you just copied between the "

Code: Select all

" brackets.

re: images. Here's what I do. There are other ways, but this works for me.

in Win XP, I hit the Pr Scr button on the keyboard, then paste into a paint type editor, either the Windows PAINT or a better utility paint.net. Then, I select the part of the screen shot I want, paste it into a separate file, and save to whatever name you want.

this part is easier in Vista or Win 7, since you can use the built-in Snipping Tool utility to make a .jpg file of the part of the screen you want.



[above image is kind of surreal - I did the exact steps to generate THAT image as in the instructions below]
I then use Google's Picasa to upload the image to the Picasa Web (you need a Google login to do this). Within Picasa web:
- select "link to this photo"
- select size (I always select the largest available)
- click "image only (no link)
- then when you mouse over the "embed image" box, it will be highlighted. Just do a CTRL-C to copy the link
- Now, back in the forum post you are creating. click on the Img button above the text entry box. paste the URL you just copied in Picasa Web between the "img - /img" brackets.

and, Shazzam, your image will be part of your post.

[MeekMark]

I'm getting the same problem on 7641 Work Units, GeoForce GT 430, NVIDIA driver version 296.10. Also had same issue with an older version of NVIDIA driver. I have it overclocked to Core Clock: 825, Memory Clock 700. I usually had Core Clock at 835 for a long time with no issues. It usually dies between frames 55 and 62, similar like this:

I haven't yet completed a 7641 WU on this PC. (Windows Vista Ultimate, 64 bit, Intel Core2 Quad CPU Q9650 @ 3 GHz.)

Code: Select all

[07:20:38] Completed   1550000 out of 2500000 steps (62%).
[07:20:38] mdrun_gpu returned 52
[07:20:38] NANs detected on GPU
[07:20:38] 
[07:20:38] Folding@home Core Shutdown: UNSTABLE_MACHINE
[07:20:41] CoreStatus = 7A (122)
[07:20:41] Sending work to server

Here's a full snippet of a run; this is from continuing from a checkpoint; but I get same results starting fresh too:

Code: Select all

--- Opening Log file [May 11 02:46:52 UTC] 


# Windows GPU Systray Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.40r1

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Folding\GPU_6_40


[02:46:52] - Ask before connecting: No
[02:46:52] - User name: MarkyMark7 (Team 1003)
[02:46:52] - User ID: xxxxx
[02:46:52] - Machine ID: 16
[02:46:52] 
[02:46:52] Gpu type=3 species=21.
[02:46:53] Loaded queue successfully.
[02:46:53] Initialization complete
[02:46:53] 
[02:46:53] + Processing work unit
[02:46:53] Core required: FahCore_15.exe
[02:46:53] Core found.
[02:46:53] Working on queue slot 02 [May 11 02:46:53 UTC]
[02:46:53] + Working ...
[02:46:53] 
[02:46:53] *------------------------------*
[02:46:53] Folding@Home GPU Core
[02:46:53] Version                2.22 (Thu Dec 8 17:08:05 PST 2011)
[02:46:53] Build host             SimbiosNvdWin7
[02:46:53] Board Type             NVIDIA/CUDA
[02:46:53] Core                   15
[02:46:53] 
[02:46:53] Window's signal control handler registered.
[02:46:53] Preparing to commence simulation
[02:46:53] - Looking at optimizations...
[02:46:53] - Files status OK
[02:46:53] sizeof(CORE_PACKET_HDR) = 512 file=<>
[02:46:53] - Expanded 548987 -> 973992 (decompressed 177.4 percent)
[02:46:53] Called DecompressByteArray: compressed_data_size=548987 data_size=973992, decompressed_data_size=973992 diff=0
[02:46:53] - Digital signature verified
[02:46:53] 
[02:46:53] Project: 7641 (Run 336, Clone 0, Gen 14)
[02:46:53] 
[02:46:53] Assembly optimizations on if available.
[02:46:53] Entering M.D.
[02:46:55] Will resume from checkpoint file work/wudata_02.ckp
[02:46:55] Tpr hash work/wudata_02.tpr:  4156174597 3316335987 1496518870 3710207025 1534817559
[02:46:55] GPU device info: vendor=0 device=0 name=<NA> match=0
[02:46:55] Working on Protein in water
[02:46:55] Client config found, loading data.
[02:46:55] Starting GUI Server
[02:48:50] Resuming from checkpoint
[02:48:50] fcCheckPointResume: retreived and current tpr file hash:
[02:48:50]    0   4156174597   4156174597
[02:48:50]    1   3316335987   3316335987
[02:48:50]    2   1496518870   1496518870
[02:48:50]    3   3710207025   3710207025
[02:48:50]    4   1534817559   1534817559
[02:48:50] fcCheckPointResume: file hashes same.
[02:48:50] fcCheckPointResume: state restored.
[02:48:50] fcCheckPointResume: name work/wudata_02.log Verified work/wudata_02.log
[02:48:50] fcCheckPointResume: name work/wudata_02.trr Verified work/wudata_02.trr
[02:48:50] fcCheckPointResume: name work/wudata_02.xtc Verified work/wudata_02.xtc
[02:48:50] fcCheckPointResume: name work/wudata_02.edr Verified work/wudata_02.edr
[02:48:50] fcCheckPointResume: state restored 2
[02:48:50] Resumed from checkpoint
[02:48:50] Setting checkpoint frequency: 25000
[02:48:50] Completed    775001 out of 2500000 steps (31%).
[03:26:50] Completed    800000 out of 2500000 steps (32%).
[04:04:26] Completed    825000 out of 2500000 steps (33%).
[04:41:37] Completed    850000 out of 2500000 steps (34%).
[05:18:51] Completed    875000 out of 2500000 steps (35%).
[05:55:40] Completed    900000 out of 2500000 steps (36%).
[06:32:46] Completed    925000 out of 2500000 steps (37%).
[07:10:51] Completed    950000 out of 2500000 steps (38%).
[07:49:23] Completed    975000 out of 2500000 steps (39%).
[08:26:48] Completed   1000000 out of 2500000 steps (40%).
[08:46:52] + Working...
[09:03:49] Completed   1025000 out of 2500000 steps (41%).
[09:40:45] Completed   1050000 out of 2500000 steps (42%).
[10:17:44] Completed   1075000 out of 2500000 steps (43%).
[10:54:44] Completed   1100000 out of 2500000 steps (44%).
[11:31:34] Completed   1125000 out of 2500000 steps (45%).
[12:51:29] Completed   1150000 out of 2500000 steps (46%).
[12:51:29] mdrun_gpu returned 52
[12:51:29] NANs detected on GPU
[12:51:29] 
[12:51:29] Folding@home Core Shutdown: UNSTABLE_MACHINE
[12:51:33] CoreStatus = 7A (122)
[12:51:33] Sending work to server
[12:51:33] Project: 7641 (Run 336, Clone 0, Gen 14)
[12:51:33] - Read packet limit of 540015616... Set to 524286976.
[12:51:33] - Error: Could not get length of results file work/wuresults_02.dat
[12:51:33] - Error: Could not read unit 02 file. Removing from queue.
[12:51:33] - Preparing to get new work unit...
[12:51:33] Cleaning up work directory
[12:51:33] + Attempting to get work packet
[12:51:33] Passkey found
[12:51:33] Gpu type=3 species=21.

[7im]

Hello MeekMark, welcome to the forum.

Please note that 76xx GPU work units are much larger than previous projects. What was once a stable overclock is unlikely to remain that way as work units continue to grow in size.

[MeekMark]

Great; I'll keep throttling it down until it works.

[HaloJones]

Don't hold your breath. Seems like some 4xx cards just struggle to fold 76xx units no matter what you clock them to, yet work perfectly for 8xxx units.

[Nathan_P]

Don't hold your breath. Seems like some 4xx cards just struggle to fold 76xx units no matter what you clock them to, yet work perfectly for 8xxx units.

Perhaps we are running into the "card not powerful enough to do the project" again. We had a similar issue with the old 66xx GPU projects about 18months ago, Low end 8xxx and 9xxx series cards didn't have enough grunt to run the WU properly and would fail - no matter what clocks they were set to.

[7im]

Slightly differrent issue, but may have same result.

[DrSpalding]

I am also having problems with Project: 7640 (Run 76, Clone 0, Gen 14) but on a GTS450. It gets a "NANs detected on GPU" error from mdrun_gpu returning 52. This results in an "UNSTABLE MACHINE" error being thrown ultimately. The card is a completely stock MSI GTS450, running at stock clocks. Here is the log:

Code: Select all

*********************** Log Started 2012-05-14T05:05:37Z ***********************
05:05:37:************************* Folding@home Client *************************
05:05:37:      Website: http://folding.stanford.edu/
05:05:37:    Copyright: (c) 2009-2012 Stanford University
05:05:37:       Author: Joseph Coffland <joseph@cauldrondevelopment.com>
05:05:37:         Args: --lifeline 3428 --command-port=36330
05:05:37:       Config: c:/fah/Data/config.xml
05:05:37:******************************** Build ********************************
05:05:37:      Version: 7.1.52
05:05:37:         Date: Mar 20 2012
05:05:37:         Time: 19:37:42
05:05:37:      SVN Rev: 3515
05:05:37:       Branch: fah/trunk/client
05:05:37:     Compiler: Intel(R) C++ MSVC 1500 mode 1200
05:05:37:      Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
05:05:37:               /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT
05:05:37:     Platform: win32 XP
05:05:37:         Bits: 32
05:05:37:         Mode: Release
05:05:37:******************************* System ********************************
05:05:37:          CPU: Intel(R) Pentium(R) 4 CPU 3.20GHz
05:05:37:       CPU ID: GenuineIntel Family 15 Model 4 Stepping 3
05:05:37:         CPUs: 2
05:05:37:       Memory: 3.00GiB
05:05:37:  Free Memory: 1.66GiB
05:05:37:      Threads: WINDOWS_THREADS
05:05:37:   On Battery: false
05:05:37:   UTC offset: -7
05:05:37:          PID: 2080
05:05:37:          CWD: c:/fah/Data
05:05:37:           OS: Windows 7 Professional
05:05:37:      OS Arch: AMD64
05:05:37:         GPUs: 1
05:05:37:        GPU 0: FERMI:1 GF106 [GeForce GTS 450]
05:05:37:         CUDA: 2.1
05:05:37:  CUDA Driver: 4010
05:05:37:Win32 Service: false
05:05:37:***********************************************************************
05:05:37:<config>
05:05:37:  <!-- Folding Slot Configuration -->
05:05:37:  <gpu v='true'/>
05:05:37:  <smp v='false'/>
05:05:37:
05:05:37:  <!-- Network -->
05:05:37:  <proxy v=':8080'/>
05:05:37:
05:05:37:  <!-- Remote Command Server -->
05:05:37:
05:05:37:  <!-- User Information -->
05:05:37:  <user v='DrSpalding'/>
05:05:37:
05:05:37:  <!-- Folding Slots -->
05:05:37:  <slot id='0' type='GPU'>
05:05:37:    <max-packet-size v='big'/>
05:05:37:  </slot>
05:05:37:</config>
05:05:37:Trying to access database...
05:05:37:Successfully acquired database lock
05:05:37:Enabled folding slot 00: READY gpu:0:\"GF106 [GeForce GTS 450]\"
05:05:37:WU01:FS00:Starting
05:05:37:WU01:FS00:Running FahCore: C:\\FAH\\Program/FAHCoreWrapper.exe c:/fah/Data/cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_15.fah/FahCore_15.exe -dir 01 -suffix 01 -version 701 -lifeline 2080 -checkpoint 15 -gpu 0
05:05:38:Server connection id=1 on 0.0.0.0:36330 from 127.0.0.1
05:05:38:WU01:FS00:Started FahCore on PID 4192
05:05:38:WU01:FS00:Core PID:4204
05:05:38:WU01:FS00:FahCore 0x15 started
05:05:39:WU01:FS00:0x15:
05:05:39:WU01:FS00:0x15:*------------------------------*
05:05:39:WU01:FS00:0x15:Folding@Home GPU Core
05:05:39:WU01:FS00:0x15:Version                2.22 (Thu Dec 8 17:08:05 PST 2011)
05:05:39:WU01:FS00:0x15:Build host             SimbiosNvdWin7
05:05:39:WU01:FS00:0x15:Board Type             NVIDIA/CUDA
05:05:39:WU01:FS00:0x15:Core                   15
05:05:39:WU01:FS00:0x15:
05:05:39:WU01:FS00:0x15:Window's signal control handler registered.
05:05:39:WU01:FS00:0x15:Preparing to commence simulation
05:05:39:WU01:FS00:0x15:- Looking at optimizations...
05:05:39:WU01:FS00:0x15:- Files status OK
05:05:39:WU01:FS00:0x15:sizeof(CORE_PACKET_HDR) = 512 file=<>
05:05:39:WU01:FS00:0x15:- Expanded 548545 -> 890896 (decompressed 162.4 percent)
05:05:39:WU01:FS00:0x15:Called DecompressByteArray: compressed_data_size=548545 data_size=890896, decompressed_data_size=890896 diff=0
05:05:39:WU01:FS00:0x15:- Digital signature verified
05:05:39:WU01:FS00:0x15:
05:05:39:WU01:FS00:0x15:Project: 7640 (Run 76, Clone 0, Gen 14)
05:05:39:WU01:FS00:0x15:
05:05:39:WU01:FS00:0x15:Assembly optimizations on if available.
05:05:39:WU01:FS00:0x15:Entering M.D.
05:05:41:WU01:FS00:0x15:Will resume from checkpoint file 01/wudata_01.ckp
05:05:41:WU01:FS00:0x15:Tpr hash 01/wudata_01.tpr:  1726458347 522663929 1396732719 3771082946 1308202816
05:05:41:WU01:FS00:0x15:GPU device info: vendor=0 device=0 name=<NA> match=0
05:05:41:WU01:FS00:0x15:Working on Generated by trjconv : 21521 p3064_lambda5_2003 t= 5500.00049 in water
05:05:41:WU01:FS00:0x15:Client config unavailable.
05:05:41:WU01:FS00:0x15:Starting GUI Server
05:05:42:Server connection id=2 on 0.0.0.0:36330 from 192.168.1.21
05:05:44:Server connection id=3 on 0.0.0.0:36330 from 192.168.1.20
05:08:21:WU01:FS00:0x15:Resuming from checkpoint
05:08:22:WU01:FS00:0x15:fcCheckPointResume: retreived and current tpr file hash:
05:08:22:WARNING:WU01:FS00:Detected clock skew, adjusting time estimates
05:08:22:WU01:FS00:0x15:   0   1726458347   1726458347
05:08:22:WU01:FS00:0x15:   1    522663929    522663929
05:08:22:WU01:FS00:0x15:   2   1396732719   1396732719
05:08:22:WU01:FS00:0x15:   3   3771082946   3771082946
05:08:22:WU01:FS00:0x15:   4   1308202816   1308202816
05:08:22:WU01:FS00:0x15:fcCheckPointResume: file hashes same.
05:08:22:WU01:FS00:0x15:fcCheckPointResume: state restored.
05:08:22:WU01:FS00:0x15:fcCheckPointResume: name 01/wudata_01.log Verified 01/wudata_01.log
05:08:22:WU01:FS00:0x15:fcCheckPointResume: name 01/wudata_01.trr Verified 01/wudata_01.trr
05:08:22:WU01:FS00:0x15:fcCheckPointResume: name 01/wudata_01.xtc Verified 01/wudata_01.xtc
05:08:22:WU01:FS00:0x15:fcCheckPointResume: name 01/wudata_01.edr Verified 01/wudata_01.edr
05:08:22:WU01:FS00:0x15:fcCheckPointResume: state restored 2
05:08:22:WU01:FS00:0x15:Resumed from checkpoint
05:08:22:WU01:FS00:0x15:Setting checkpoint frequency: 25000
05:08:22:WU01:FS00:0x15:Completed    900001 out of 2500000 steps (36%).
05:15:59:Server connection id=4 on 0.0.0.0:36330 from 192.168.1.20
05:27:56:WU01:FS00:0x15:Completed    925000 out of 2500000 steps (37%).
05:47:15:WU01:FS00:0x15:Completed    950000 out of 2500000 steps (38%).
06:06:28:WU01:FS00:0x15:Completed    975000 out of 2500000 steps (39%).
06:25:42:WU01:FS00:0x15:Completed   1000000 out of 2500000 steps (40%).
06:44:55:WU01:FS00:0x15:Completed   1025000 out of 2500000 steps (41%).
07:04:15:WU01:FS00:0x15:Completed   1050000 out of 2500000 steps (42%).
07:23:26:WU01:FS00:0x15:Completed   1075000 out of 2500000 steps (43%).
07:42:44:WU01:FS00:0x15:Completed   1100000 out of 2500000 steps (44%).
08:02:02:WU01:FS00:0x15:Completed   1125000 out of 2500000 steps (45%).
08:21:23:WU01:FS00:0x15:Completed   1150000 out of 2500000 steps (46%).
08:40:48:WU01:FS00:0x15:Completed   1175000 out of 2500000 steps (47%).
09:05:53:WU01:FS00:0x15:Completed   1200000 out of 2500000 steps (48%).
09:05:54:WU01:FS00:0x15:mdrun_gpu returned 52
09:05:54:WU01:FS00:0x15:NANs detected on GPU
09:05:54:WU01:FS00:0x15:
09:05:54:WU01:FS00:0x15:Folding@home Core Shutdown: UNSTABLE_MACHINE
09:05:54:WU01:FS00:FahCore returned: UNSTABLE_MACHINE (122 = 0x7a)
09:05:54:WU01:FS00:Starting
09:05:54:WU01:FS00:Running FahCore: C:\\FAH\\Program/FAHCoreWrapper.exe c:/fah/Data/cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_15.fah/FahCore_15.exe -dir 01 -suffix 01 -version 701 -lifeline 2080 -checkpoint 15 -gpu 0
09:05:54:WU01:FS00:Started FahCore on PID 1584
09:05:54:WU01:FS00:Core PID:3764
09:05:54:WU01:FS00:FahCore 0x15 started
09:05:55:WU01:FS00:0x15:
09:05:55:WU01:FS00:0x15:*------------------------------*
09:05:55:WU01:FS00:0x15:Folding@Home GPU Core
09:05:55:WU01:FS00:0x15:Version                2.22 (Thu Dec 8 17:08:05 PST 2011)
09:05:55:WU01:FS00:0x15:Build host             SimbiosNvdWin7
09:05:55:WU01:FS00:0x15:Board Type             NVIDIA/CUDA
09:05:55:WU01:FS00:0x15:Core                   15
09:05:55:WU01:FS00:0x15:
09:05:55:WU01:FS00:0x15:Window's signal control handler registered.
09:05:55:WU01:FS00:0x15:Preparing to commence simulation
09:05:55:WU01:FS00:0x15:- Looking at optimizations...
09:05:55:WU01:FS00:0x15:DeleteFrameFiles: successfully deleted file=01/wudata_01.ckp
09:05:55:WU01:FS00:0x15:- Created dyn
09:05:55:WU01:FS00:0x15:- Files status OK
09:05:55:WU01:FS00:0x15:sizeof(CORE_PACKET_HDR) = 512 file=<>
09:05:55:WU01:FS00:0x15:- Expanded 548545 -> 890896 (decompressed 162.4 percent)
09:05:55:WU01:FS00:0x15:Called DecompressByteArray: compressed_data_size=548545 data_size=890896, decompressed_data_size=890896 diff=0
09:05:55:WU01:FS00:0x15:- Digital signature verified
09:05:55:WU01:FS00:0x15:
09:05:55:WU01:FS00:0x15:Project: 7640 (Run 76, Clone 0, Gen 14)
09:05:55:WU01:FS00:0x15:
09:05:55:WU01:FS00:0x15:Assembly optimizations on if available.
09:05:55:WU01:FS00:0x15:Entering M.D.
09:05:57:WU01:FS00:0x15:Tpr hash 01/wudata_01.tpr:  1726458347 522663929 1396732719 3771082946 1308202816
09:05:57:WU01:FS00:0x15:GPU device info: vendor=0 device=0 name=<NA> match=0
09:05:57:WU01:FS00:0x15:Working on Generated by trjconv : 21521 p3064_lambda5_2003 t= 5500.00049 in water
09:05:57:WU01:FS00:0x15:Client config unavailable.
09:05:58:WU01:FS00:0x15:Starting GUI Server
09:08:24:WU01:FS00:0x15:Setting checkpoint frequency: 25000
09:08:24:WU01:FS00:0x15:Completed         3 out of 2500000 steps (0%).
09:08:25:WARNING:WU01:FS00:Detected clock skew, adjusting time estimates
09:27:26:WU01:FS00:0x15:Completed     25000 out of 2500000 steps (1%).
09:46:33:WU01:FS00:0x15:Completed     50000 out of 2500000 steps (2%).
10:05:52:WU01:FS00:0x15:Completed     75000 out of 2500000 steps (3%).
10:25:14:WU01:FS00:0x15:Completed    100000 out of 2500000 steps (4%).
10:29:27:Server connection id=5 on 0.0.0.0:36330 from 192.168.1.21
10:29:32:Server connection id=2 ended
10:44:23:WU01:FS00:0x15:Completed    125000 out of 2500000 steps (5%).
11:03:31:WU01:FS00:0x15:Completed    150000 out of 2500000 steps (6%).
******************************** Date: 14/05/12 ********************************
11:22:40:WU01:FS00:0x15:Completed    175000 out of 2500000 steps (7%).
11:41:49:WU01:FS00:0x15:Completed    200000 out of 2500000 steps (8%).
12:01:00:WU01:FS00:0x15:Completed    225000 out of 2500000 steps (9%).
12:20:10:WU01:FS00:0x15:Completed    250000 out of 2500000 steps (10%).
12:39:24:WU01:FS00:0x15:Completed    275000 out of 2500000 steps (11%).
12:58:45:WU01:FS00:0x15:Completed    300000 out of 2500000 steps (12%).
13:18:03:WU01:FS00:0x15:Completed    325000 out of 2500000 steps (13%).
13:37:24:WU01:FS00:0x15:Completed    350000 out of 2500000 steps (14%).
13:56:34:WU01:FS00:0x15:Completed    375000 out of 2500000 steps (15%).
14:15:50:WU01:FS00:0x15:Completed    400000 out of 2500000 steps (16%).
14:35:06:WU01:FS00:0x15:Completed    425000 out of 2500000 steps (17%).
14:54:16:WU01:FS00:0x15:Completed    450000 out of 2500000 steps (18%).
15:13:24:WU01:FS00:0x15:Completed    475000 out of 2500000 steps (19%).
15:32:36:WU01:FS00:0x15:Completed    500000 out of 2500000 steps (20%).
15:51:56:WU01:FS00:0x15:Completed    525000 out of 2500000 steps (21%).

Is there anything I can do to this one? How do I purge the v7 client to get it to move on to a different WU?

[7im]

Close fahcontrol. Start, all programs, fahclient, data folder, work folder, delete the slot.

And yes, you can test downclocking the GPU below stock speeds to see if that helps avoid the NaN error, which is usually a data error caused by hardware pushed too hard. Also try to increase cooling as that may help as well.

[DrSpalding]

The GPU card may need to be cleaned out--it's probably been a few months since I blew out the dust. I'll do that this morning to see if it helps before blowing away the slot. To do it by hand via a command prompt, is it just deleting the directory named "00" or "01" in the "\FAH\Program\work"?

[DrSpalding]

Cleaning the card did not help. It failed the same WU again (Project: 7640 (Run 76, Clone 0, Gen 14)) so I nuked the slot and moved on to some 80xx WU and it is working just fine.

[bruce]

There seems to be a patter of troubles with p76xx on Fermi. The Pande Group is aware of the problem.