Is there a problem with acemd 6.69?

Message boards : Graphics cards (GPUs) : Is there a problem with acemd 6.69?

Author	Message
Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12644 - Posted: 23 Sep 2009 \| 2:16:54 UTC
	I have had 4 work units in a row fail with error 107374. When I had a bit of a look I noticed that the last successful units were running v6.67. Hardware Intel i7, 4GB RAM, NVIDIA GTX295 Driver 190.62 I've reset the project after the first two failures, the next pair of workunits also had the same computation error - 107374 Any ideas?
	ID: 12644 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12660 - Posted: 23 Sep 2009 \| 4:49:42 UTC - in response to Message 12644.
	More have failed with same error - 107374. I'm running Vista 64 bit, GTX295 190.62 , BOINC 6.6.36 Error details : Problem signature: Problem Event Name: APPCRASH Application Name: acemd_6.69_windows_intelx86__cuda.exe Application Version: 0.0.0.0 Application Timestamp: 4ab753ae Fault Module Name: cudart.dll!cudaRuntimeGetVersion Fault Module Version: 6.0.6002.18005 Fault Module Timestamp: 49e03824 Exception Code: c0000139 Exception Offset: 0006f04e OS Version: 6.0.6002.2.2.0.256.1 Locale ID: 3081 Additional Information 1: 9d13 Additional Information 2: 1abee00edb3fc1158f9ad6f44f0f6be8 Additional Information 3: 9d13 Additional Information 4: 1abee00edb3fc1158f9ad6f44f0f6be8
	ID: 12660 \| Rating: 0 \| rate: / Reply Quote

RyanChen Send message Joined: 2 Dec 08 Posts: 5 Credit: 3,027,593 RAC: 0 Level Scientific publications	Message 12666 - Posted: 23 Sep 2009 \| 6:39:21 UTC - in response to Message 12660.
	Did you solve the problem? I encountered the similar problem, after reset the project and reboot did not help! I made a trial to replace the cudart.dll & cufft.dll which placed in GPUGRID project folder with version CUDA 2.3 runtime lib. I downloaded from http://tinyurl.com/lwr6av Amazingly it works without exit code!!! I'm not sure if it is right method to solve the problem, so remember to backup the files and do it at your own risk.
	ID: 12666 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12669 - Posted: 23 Sep 2009 \| 7:50:46 UTC - in response to Message 12666.
	Sounds like you have a simiar problem to me. Resets and reboots did not fix it for me. I will try the CUDA 2.3 runtime and see what happens, althought I suspect there is something wrong with the code being delivered by GPUGRID. Or are we just tangled up in the transition from 2.2 to 2.3??
	ID: 12669 \| Rating: 0 \| rate: / Reply Quote

GDF Volunteer moderator Project administrator Project developer Project tester Volunteer developer Volunteer tester Project scientist Send message Joined: 14 Mar 07 Posts: 1957 Credit: 629,356 RAC: 0 Level Scientific publications	Message 12672 - Posted: 23 Sep 2009 \| 8:16:29 UTC - in response to Message 12669.
	New application coming out in a couple of hours. gdf
	ID: 12672 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12673 - Posted: 23 Sep 2009 \| 8:24:32 UTC - in response to Message 12672.
	Cool. Will that mean I can leave my toy running the current drivers from NVIDIA, ie 190.62?
	ID: 12673 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12675 - Posted: 23 Sep 2009 \| 8:34:49 UTC - in response to Message 12666.
	Did you solve the problem? I encountered the similar problem, after reset the project and reboot did not help! I made a trial to replace the cudart.dll & cufft.dll which placed in GPUGRID project folder with version CUDA 2.3 runtime lib. I downloaded from http://tinyurl.com/lwr6av Amazingly it works without exit code!!! I'm not sure if it is right method to solve the problem, so remember to backup the files and do it at your own risk. Just for those thinking about playing with CUDA versions, I downloaded the CUDA V2.3 driver from CUDA Zone (I won't use drivers from sources other than the manufacturer's web site) and it continued to get the same error. So it is probably not worth while. I was going to download the 2.2 driver, then saw the msg from GDF so I will wait and see what happens a bit later when he has organised his magic :) I have a preference for running science related to health rather than things like Milkyway or SETI. Although I do run SETI, but not on my GPUs, this is because I have been running it for most of 10 yrs.
	ID: 12675 \| Rating: 0 \| rate: / Reply Quote

GDF Volunteer moderator Project administrator Project developer Project tester Volunteer developer Volunteer tester Project scientist Send message Joined: 14 Mar 07 Posts: 1957 Credit: 629,356 RAC: 0 Level Scientific publications	Message 12679 - Posted: 23 Sep 2009 \| 9:04:05 UTC - in response to Message 12675. Last modified: 23 Sep 2009 \| 9:04:25 UTC
	New application out. I am not sure that it will solve the problem, but it is a try. acemd_6.71 gdf
	ID: 12679 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12681 - Posted: 23 Sep 2009 \| 9:39:17 UTC - in response to Message 12679.
	Hi GDF, I have reset the project and let it schedule work and I am now getting : AECMD beta version is not available for this computer Full-atom molecular dynamics on Cell Processor is not available for your type of computer Dave
	ID: 12681 \| Rating: 0 \| rate: / Reply Quote

ignasi Send message Joined: 10 Apr 08 Posts: 254 Credit: 16,836,000 RAC: 0 Level Scientific publications	Message 12683 - Posted: 23 Sep 2009 \| 9:39:46 UTC - in response to Message 12681.
	try again now. ignasi
	ID: 12683 \| Rating: 0 \| rate: / Reply Quote

Dave_In_Oz Send message Joined: 13 Jul 09 Posts: 32 Credit: 287,042,950 RAC: 0 Level Scientific publications	Message 12685 - Posted: 23 Sep 2009 \| 9:48:56 UTC - in response to Message 12683.
	Hi Ignasi, Looking good, I reset project again and it successfully downloaded and started 2 tasks on the GTX295. They are approaching 1% done which is further than they were getting earlier. Thanks for your help Dave
	ID: 12685 \| Rating: 0 \| rate: / Reply Quote

dataman Send message Joined: 18 Sep 08 Posts: 36 Credit: 100,352,867 RAC: 0 Level Scientific publications	Message 12688 - Posted: 23 Sep 2009 \| 16:28:13 UTC
	The 6.71 app seems to be working fine now on 6.10.6 and GTX260 cards. :) Win7 64-bit, BOINC 6.10.6, nVidia 190.62, CUDA 2.3 Thanks! ____________
	ID: 12688 \| Rating: 0 \| rate: / Reply Quote

GDF Volunteer moderator Project administrator Project developer Project tester Volunteer developer Volunteer tester Project scientist Send message Joined: 14 Mar 07 Posts: 1957 Credit: 629,356 RAC: 0 Level Scientific publications	Message 12689 - Posted: 23 Sep 2009 \| 16:31:14 UTC - in response to Message 12688.
	Wow. I saw that you have many GTX260. Did you have any problem with the Nvidia FFT bug? gdf
	ID: 12689 \| Rating: 0 \| rate: / Reply Quote

dataman Send message Joined: 18 Sep 08 Posts: 36 Credit: 100,352,867 RAC: 0 Level Scientific publications	Message 12690 - Posted: 23 Sep 2009 \| 16:40:25 UTC - in response to Message 12689. Last modified: 23 Sep 2009 \| 17:31:23 UTC
	Wow. I saw that you have many GTX260. Did you have any problem with the Nvidia FFT bug? gdf Not really. Until the 6.69 came out yesterday, I had maybe 1 failed wu per day over 7 GTX260's but it also happened on the GTX285 too. When 6.69 came out, every wu failed on every machine. I have XP SP3 running Ok now and 1 Win7. I will convert the Vista 64-bit today. In all, I have 6 Win7, 1 Vista and 1 XP ... 7 GTX260's and 1 GTX285. Thanks for the quick resolution!! EDIT: The Vista machine is working fine now also. :) Now I will put the rest of the machines on the config I posted earlier. ____________
	ID: 12690 \| Rating: 0 \| rate: / Reply Quote

Rabinovitch Send message Joined: 25 Aug 08 Posts: 143 Credit: 64,937,578 RAC: 0 Level Scientific publications	Message 12698 - Posted: 24 Sep 2009 \| 2:19:19 UTC
	Sehr Gut! I will try to get a couple of Wus for testing too, 'cause my PC is under Win 7 x64 too and equipped with GTX260 (unfortunately, single). Now need to wait for Collatz Wus to be finished... :-) ____________ From Siberia with love!
	ID: 12698 \| Rating: 0 \| rate: / Reply Quote

Rabinovitch Send message Joined: 25 Aug 08 Posts: 143 Credit: 64,937,578 RAC: 0 Level Scientific publications	Message 12700 - Posted: 24 Sep 2009 \| 4:30:42 UTC
	First WU under 6.71 app failed on Win 7 x64 (GTX260): 24.09.2009 9:23:55 GPUGRID Starting task 838-GIANNI_BIND001-36-100-RND2858_2 using acemd version 671 24.09.2009 11:04:23 GPUGRID Computation for task 838-GIANNI_BIND001-36-100-RND2858_2 finished 24.09.2009 11:04:23 GPUGRID Output file 838-GIANNI_BIND001-36-100-RND2858_2_1 for task 838-GIANNI_BIND001-36-100-RND2858_2 absent 24.09.2009 11:04:23 GPUGRID Output file 838-GIANNI_BIND001-36-100-RND2858_2_2 for task 838-GIANNI_BIND001-36-100-RND2858_2 absent 24.09.2009 11:04:23 GPUGRID Output file 838-GIANNI_BIND001-36-100-RND2858_2_3 for task 838-GIANNI_BIND001-36-100-RND2858_2 absent 24.09.2009 11:04:30 GPUGRID Started upload of 838-GIANNI_BIND001-36-100-RND2858_2_0 24.09.2009 11:04:30 GPUGRID Started upload of 838-GIANNI_BIND001-36-100-RND2858_2_4 ____________ From Siberia with love!
	ID: 12700 \| Rating: 0 \| rate: / Reply Quote

Paul D. Buck Send message Joined: 9 Jun 08 Posts: 1050 Credit: 37,321,185 RAC: 0 Level Scientific publications	Message 12701 - Posted: 24 Sep 2009 \| 6:20:00 UTC
	I guess I am just slow to react, but, I just noticed that I have quite a few failed tasks with this version in the last two days. Across, I think all of my systems. THough I should note that it looks like I have one in progress that is going to complete as it is 90% plus complete ... So, the following tasks had the 139 error: Computer 19410: GTX 295 cards - 757-GIANNI_BIND001-30-100-RND7538_2 - 492-GIANNI_BINDT1-7-10-RND6252_1 - m35000-IBUCH_1_pYEEI_2209-0-20-RND5275_1 - 72-KASHIF_HIVPR_sub_so_ba1-33-100-RND2448_1 - PMEcube25-OTTO_HERG4-4-40-RND3779_2 - PME191-TONI_HERG2-6-40-RND1676_1 - 131-KASHIF_HIVPR_sub_so_ba2-34-100-RND4021_1 - 171-GIANNI_BIND001-27-100-RND7217_1 - 778-GIANNI_BIND001-44-100-RND0745_2 - 717-GIANNI_BIND001-38-100-RND3246_2 - 196-KASHIF_HIVPR_sub_so_ba2-37-100-RND7121_1 - 19-GIANNI_CLOZ-16-50-RND8348_2 Computer 36139: GTX 260 cards - 570-GIANNI_BIND001-35-100-RND1972_2 - PMEcube45-OTTO_HERG5-8-40-RND4347_2 - 100-GIANNI_VIL-0-100-RND1180_1 - PMEcube60-TONI_HERG2-10-40-RND5555_1 I will note that I have not see a string of failures like this in like forever ...
	ID: 12701 \| Rating: 0 \| rate: / Reply Quote

GDF Volunteer moderator Project administrator Project developer Project tester Volunteer developer Volunteer tester Project scientist Send message Joined: 14 Mar 07 Posts: 1957 Credit: 629,356 RAC: 0 Level Scientific publications	Message 12704 - Posted: 24 Sep 2009 \| 7:42:04 UTC - in response to Message 12701.
	Hi Paul, your are filing on the previous application (the first release). Then we have updloaded another two. 6.70/6.71. gdf
	ID: 12704 \| Rating: 0 \| rate: / Reply Quote

Paul D. Buck Send message Joined: 9 Jun 08 Posts: 1050 Credit: 37,321,185 RAC: 0 Level Scientific publications	Message 12719 - Posted: 24 Sep 2009 \| 15:56:48 UTC - in response to Message 12704.
	Hi Paul, your are filing on the previous application (the first release). Then we have updloaded another two. 6.70/6.71. gdf Only because the deaths were on the date that I was filing the report, yesterday. And I still had one in hand from that version. I knew it was probably a waste of time, but, one never knows ... error on the side of pointing out problems is not a bad rule ...
	ID: 12719 \| Rating: 0 \| rate: / Reply Quote

Post to thread

Message boards : Graphics cards (GPUs) : Is there a problem with acemd 6.69?

	About	Science	Volunteers	Performance	Forum	Join us	Donate