Beta BOINC client 6.6.20 available

Message boards : Graphics cards (GPUs) : Beta BOINC client 6.6.20 available

Author	Message
Neil A Send message Joined: 9 Oct 08 Posts: 50 Credit: 12,676,739 RAC: 0 Level Scientific publications	Message 8001 - Posted: 30 Mar 2009 \| 20:28:22 UTC
	A new release is available for download and testing... this is a beta release for testing only ..... Change log to follow shortly. ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 8001 \| Rating: 0 \| rate: / Reply Quote

Stefan Ledwina Send message Joined: 16 Jul 07 Posts: 464 Credit: 135,911,881 RAC: 83 Level Scientific publications	Message 8002 - Posted: 30 Mar 2009 \| 20:39:55 UTC
	Change Log: - MGR: Remove erroneous call of locale in main thread - manager: display versions as x.xx - manager: fix message - client: typo in work fetch code; may cause minor errors ____________ pixelicious.at - my little photoblog
	ID: 8002 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8003 - Posted: 30 Mar 2009 \| 20:53:49 UTC - in response to Message 8002.
	Seems like development is converging to something, if that's the entire change log :) MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8003 \| 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 8004 - Posted: 30 Mar 2009 \| 21:47:20 UTC - in response to Message 8003.
	Seems like development is converging to something, if that's the entire change log :) Or they are getting tired. There are still a few wok fetch issues. And now the language files are cranked. My concern is that they will release it even before it is well tested.
	ID: 8004 \| Rating: 0 \| rate: / Reply Quote

Clownius Send message Joined: 19 Feb 09 Posts: 37 Credit: 30,657,566 RAC: 0 Level Scientific publications	Message 8008 - Posted: 31 Mar 2009 \| 7:32:53 UTC
	Isn't that Microsoft's solution? Open source doesn't usually do things quite that silly.
	ID: 8008 \| Rating: 0 \| rate: / Reply Quote

MarkJ Volunteer moderator Volunteer tester Send message Joined: 24 Dec 08 Posts: 738 Credit: 200,909,904 RAC: 0 Level Scientific publications	Message 8012 - Posted: 31 Mar 2009 \| 11:45:57 UTC - in response to Message 8004.
	Seems like development is converging to something, if that's the entire change log :) Or they are getting tired. There are still a few wok fetch issues. And now the language files are cranked. My concern is that they will release it even before it is well tested. Oh come on Paul I know you are just itching to test the Swedish language version. Or was it the French version? LOL. It does look like its fairly well behaved these days despite various suggestions falling on deaf ears. Just upgraded most of boxes to 18 too. Will try 20 on one of them and see how it handles things. ____________ BOINC blog
	ID: 8012 \| 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 8014 - Posted: 31 Mar 2009 \| 16:33:04 UTC - in response to Message 8012.
	Seems like development is converging to something, if that's the entire change log :) Or they are getting tired. There are still a few wok fetch issues. And now the language files are cranked. My concern is that they will release it even before it is well tested. Oh come on Paul I know you are just itching to test the Swedish language version. Or was it the French version? LOL. It does look like its fairly well behaved these days despite various suggestions falling on deaf ears. Just upgraded most of boxes to 18 too. Will try 20 on one of them and see how it handles things. Actually it was the FORTRAN and Pascal language versions ... I just installed 6.6.20 on my Mac Pro to start my own look-see, and already asked a niggle question which will be ignored as you noted. I don't feel any insane pressure to move my CUDA machines yet 6.5.0 is working well enough that there is no immediate need or reason.
	ID: 8014 \| Rating: 0 \| rate: / Reply Quote

Neil A Send message Joined: 9 Oct 08 Posts: 50 Credit: 12,676,739 RAC: 0 Level Scientific publications	Message 8017 - Posted: 31 Mar 2009 \| 17:20:59 UTC
	Hello gentlemen, As I understand it, the BOINC development team would like to get a GA release out the door pretty soon. I agree with the comments that 6.6 is coming along nicely after some pretty difficult times. I have a number of systems running 6.6.18 and 6.6.20 and they are behaving nicely. I only run WCG on CPU and GPUGRID on GPU, so am not experiencing the problems of those who run multiple concurrent CPU or GPU projects. That seems to be one of the main areas still being tweaked up. ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 8017 \| Rating: 0 \| rate: / Reply Quote

MarkJ Volunteer moderator Volunteer tester Send message Joined: 24 Dec 08 Posts: 738 Credit: 200,909,904 RAC: 0 Level Scientific publications	Message 8030 - Posted: 1 Apr 2009 \| 8:20:19 UTC - in response to Message 8017.
	Hello gentlemen, As I understand it, the BOINC development team would like to get a GA release out the door pretty soon. I agree with the comments that 6.6 is coming along nicely after some pretty difficult times. I have a number of systems running 6.6.18 and 6.6.20 and they are behaving nicely. I only run WCG on CPU and GPUGRID on GPU, so am not experiencing the problems of those who run multiple concurrent CPU or GPU projects. That seems to be one of the main areas still being tweaked up. It looks like the GPU side is well behaved and the CPU is possibly not requesting enough work to keep a cache full. I have run with mixed cpu and gpu with Seti and that seems to get along quite nicely with the same app names (but different version numbers). I also have different projects using different resources. GPUgrid seems to hog the gpu a lot because of the short deadlines. The gpu apps run in strict deadline order so whatever comes in with an earlier deadline gets the gpu and the others have to wait. I need more gpu's LOL. Actually i'm hanging out to see what the GTX212 is like as I was hoping to put one in each of my quaddies. ____________ BOINC blog
	ID: 8030 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8127 - Posted: 3 Apr 2009 \| 15:54:14 UTC Last modified: 3 Apr 2009 \| 16:09:43 UTC
	I tried 6.6.20 on a machine and it can't get work from GPU-Grid: 03.04.2009 16:21:30 GPUGRID Sending scheduler request: Requested by project. 03.04.2009 16:21:30 GPUGRID Requesting new tasks 03.04.2009 16:21:35 GPUGRID Scheduler request completed: got 0 new tasks 03.04.2009 16:21:35 GPUGRID Message from server: No work sent 03.04.2009 16:21:35 GPUGRID Message from server: Full-atom molecular dynamics on Cell processor is not available for your type of computer. 03.04.2009 16:21:35 GPUGRID Message from server: CUDA app exists for Full-atom molecular dynamics but no CUDA work requested It's a Vista 64 PC which was working with 6.4.5. Driver is now 185.66, but the same happened with a 181.xx version, whch was installed previously. Not sure if I could provide you with any other useful information. Edit: downgrade to 6.5.0 solved the probem.. just in time for the week end. MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8127 \| 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 8131 - Posted: 3 Apr 2009 \| 18:19:32 UTC
	Did you have work? Or were you dry? For whatever reason BOINC did not think that you needed CUDA work. Not sure why or even if it knew you had CUDA card installed. The logs are long gone so no way to tell for sure ... There are a couple flags you can turn on to try to see what was happening, the sad thing is that on my systems for example you get so much information (lots of projects attached) that it is almost impossible to figure out what is important. They also removed the information that would tell you what kind of request BOINC was making, several of us have suggested on the Dev and Alpha mailing lists that this is a mistake (done to "simplify" and make "friendly" the messages tab that they claim people find confusing - wait till we can't trouble shoot and see how confusing it gets, but I digress) ... it would be NICE if you would add your voice to the chorus so that we can get this critical information back ... PM me with your e-mail and I will send a pre-built cc_config you can use ...
	ID: 8131 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8139 - Posted: 3 Apr 2009 \| 20:25:13 UTC - in response to Message 8131.
	After thinking about it for a few more seconds I think Boinc 6.6.20 might have tried to get cpu-work from GPU-Grid. Anything else just doesn't make sense. Is there some server flag to tell the new client "man, you can ask me a hundred times and I still ain't going to give you any cpu-work!"? And some more info: - running 3 projects: GPU-Grid (share 70), ABC (share 100), Seti (share 100, no new work) - cache set to 0.1 + 0.2 days - 1.5 GPU WUs left Not sure how large the ABC cache was, it was not very much though. MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8139 \| Rating: 0 \| rate: / Reply Quote

MarkJ Volunteer moderator Volunteer tester Send message Joined: 24 Dec 08 Posts: 738 Credit: 200,909,904 RAC: 0 Level Scientific publications	Message 8150 - Posted: 4 Apr 2009 \| 3:25:41 UTC - in response to Message 8139.
	After thinking about it for a few more seconds I think Boinc 6.6.20 might have tried to get cpu-work from GPU-Grid. Anything else just doesn't make sense. Is there some server flag to tell the new client "man, you can ask me a hundred times and I still ain't going to give you any cpu-work!"? And some more info: - running 3 projects: GPU-Grid (share 70), ABC (share 100), Seti (share 100, no new work) - cache set to 0.1 + 0.2 days - 1.5 GPU WUs left Not sure how large the ABC cache was, it was not very much though. MrS Turn on the <sched_op_debug> flag in cc_config and you can see some more details, which makes it a bit clearer. It looks like its doing a cpu request, as you worked out. It should then try a cuda work request request. Interestingly GPUgrid is the only project, that I run, that give a message for the wrong sort of work. Seti handles both, Einstein ignores the cuda ones (it says 0 work sent). Orbit (which doesn't have any of either sort) ignores the cuda and from memory so does CPDN. ____________ BOINC blog
	ID: 8150 \| Rating: 0 \| rate: / Reply Quote

Stefan Ledwina Send message Joined: 16 Jul 07 Posts: 464 Credit: 135,911,881 RAC: 83 Level Scientific publications	Message 8152 - Posted: 4 Apr 2009 \| 5:42:34 UTC - in response to Message 8150.
	And maybe try to reset the debts if you try 6.6.20 again. You can easily do that with <zero_debts>1</zero_debts> (in the <options> part)in your cc_config.xml. After the first start of BOINC all debts should be 0 and you can then delete this line in the config file, or change the 1 to 0... ____________ pixelicious.at - my little photoblog
	ID: 8152 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8154 - Posted: 4 Apr 2009 \| 8:09:32 UTC
	So it seems like this message is not common for 6.6.20? MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8154 \| Rating: 0 \| rate: / Reply Quote

Post to thread

Message boards : Graphics cards (GPUs) : Beta BOINC client 6.6.20 available

	About	Science	Volunteers	Performance	Forum	Join us	Donate