Beta BOINC Client 6.6.7 available

Message boards : Graphics cards (GPUs) : Beta BOINC Client 6.6.7 available

Author	Message
Neil A Send message Joined: 9 Oct 08 Posts: 50 Credit: 12,676,739 RAC: 0 Level Scientific publications	Message 6661 - Posted: 15 Feb 2009 \| 20:27:41 UTC
	The above captioned has been posted for download and testing. This is a beta build and should only be used by those willing to test the logic and possibly put up with bugs. One of the major items in this build is a revamped work fetch routine that appears to separate out GPU vs CPU resouces and tracks debt separately. I will post the list of fixes in this build and 6.6.6 once I receive it. I am running 6.6.7 on a quad core with 2xGTX 260 Core 216's. So far so good. ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 6661 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6662 - Posted: 15 Feb 2009 \| 20:45:30 UTC
	Changelog 6.6.7
	ID: 6662 \| Rating: 0 \| rate: / Reply Quote

Wassertropfen Send message Joined: 14 Aug 08 Posts: 15 Credit: 13,774,919 RAC: 0 Level Scientific publications	Message 6669 - Posted: 16 Feb 2009 \| 0:23:48 UTC - in response to Message 6662. Last modified: 16 Feb 2009 \| 0:24:00 UTC
	How to get new WU? The fckng 6.6.* clients trys to get from the 7 non cuda app work but not from GpuGrid: :( See here: 16.02.2009 01:23:00 GPUGRID CUDA work request: 0.00 seconds; 0 idle GPUs Yes, no cuda task is running and gpugrid runs out of work. :( ____________ Constant dripping wears away the stone. :)
	ID: 6669 \| 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 6672 - Posted: 16 Feb 2009 \| 3:22:39 UTC - in response to Message 6669.
	How to get new WU? Use the last known working version ... 6.5.0 ...
	ID: 6672 \| 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 6673 - Posted: 16 Feb 2009 \| 12:05:47 UTC
	As mentioned under the BOINC 6.6.5 thread it doesn't honour network times (or cpu times), that is it assumes it can crunch 24/7 and use the network 24/7. Supposedly fixed in the next version, whenever that comes out. ____________ BOINC blog
	ID: 6673 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6689 - Posted: 16 Feb 2009 \| 21:39:28 UTC - in response to Message 6669.
	How to get new WU? The fckng 6.6.* clients trys to get from the 7 non cuda app work but not from GpuGrid: :( See here: 16.02.2009 01:23:00 GPUGRID CUDA work request: 0.00 seconds; 0 idle GPUs Yes, no cuda task is running and gpugrid runs out of work. :( There are 2 fu**ing scheduling policies now, the normal RPC request and a project scheduling request (you can check it over the properties button). The second work fetch is activated while there is a no new work message from the first (totally useless)
	ID: 6689 \| 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 6698 - Posted: 17 Feb 2009 \| 0:02:17 UTC
	I have had some luck with 6.6.7 over the past few days. I am getting good response from 6.6.7 on a quad core with 2 x GTX 260 Core 216's. Both WCG and GPU work is being fetched and a reasonable sized work buffer maintained based on my preferences. CPU and GPU is working 24x7. Not sure why Rebirther is having problems, but perhaps a different configuration or pool of pending work units. I do hope he manages to get it sorted out asap. Neil ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 6698 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6699 - Posted: 17 Feb 2009 \| 0:38:45 UTC - in response to Message 6698.
	I have had some luck with 6.6.7 over the past few days. I am getting good response from 6.6.7 on a quad core with 2 x GTX 260 Core 216's. Both WCG and GPU work is being fetched and a reasonable sized work buffer maintained based on my preferences. CPU and GPU is working 24x7. Not sure why Rebirther is having problems, but perhaps a different configuration or pool of pending work units. I do hope he manages to get it sorted out asap. Neil Switched back to 6.6.0 after some tests and nasty bugs ;) Problem solved by this way :)
	ID: 6699 \| Rating: 0 \| rate: / Reply Quote

Post to thread

Message boards : Graphics cards (GPUs) : Beta BOINC Client 6.6.7 available

	About	Science	Volunteers	Performance	Forum	Join us	Donate

Author	Message
Neil A Send message Joined: 9 Oct 08 Posts: 50 Credit: 12,676,739 RAC: 0 Level Scientific publications	Message 6661 - Posted: 15 Feb 2009 \| 20:27:41 UTC
	The above captioned has been posted for download and testing. This is a beta build and should only be used by those willing to test the logic and possibly put up with bugs. One of the major items in this build is a revamped work fetch routine that appears to separate out GPU vs CPU resouces and tracks debt separately. I will post the list of fixes in this build and 6.6.6 once I receive it. I am running 6.6.7 on a quad core with 2xGTX 260 Core 216's. So far so good. ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 6661 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6662 - Posted: 15 Feb 2009 \| 20:45:30 UTC
	Changelog 6.6.7
	ID: 6662 \| Rating: 0 \| rate: / Reply Quote

Wassertropfen Send message Joined: 14 Aug 08 Posts: 15 Credit: 13,774,919 RAC: 0 Level Scientific publications	Message 6669 - Posted: 16 Feb 2009 \| 0:23:48 UTC - in response to Message 6662. Last modified: 16 Feb 2009 \| 0:24:00 UTC
	How to get new WU? The fckng 6.6.* clients trys to get from the 7 non cuda app work but not from GpuGrid: :( See here: 16.02.2009 01:23:00 GPUGRID CUDA work request: 0.00 seconds; 0 idle GPUs Yes, no cuda task is running and gpugrid runs out of work. :( ____________ Constant dripping wears away the stone. :)
	ID: 6669 \| 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 6672 - Posted: 16 Feb 2009 \| 3:22:39 UTC - in response to Message 6669.
	How to get new WU? Use the last known working version ... 6.5.0 ...
	ID: 6672 \| 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 6673 - Posted: 16 Feb 2009 \| 12:05:47 UTC
	As mentioned under the BOINC 6.6.5 thread it doesn't honour network times (or cpu times), that is it assumes it can crunch 24/7 and use the network 24/7. Supposedly fixed in the next version, whenever that comes out. ____________ BOINC blog
	ID: 6673 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6689 - Posted: 16 Feb 2009 \| 21:39:28 UTC - in response to Message 6669.
	How to get new WU? The fckng 6.6.* clients trys to get from the 7 non cuda app work but not from GpuGrid: :( See here: 16.02.2009 01:23:00 GPUGRID CUDA work request: 0.00 seconds; 0 idle GPUs Yes, no cuda task is running and gpugrid runs out of work. :( There are 2 fu**ing scheduling policies now, the normal RPC request and a project scheduling request (you can check it over the properties button). The second work fetch is activated while there is a no new work message from the first (totally useless)
	ID: 6689 \| 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 6698 - Posted: 17 Feb 2009 \| 0:02:17 UTC
	I have had some luck with 6.6.7 over the past few days. I am getting good response from 6.6.7 on a quad core with 2 x GTX 260 Core 216's. Both WCG and GPU work is being fetched and a reasonable sized work buffer maintained based on my preferences. CPU and GPU is working 24x7. Not sure why Rebirther is having problems, but perhaps a different configuration or pool of pending work units. I do hope he manages to get it sorted out asap. Neil ____________ Crunching for the benefit of humanity and in memory of my dad and other family members.
	ID: 6698 \| Rating: 0 \| rate: / Reply Quote

rebirther Send message Joined: 7 Jul 07 Posts: 53 Credit: 3,048,781 RAC: 0 Level Scientific publications	Message 6699 - Posted: 17 Feb 2009 \| 0:38:45 UTC - in response to Message 6698.
	I have had some luck with 6.6.7 over the past few days. I am getting good response from 6.6.7 on a quad core with 2 x GTX 260 Core 216's. Both WCG and GPU work is being fetched and a reasonable sized work buffer maintained based on my preferences. CPU and GPU is working 24x7. Not sure why Rebirther is having problems, but perhaps a different configuration or pool of pending work units. I do hope he manages to get it sorted out asap. Neil Switched back to 6.6.0 after some tests and nasty bugs ;) Problem solved by this way :)
	ID: 6699 \| Rating: 0 \| rate: / Reply Quote