| Author |
Message |
|---|
Neil ASend message
Joined: 9 Oct 08
Posts: 50
Credit: 12,676,739
RAC: 0
Level
Scientific publications
   |
|
Updates for this release can be found here...
Link
Beta release.....
____________
Crunching for the benefit of humanity and in memory of my dad and other family members. |
|
|
|
|
|
Interesting change logs ...
It looks like Rom is actually trying to work on some of the long standing issues that make BOINC look like a child's toy ... |
|
|
|
|
|
Hmm I've updated and now it doesn't seem to be able to complete the CPU benchmarks. Is anybody else seeing this issue?
12/03/2009 2:08:26 AM Running CPU benchmarks
12/03/2009 2:08:27 AM Suspending computation - running CPU benchmarks
12/03/2009 2:08:58 AM FP benchmark ran only 0.031250 sec; ignoring
12/03/2009 2:08:59 AM Resuming computation
|
|
|
|
|
|
2009-03-12 11:09:12 Starting BOINC client version 6.6.15 for windows_intelx86
2009-03-12 11:09:12 log flags: task, file_xfer, sched_ops
2009-03-12 11:09:12 Libraries: libcurl/7.19.4 OpenSSL/0.9.8j zlib/1.2.3
2009-03-12 11:09:12 Running as a daemon
2009-03-12 11:09:12 Data directory: C:\Documents and Settings\All Users\Dane aplikacji\BOINC
2009-03-12 11:09:12 Running under account boinc_master
2009-03-12 11:09:12 Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad CPU Q8200 @ 2.33GHz [x86 Family 6 Model 23 Stepping 7]
2009-03-12 11:09:12 Processor features: fpu tsc pae nx sse sse2 mmx
2009-03-12 11:09:12 OS: Microsoft Windows XP: Professional x86 Editon, Dodatek Service Pack 3, (05.01.2600.00)
2009-03-12 11:09:12 Memory: 2.00 GB physical, 3.80 GB virtual
2009-03-12 11:09:12 Disk: 32.00 GB total, 16.51 GB free
2009-03-12 11:09:12 Local time is UTC +1 hours
2009-03-12 11:09:12 CUDA device: GeForce 8600 GT (driver version 18206, CUDA version 1.1, 256MB, est. 14GFLOPS)
2009-03-12 11:09:12 Not using a proxy
2009-03-12 11:09:12 Version change (6.4.5 -> 6.6.15)
2009-03-12 11:09:12 rosetta@home URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 1006611; location: (none); project prefs: default
2009-03-12 11:09:12 GPUGRID URL: http://www.gpugrid.net/; Computer ID: 28766; location: (none); project prefs: default
2009-03-12 11:09:12 General prefs: from http://milkyway.cs.rpi.edu/milkyway/ (last modified 09-Nov-2008 21:01:32)
2009-03-12 11:09:12 Host location: none
2009-03-12 11:09:12 General prefs: using your defaults
2009-03-12 11:09:12 Reading preferences override file
2009-03-12 11:09:12 Preferences limit memory usage when active to 1023.49MB
2009-03-12 11:09:12 Preferences limit memory usage when idle to 1842.28MB
2009-03-12 11:09:13 Preferences limit disk usage to 16.00GB
2009-03-12 11:09:13 Running CPU benchmarks
2009-03-12 11:09:13 Suspending computation - running CPU benchmarks
2009-03-12 11:09:44 Benchmark results:
2009-03-12 11:09:44 Number of CPUs: 4
2009-03-12 11:09:44 3186 floating point MIPS (Whetstone) per CPU
2009-03-12 11:09:44 6490 integer MIPS (Dhrystone) per CPU
2009-03-12 11:09:45 GPUGRID Restarting task eJY2397-SH2_US_8-0-10-SH2_US_8520000_0 using acemd version 662
2009-03-12 11:09:47 rosetta@home Restarting task h001__BOINC_ABRELAX_SAVE_ALL_OUT_RANGE_t460_IGNORE_THE_REST-S25-7-S3-7--h001_-_7725_4926_0 using minirosetta version 154
2009-03-12 11:09:50 rosetta@home Restarting task 1c9oA_BOINC_ABINITIO_IGNORE_THE_REST-MOO34-S25-18-S3-8--1c9oA-_7867_81_0 using minirosetta version 154
2009-03-12 11:09:53 rosetta@home Restarting task abinitio_nohomfrag_129_B_1wjgA_SAVE_ALL_OUT_7581_1625_0 using minirosetta version 154
2009-03-12 11:09:56 rosetta@home Restarting task h001__BOINC_ABRELAX_SAVE_ALL_OUT_RANGE_t460_IGNORE_THE_REST-S25-5-S3-7--h001_-_7725_5167_0 using minirosetta version 154
after downloading 1 work this apera:
2009-03-12 11:15:46 GPUGRID Sending scheduler request: To fetch work.
2009-03-12 11:15:46 GPUGRID Requesting new tasks
2009-03-12 11:15:51 GPUGRID Scheduler request failed: HTTP internal server error
So I have 4 WU in GPUGRID and this is it, maby this message is "wrong".
Since I have 8600GT i need manually abort 3 task when finished one, due to time limit...
It should be longer time limit or ability too set e.g. " download 1 wu". |
|
|
MarkJ  Volunteer moderator
Volunteer tester
Send message
Joined: 24 Dec 08
Posts: 738
Credit: 200,909,904
RAC: 0
Level
Scientific publications
   |
2009-03-12 11:09:12 Starting BOINC client version 6.6.15 for windows_intelx86
2009-03-12 11:09:12 log flags: task, file_xfer, sched_ops
2009-03-12 11:09:12 Libraries: libcurl/7.19.4 OpenSSL/0.9.8j zlib/1.2.3
2009-03-12 11:09:12 Running as a daemon
2009-03-12 11:09:12 Data directory: C:\Documents and Settings\All Users\Dane aplikacji\BOINC
2009-03-12 11:09:12 Running under account boinc_master
2009-03-12 11:09:12 Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad CPU Q8200 @ 2.33GHz [x86 Family 6 Model 23 Stepping 7]
2009-03-12 11:09:12 Processor features: fpu tsc pae nx sse sse2 mmx
2009-03-12 11:09:12 OS: Microsoft Windows XP: Professional x86 Editon, Dodatek Service Pack 3, (05.01.2600.00)
2009-03-12 11:09:12 Memory: 2.00 GB physical, 3.80 GB virtual
2009-03-12 11:09:12 Disk: 32.00 GB total, 16.51 GB free
2009-03-12 11:09:12 Local time is UTC +1 hours
2009-03-12 11:09:12 CUDA device: GeForce 8600 GT (driver version 18206, CUDA version 1.1, 256MB, est. 14GFLOPS)
2009-03-12 11:09:12 Not using a proxy
2009-03-12 11:09:12 Version change (6.4.5 -> 6.6.15)
2009-03-12 11:09:12 rosetta@home URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 1006611; location: (none); project prefs: default
2009-03-12 11:09:12 GPUGRID URL: http://www.gpugrid.net/; Computer ID: 28766; location: (none); project prefs: default
2009-03-12 11:09:12 General prefs: from http://milkyway.cs.rpi.edu/milkyway/ (last modified 09-Nov-2008 21:01:32)
2009-03-12 11:09:12 Host location: none
2009-03-12 11:09:12 General prefs: using your defaults
2009-03-12 11:09:12 Reading preferences override file
2009-03-12 11:09:12 Preferences limit memory usage when active to 1023.49MB
2009-03-12 11:09:12 Preferences limit memory usage when idle to 1842.28MB
2009-03-12 11:09:13 Preferences limit disk usage to 16.00GB
2009-03-12 11:09:13 Running CPU benchmarks
2009-03-12 11:09:13 Suspending computation - running CPU benchmarks
2009-03-12 11:09:44 Benchmark results:
2009-03-12 11:09:44 Number of CPUs: 4
2009-03-12 11:09:44 3186 floating point MIPS (Whetstone) per CPU
2009-03-12 11:09:44 6490 integer MIPS (Dhrystone) per CPU
2009-03-12 11:09:45 GPUGRID Restarting task eJY2397-SH2_US_8-0-10-SH2_US_8520000_0 using acemd version 662
2009-03-12 11:09:47 rosetta@home Restarting task h001__BOINC_ABRELAX_SAVE_ALL_OUT_RANGE_t460_IGNORE_THE_REST-S25-7-S3-7--h001_-_7725_4926_0 using minirosetta version 154
2009-03-12 11:09:50 rosetta@home Restarting task 1c9oA_BOINC_ABINITIO_IGNORE_THE_REST-MOO34-S25-18-S3-8--1c9oA-_7867_81_0 using minirosetta version 154
2009-03-12 11:09:53 rosetta@home Restarting task abinitio_nohomfrag_129_B_1wjgA_SAVE_ALL_OUT_7581_1625_0 using minirosetta version 154
2009-03-12 11:09:56 rosetta@home Restarting task h001__BOINC_ABRELAX_SAVE_ALL_OUT_RANGE_t460_IGNORE_THE_REST-S25-5-S3-7--h001_-_7725_5167_0 using minirosetta version 154
after downloading 1 work this apera:
2009-03-12 11:15:46 GPUGRID Sending scheduler request: To fetch work.
2009-03-12 11:15:46 GPUGRID Requesting new tasks
2009-03-12 11:15:51 GPUGRID Scheduler request failed: HTTP internal server error
So I have 4 WU in GPUGRID and this is it, maby this message is "wrong".
Since I have 8600GT i need manually abort 3 task when finished one, due to time limit...
It should be longer time limit or ability too set e.g. " download 1 wu".
Running as a damon (service) isn't meant to work, so that might account for some of your problem.
Regarding the download of 4 tasks, that is a project setting and nothing to do with BOINC. The people running GPUgrid are aware that it excludes the slower cards like yours, as it cannot possibly complete the work in the time allowed. However they need the data back quickly so thats why they have short deadlines.
____________
BOINC blog |
|
|
|
|
The people running GPUgrid are aware that it excludes the slower cards like yours, as it cannot possibly complete the work in the time allowed.
I don't agree with your statment.
8600GT need 2 days to comlete WU. The problem is, in this case, that I have 4 cores, if i have 1 core 1 WU will be downloaded and proceed so it is no problem. With 4 cores 4 WU are downloaded and it can not be procced before dedline, so 3 of it must be manually removed... and this is real problem, not speed of my VGA. |
|
|
MarkJ Volunteer moderator
Volunteer tester
Send message
Joined: 24 Dec 08
Posts: 738
Credit: 200,909,904
RAC: 0
Level
Scientific publications
|
The people running GPUgrid are aware that it excludes the slower cards like yours, as it cannot possibly complete the work in the time allowed.
I don't agree with your statment.
8600GT need 2 days to comlete WU. The problem is, in this case, that I have 4 cores, if i have 1 core 1 WU will be downloaded and proceed so it is no problem. With 4 cores 4 WU are downloaded and it can not be procced before dedline, so 3 of it must be manually removed... and this is real problem, not speed of my VGA.
Well if your VGA was faster you wouldn't have an issue with returning 4 wu within the 4 day limit. The deadline is a project setting, they could have it further out, but they choose to set it to 4 days. Other projects use longer deadlines as they choose.
You are correct in that BOINC (on the server side) will grant you 1 wu per cpu core. As far as I know they can amend that, but probably can't set it lower than 1.
____________
BOINC blog |
|
|
uBronan
Send message
Joined: 1 Feb 09
Posts: 139
Credit: 575,023
RAC: 0
Level
Scientific publications
 |
|
Well i do manage them by hand,Its called babysitting a project ;)
In terms of gpugrid i have a fairly slow card (9600 GT 64 shaders).
So i just select by no more new work if the downloaded units are reported downloading.
Its the best and safe method, i leave max 2 units available for processing
If a workunit shows a estimate time of under 50+ hours i do 2 if higher say 60+ i choose 1 unit in that way i make sure not to download units i cant proces in time.
Anyway i have been doing this since i started this project and works fine because once in say 10 hours you have a change the project finish a unit, enough time to check
|
|
|
MarkJ Volunteer moderator
Volunteer tester
Send message
Joined: 24 Dec 08
Posts: 738
Credit: 200,909,904
RAC: 0
Level
Scientific publications
|
|
It would appear that BOINC is requesting cuda work only from GPUgrid, at least under 6.6.15
Host Project Date Message
Qui-Gon GPUGRID 16/03/2009 8:15:52 PM [sched_op_debug] Starting scheduler request
Qui-Gon GPUGRID 16/03/2009 8:15:52 PM Sending scheduler request: Requested by project.
Qui-Gon GPUGRID 16/03/2009 8:15:52 PM Requesting new tasks
Qui-Gon GPUGRID 16/03/2009 8:15:52 PM [sched_op_debug] CPU work request: 0.00 seconds; 0 idle CPUs
Qui-Gon GPUGRID 16/03/2009 8:15:52 PM [sched_op_debug] CUDA work request: 64800.00 seconds; 1 idle GPUs
Host Project Date Message
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM Scheduler request completed: got 1 new tasks
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM [sched_op_debug] Server version 607
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM Project requested delay of 31 seconds
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM [sched_op_debug] estimated total CPU job duration: 0 seconds
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM [sched_op_debug] estimated total CUDA job duration: 119254 seconds
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM [sched_op_debug] Deferring communication for 31 sec
Qui-Gon GPUGRID 16/03/2009 8:15:58 PM [sched_op_debug] Reason: requested by project
Host Project Date Message
Qui-Gon GPUGRID 16/03/2009 8:16:00 PM Started download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-LICENSE
Qui-Gon GPUGRID 16/03/2009 8:16:00 PM Started download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-COPYRIGHT
Qui-Gon GPUGRID 16/03/2009 8:16:01 PM Finished download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-LICENSE
Qui-Gon GPUGRID 16/03/2009 8:16:01 PM Finished download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-COPYRIGHT
Qui-Gon GPUGRID 16/03/2009 8:16:01 PM Started download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-Sh17894-SH2_US_8-9-50-SH2_US_8190000_1
Qui-Gon GPUGRID 16/03/2009 8:16:01 PM Started download of Sh17894-SH2_US_8-10-50-SH2_US_8190000-Sh17894-SH2_US_8-9-50-SH2_US_8190000_2
____________
BOINC blog |
|
|
/ 
Reply
Quote
Paul D. Buck
