News:

Follow BoincTasks on Twitter Facebook        Visit our website here.
BoincTasks cloud login is working again

Main Menu

Testing Version 3.60

Started by fred, May 19, 2011, 05:31:02 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

fred

Add: Independent temperature regulator for each GPU.
Add: When a CPU signature changes a message box will be shown with a warning. At the same time the number of temperature sensors is checked.
Add: An option in Preferences, to allowing opening the dialog, as administrator only.
Changed: Every GPU now has it's own regulator thread, with a randomized offset.
Changed: Minimum CPU and GPU regulator 2->1%.
Fixed: In BOINC connection mode, TThrottle didn't always reconnect when a connection was lost. (Due to a restart of the BOINC client).
Fixed: Graphs, draw the horizontal and vertical text last, to avoid them to be overdrawn by other items.

New translation: Preference tab: Allow TThrottle to be opened as administrator only.
New translation: Administrator access only = balloon text when option above is active.

Pepo

#1
On one machine I'm still running TTh 3.60. Approx. at the time I've upgraded BT 1.05->1.06 (it could be around 11:54 +- 5 minutes, but not exactly) the TTh's CPU usage and I/O suddenly went up approximately twice:

   

At the same time the level of throttling was rising, but just very slowly, no reason for such change:

   

The last two entries in log are:
Quote----------------------------------------------------------------------------------------------------------------------------------------------
14 June 2011 - 11:14:58  Počet zhodných programov (procesov): 8
CPU:1 - PID:7068 (3)   Slot:14   http://cpdnbeta.oerc.ox.ac.uk/   famous_s3fk_599_200_000222629
CPU:1, GPU:0, PID:5024 (4)   Child:   famous_um_6.11_windows_intelx86.exe
GPU:1 -11 PID:15276 (3)   Slot:17   http://www.gpugrid.net/   5-GIANNI_TESTDHFR10-0-5-RND7358
CPU:1 - PID:11996 (3)   Slot:16   http://www.primegrid.com/   llrCUL_91107183
CPU:1, GPU:0, PID:11884 (1)   Child:   primegrid_cllr.exe
CPU:1 - PID:14560 (3)   Slot:0   http://wuprop.boinc-af.org/   wu_1307636806_35865
CPU:1 - PID:15964 (6)   Slot:3   http://qcn.stanford.edu/sensor/   qcnb_007059
CPU:1 - PID:7820 (6)   Slot:4   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_98
CPU:1 - PID:14268 (6)   Slot:6   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_97
CPU:1 - PID:13296 (3)   Slot:9   http://www.worldcommunitygrid.org/   X0000111761498200907271408
----------------------------------------------------------------------------------------------------------------------------------------------
14 June 2011 - 11:53:37  Počet zhodných programov (procesov): 7
CPU:1 - PID:7068 (3)   Slot:14   http://cpdnbeta.oerc.ox.ac.uk/   famous_s3fk_599_200_000222629
CPU:1, GPU:0, PID:5024 (4)   Child:   famous_um_6.11_windows_intelx86.exe
GPU:1 -11 PID:15276 (3)   Slot:17   http://www.gpugrid.net/   5-GIANNI_TESTDHFR10-0-5-RND7358
CPU:1 - PID:11996 (3)   Slot:16   http://www.primegrid.com/   llrCUL_91107183
CPU:1, GPU:0, PID:11884 (1)   Child:   primegrid_cllr.exe
CPU:1 - PID:15964 (6)   Slot:3   http://qcn.stanford.edu/sensor/   qcnb_007059
CPU:1 - PID:7820 (6)   Slot:4   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_98
CPU:1 - PID:14268 (6)   Slot:6   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_97
CPU:1 - PID:13296 (3)   Slot:9   http://www.worldcommunitygrid.org/   X0000111761498200907271408
----------------------------------------------------------------------------------------------------------------------------------------------
14 June 2011 - 11:54:39  Počet zhodných programov (procesov): 8
CPU:1 - PID:7068 (3)   Slot:14   http://cpdnbeta.oerc.ox.ac.uk/   famous_s3fk_599_200_000222629
CPU:1, GPU:0, PID:5024 (4)   Child:   famous_um_6.11_windows_intelx86.exe
GPU:1 -11 PID:15276 (3)   Slot:17   http://www.gpugrid.net/   5-GIANNI_TESTDHFR10-0-5-RND7358
CPU:1 - PID:11996 (3)   Slot:16   http://www.primegrid.com/   llrCUL_91107183
CPU:1, GPU:0, PID:11884 (1)   Child:   primegrid_cllr.exe
CPU:1 - PID:15964 (6)   Slot:3   http://qcn.stanford.edu/sensor/   qcnb_007059
CPU:1 - PID:7820 (6)   Slot:4   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_98
CPU:1 - PID:14268 (6)   Slot:6   http://www.freehal.net/freehal_at_home/   fh_nci_0_30723189_97
CPU:1 - PID:13296 (3)   Slot:9   http://www.worldcommunitygrid.org/   X0000111761498200907271408
CPU:1 - PID:10296 (3)   Slot:0   http://wuprop.boinc-af.org/   wu_1307636806_40395
----------------------------------------------------------------------------------------------------------------------------------------------
where the second log entry is approx. the time when it started happening.

Relevant BOINC log entries:
Quote14.06.2011 11:53:27 | WUProp@Home | [task] result wu_1307636806_35865_0 checkpointed
14.06.2011 11:53:31 | WUProp@Home | Computation for task wu_1307636806_35865_0 finished
14.06.2011 11:53:39 | FreeHAL@home | [task] result fh_nci_0_30723189_97_0 checkpointed
14.06.2011 11:53:41 | WUProp@Home | Sending scheduler request: To fetch work.
14.06.2011 11:53:41 | WUProp@Home | Reporting 1 completed tasks, requesting new tasks for CPU
14.06.2011 11:53:42 | WUProp@Home | Scheduler request completed: got 1 new tasks
14.06.2011 11:53:42 | WUProp@Home | [task] result state=NEW for wu_1307636806_40395_0 from handle_scheduler_reply
14.06.2011 11:53:43 | FreeHAL@home | [task] result fh_nci_0_30723189_98_0 checkpointed
14.06.2011 11:53:44 | FreeHAL@home | [task] result fh_nci_0_30723189_97_0 checkpointed
14.06.2011 11:53:45 | WUProp@Home | [task] result state=FILES_DOWNLOADED for wu_1307636806_40395_0 from CS::update_results
14.06.2011 11:53:45 | WUProp@Home | [task] task_state=EXECUTING for wu_1307636806_40395_0 from start
14.06.2011 11:53:45 | WUProp@Home | Starting task wu_1307636806_40395_0 using data_collect version 243
14.06.2011 11:53:47 | FreeHAL@home | [task] result fh_nci_0_30723189_98_0 checkpointed
14.06.2011 11:53:48 | WUProp@Home | [task] result wu_1307636806_40395_0 checkpointed

I don't have any explanation for what could happen.




I've tried to suspend the WUProp task (nCi task with 3 threads, zero CPU usage, occasional I/O blip in few minutes). As TTh noticed it, its CPU throttle graph went up from 62% to 78% and CPU usage jumped down to previous level, after resuming the task TTh's CPU usage again jumped up to higher level and CPU throttle graph went to 63%.

I'll check this once more after installing TTh 4.xx series.
Peter

fred

Quote from: Pepo on June 14, 2011, 10:51:30 AM
I'll check this once more after installing TTh 4.xx series.
I only support V 4.2. ;D

Pepo

Quote from: fred on June 14, 2011, 10:54:00 AM
Quote from: Pepo on June 14, 2011, 10:51:30 AM
I'll check this once more after installing TTh 4.xx series.
I only support V 4.2. ;D
Then I should've noticed it on 3.41? :))

Later I'll test 4.20 too.
Peter