Anything relating to AstroPulse tasks

Message boards : Number crunching : Anything relating to AstroPulse tasks
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 120 · Next

AuthorMessage
Profile JaundicedEye
Avatar

Send message
Joined: 14 Mar 12
Posts: 5375
Credit: 30,870,693
RAC: 1
United States
Message 1699888 - Posted: 9 Jul 2015, 17:46:23 UTC - in response to Message 1699877.  

Aps have been split quite nicely for almost a day and my GPU caches have been filling so slowly that I have yet been able to get any for my CPU cache.

+1

"Sour Grapes make a bitter Whine." <(0)>
ID: 1699888 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1699907 - Posted: 9 Jul 2015, 18:54:39 UTC
Last modified: 9 Jul 2015, 18:57:05 UTC

I have a couple of AP tasks for intel gpu.
They all run about 10 times longer, use very little GPU (4-6%) and deliver only
about 10% of credits per unit of time than all the others of the same type.
Has anyone experienced something similar?
http://setiathome.berkeley.edu/workunit.php?wuid=1824192084
ID: 1699907 · Report as offensive
BetelgeuseFive Project Donor
Volunteer tester

Send message
Joined: 6 Jul 99
Posts: 158
Credit: 17,117,787
RAC: 19
Netherlands
Message 1699916 - Posted: 9 Jul 2015, 19:22:42 UTC - in response to Message 1699907.  

That task is a Multibeam on NVIDIA GPU, not an Astropulse on Intel GPU.
I don't see anything wrong with it and there are not enough other Multibeam NVIDIA tasks from the same computer to compare it with.

Tom


I have a couple of AP tasks for intel gpu.
They all run about 10 times longer, use very little GPU (4-6%) and deliver only
about 10% of credits per unit of time than all the others of the same type.
Has anyone experienced something similar?
http://setiathome.berkeley.edu/workunit.php?wuid=1824192084
ID: 1699916 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1699917 - Posted: 9 Jul 2015, 19:26:45 UTC

ID: 1699917 · Report as offensive
BetelgeuseFive Project Donor
Volunteer tester

Send message
Joined: 6 Jul 99
Posts: 158
Credit: 17,117,787
RAC: 19
Netherlands
Message 1699919 - Posted: 9 Jul 2015, 19:40:40 UTC - in response to Message 1699917.  

Did you install a new driver ?
According to the output a recompile was performed after a task restart:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
INFO: can't open binary kernel file: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\AstroPulse_Kernels_r2737.cl_IntelRHDGraphics4000.bin_V7_TWIN_FFA_1018104226, continue with recompile...
Info : Building Program (binary, clBuildProgram):main kernels: OK code 0

Maybe there is something wrong with this new driver ???

Tom

Sorry, wrong link.
This is the one
http://setiathome.berkeley.edu/workunit.php?wuid=1828604364
ID: 1699919 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1699921 - Posted: 9 Jul 2015, 19:47:29 UTC

I installed a new driver, but for nvidia.
ID: 1699921 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1699923 - Posted: 9 Jul 2015, 20:06:26 UTC - in response to Message 1699919.  

Did you install a new driver ?
According to the output a recompile was performed after a task restart:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
INFO: can't open binary kernel file: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\AstroPulse_Kernels_r2737.cl_IntelRHDGraphics4000.bin_V7_TWIN_FFA_1018104226, continue with recompile...
Info : Building Program (binary, clBuildProgram):main kernels: OK code 0

Maybe there is something wrong with this new driver ???

Tom

Sorry, wrong link.
This is the one
http://setiathome.berkeley.edu/workunit.php?wuid=1828604364

All of their other tasks look relatively normal & are also using the v4226 driver. The "Number of app instances per device set to:3" does give me pause.
Another issue I see at the moment is that host is currently not reporting that it has an iGPU to BOINC. Only a GT 720M.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1699923 · Report as offensive
Profile Blurf
Volunteer tester

Send message
Joined: 2 Sep 06
Posts: 8962
Credit: 12,678,685
RAC: 0
United States
Message 1699926 - Posted: 9 Jul 2015, 20:23:53 UTC

Pulled a full sheet of AP's today...nice


ID: 1699926 · Report as offensive
Profile TimeLord04
Volunteer tester
Avatar

Send message
Joined: 9 Mar 06
Posts: 21140
Credit: 33,933,039
RAC: 23
United States
Message 1699928 - Posted: 9 Jul 2015, 20:26:35 UTC

I've got 7 APs on each machine - 14 total!!! :-)
TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos
ID: 1699928 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1699937 - Posted: 9 Jul 2015, 21:21:36 UTC - in response to Message 1699923.  

Somehow nv driver 353.30 messes up my intel-gpu tasks.
Once installed and settled, every new task started seems to run fine.
However, every intel-gpu tasks, that was started progresses very slowly with very little gpu utilization.
Unless there is a way to "reset" a single task, i guess, i just have to sit it out.
ID: 1699937 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1699954 - Posted: 9 Jul 2015, 22:02:11 UTC - in response to Message 1699937.  
Last modified: 9 Jul 2015, 22:03:11 UTC

Somehow nv driver 353.30 messes up my intel-gpu tasks.
Once installed and settled, every new task started seems to run fine.
However, every intel-gpu tasks, that was started progresses very slowly with very little gpu utilization.
Unless there is a way to "reset" a single task, i guess, i just have to sit it out.


I'd try suspend and resume for Seti at home in Boinc manager projects tab. Or suspend and resume for the task.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1699954 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1699957 - Posted: 9 Jul 2015, 22:09:11 UTC - in response to Message 1699954.  

Tried it-no change
ID: 1699957 · Report as offensive
Profile Louis Loria II
Volunteer tester
Avatar

Send message
Joined: 20 Oct 03
Posts: 259
Credit: 9,208,040
RAC: 24
United States
Message 1699970 - Posted: 9 Jul 2015, 22:59:20 UTC - in response to Message 1699928.  

Picked up 27 today myself. half an hour to 45mins each so far. Seems pretty quick to me considering the starting estimate is about 53hrs. Why so high an estimate? Anyhow, I certainly have the impression APs are sought after. Glad to have some...
ID: 1699970 · Report as offensive
Speedy
Volunteer tester
Avatar

Send message
Joined: 26 Jun 04
Posts: 1639
Credit: 12,921,799
RAC: 89
New Zealand
Message 1699973 - Posted: 9 Jul 2015, 23:21:39 UTC - in response to Message 1699970.  
Last modified: 9 Jul 2015, 23:33:21 UTC

Picked up 27 today myself. half an hour to 45mins each so far. Seems pretty quick to me considering the starting estimate is about 53hrs. Why so high an estimate? Anyhow, I certainly have the impression APs are sought after. Glad to have some...

After you return around 12 results the results estimated run-time will normalise. Reasons for such high runtimes at present I believe is because Boinc is getting used to running these tasks

Edited: Added the word estimate
ID: 1699973 · Report as offensive
Speedy
Volunteer tester
Avatar

Send message
Joined: 26 Jun 04
Posts: 1639
Credit: 12,921,799
RAC: 89
New Zealand
Message 1699976 - Posted: 9 Jul 2015, 23:24:21 UTC

There are plenty of tapes waiting to be split however I feel 2 June AC maybe be stuck as it is sitting at (14) and has been for at least the last hour more like 2
ID: 1699976 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1699977 - Posted: 9 Jul 2015, 23:24:29 UTC - in response to Message 1699973.  

Picked up 27 today myself. half an hour to 45mins each so far. Seems pretty quick to me considering the starting estimate is about 53hrs. Why so high an estimate? Anyhow, I certainly have the impression APs are sought after. Glad to have some...

After you return around 12 results the results run-time will normalise.

The run-time will be unchanged, but the estimated run-time will normalise.
ID: 1699977 · Report as offensive
Speedy
Volunteer tester
Avatar

Send message
Joined: 26 Jun 04
Posts: 1639
Credit: 12,921,799
RAC: 89
New Zealand
Message 1699980 - Posted: 9 Jul 2015, 23:30:46 UTC - in response to Message 1699977.  

Thank you Richard for picking up my incorrect wording
ID: 1699980 · Report as offensive
OTS
Volunteer tester

Send message
Joined: 6 Jan 08
Posts: 369
Credit: 20,533,537
RAC: 0
United States
Message 1699983 - Posted: 9 Jul 2015, 23:42:43 UTC - in response to Message 1699781.  

Damn!!

I should've taken my CPU MB top up 1 day earlier than I did. :-O

I could've had both my CPU's back to crunching AP's by now otherwise. :-(

Cheers.



I am not sure if it would have made much difference when all is said and done at the end of this AP run. I kept my cache low in case APs started coming but the CPU is crunching faster than then APs are downloading so I have to periodically add a small amount of MBs just to keep the CPU busy between APs. :(
ID: 1699983 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1699985 - Posted: 10 Jul 2015, 0:02:47 UTC - in response to Message 1699983.  

Damn!!

I should've taken my CPU MB top up 1 day earlier than I did. :-O

I could've had both my CPU's back to crunching AP's by now otherwise. :-(

Cheers.



I am not sure if it would have made much difference when all is said and done at the end of this AP run. I kept my cache low in case APs started coming but the CPU is crunching faster than then APs are downloading so I have to periodically add a small amount of MBs just to keep the CPU busy between APs. :(


Same here....I get a few APs and then the caches fill with MB work. All my rigs are set to accept MB anytime AP is not available.

This AP run is not going like the last. The splitting rate is not making a lot of headway on demand. I have been floating up and down from about 600 APs in progress all day now. I am processing them about as fast as I can snag them.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1699985 · Report as offensive
OTS
Volunteer tester

Send message
Joined: 6 Jan 08
Posts: 369
Credit: 20,533,537
RAC: 0
United States
Message 1699992 - Posted: 10 Jul 2015, 0:37:39 UTC - in response to Message 1699985.  

Damn!!

I should've taken my CPU MB top up 1 day earlier than I did. :-O

I could've had both my CPU's back to crunching AP's by now otherwise. :-(

Cheers.



I am not sure if it would have made much difference when all is said and done at the end of this AP run. I kept my cache low in case APs started coming but the CPU is crunching faster than then APs are downloading so I have to periodically add a small amount of MBs just to keep the CPU busy between APs. :(


Same here....I get a few APs and then the caches fill with MB work. All my rigs are set to accept MB anytime AP is not available.


The problem with that approach, particularly with a slow CPU, is that you easily reach your 100 WU limit with MBs and then there is no room for APs even if one should come your way. I like to download 20-30 MBs at a time and then set it for AP only so there is always plenty of room for APs, even 4 or 5 at time ( I can dream, can’t I) if they are available. Of course I only have the one machine and not your farm to manage so it is not that hard to monitor. :)


This AP run is not going like the last. The splitting rate is not making a lot of headway on demand. I have been floating up and down from about 600 APs in progress all day now. I am processing them about as fast as I can snag them.
ID: 1699992 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 120 · Next

Message boards : Number crunching : Anything relating to AstroPulse tasks


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.