Linux CUDA 'Special' App finally available, featuring Low CPU use

Message boards : Number crunching : Linux CUDA 'Special' App finally available, featuring Low CPU use
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 49 · 50 · 51 · 52 · 53 · 54 · 55 . . . 83 · Next

AuthorMessage
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1886562 - Posted: 28 Aug 2017, 4:04:31 UTC

OK, registered for the BoincTasks forum and posted the question in the forum. I will post back when I receive a response.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1886562 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13720
Credit: 208,696,464
RAC: 304
Australia
Message 1886566 - Posted: 28 Aug 2017, 4:30:43 UTC - in response to Message 1886553.  

Stand back I'm thinking again,
Now that we have proven how tightly integrated BoincTasks and BOINC can become when it comes to scheduling requests, it brings to mind that some people have problems getting tasks and some don't - Is it BoincTasks that is causing it????

Not for me. Installed BOINC, attached to Seti, installed Lunatics. That's it. No other 3rd party programmes involved.
It's all due to the Dec Scheduler issue that stopped people getting any v8 work when they had only AP selected and v8 only when there was no work for their selected application (AP).
Grant
Darwin NT
ID: 1886566 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1886573 - Posted: 28 Aug 2017, 5:37:41 UTC - in response to Message 1886566.  

Grant, we were discussing my task download limitations per request. Not the ongoing issue of just getting work which as you say goes back to December of last year. Two separate issues. My task download per request limitation was definitely tied into BoincTasks.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1886573 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1886597 - Posted: 28 Aug 2017, 11:54:34 UTC - in response to Message 1886553.  

Stand back I'm thinking again,
Now that we have proven how tightly integrated BoincTasks and BOINC can become when it comes to scheduling requests, it brings to mind that some people have problems getting tasks and some don't - Is it BoincTasks that is causing it????

Does the scheduler request go through BT before it is sent to the server, or are the setting loaded when network communication is established?
What if .....



. . I have the download problem too but I am not running BOINC Tasks on the machines with the issue. So I am guessing ... NO!

Stephen

:(
ID: 1886597 · Report as offensive
Profile Darrell
Volunteer tester
Avatar

Send message
Joined: 14 Mar 03
Posts: 267
Credit: 1,418,681
RAC: 0
United States
Message 1886637 - Posted: 28 Aug 2017, 15:54:09 UTC

Ok peoples, let us clearup a misnomer. Boinc Manager and Boinc Tasks are simply Graphical User Interfaces. They have absolutely nothing to do with scheduling and server contacting. They are graphical versions of the Boinccmd.exe program. All scheduling, contacting the servers, etc takes place within the Boinc client program (Boinc.exe).
... and still I fear, and still I dare not laugh at the Mad Man!

Queen - The Prophet's Song
ID: 1886637 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1886658 - Posted: 28 Aug 2017, 18:35:39 UTC - in response to Message 1886637.  

OK, let's label the elements correctly. The client (BOINC.EXE) is what communicates with the scheduling servers. Yes, BoincTasks is a graphical management tool. It does have the ability to limit work by overriding the usual number of days work buffer and limit by the number of tasks. This is from Brent's post.

WU regulator
As of 1.43, this function can now be found in: C:\Program Files\eFMer\BoincTasks\examples config.xml.
Instead of having a work buffer for a number of days, BoincTasks has the option for a work buffer of xx Work Units.


Evidently I must have set this in BoincTasks a long while ago when I was having issues with Einstein inundating me with work. I had a backup file (.OLD) of the BoincTasks configuration file that had a hard limit of 400 tasks onboard it seems. That is what was causing my download issues. This is the response from BoincTask's developer, Fred.

BoincTasks simply sets no more work on the computer that reaches the maximum number of tasks. By giving the BOINC client a no more work instruction.
So this will NOT get you any more work.

If you don't get enough work check BoincTasks -> Extra -> BOINC preference
In the network tab set Minimum work buffer to 10 and the Additional work buffer to 10.
Repeat for all computers.

The only way I know to get more work on a single computer is by starting multiple BOINC clients.
https://efmer.com/linux-running-multiple-boinc-clients/
This is a great way to get plenty of work.

I see you reschedule work, that is a great way to mess things up and may be the cause of your problems.


He didn't understand my detailed post or glossed over it quickly based on his response. So what he is saying that by using the max task limit feature in BoincTasks it sets NNT for the client to not request work. That is not what was happening with my case though. I continued to get work at all times, as long as the project had work of course. My problem came from the small quantities received per request.

At this time, I have no clue why those two files had any effect but their removal has quickly erased the problem.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1886658 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1886679 - Posted: 28 Aug 2017, 21:13:34 UTC - in response to Message 1886495.  

. . Also, are you running Linux and the repository version of BOINC? Because there is a script you can run that restarts the BOINC client as a service without rebooting the system. It has saved me a lot of irritation on my rig that is running the above configuration. It is a small simple script.

. . You could even just issue the commands form the terminal each time.

:(

Yes I'm running repository BOINC, but I'm using Mint and I think I've read you can't stop the service without rebooting. Otherwise a script seems perfect.


'service boinc stop' to stop
'service boinc start' to start
'service boinc status' to get the status

Thanks W3Perl & Brent, works fine.

So does anyone have that script? I'm really not in the mood to do it manually...
ID: 1886679 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1886689 - Posted: 28 Aug 2017, 22:24:12 UTC - in response to Message 1886658.  


He didn't understand my detailed post or glossed over it quickly based on his response. So what he is saying that by using the max task limit feature in BoincTasks it sets NNT for the client to not request work. That is not what was happening with my case though. I continued to get work at all times, as long as the project had work of course. My problem came from the small quantities received per request.

At this time, I have no clue why those two files had any effect but their removal has quickly erased the problem.


. . Well Keith as they say, "the proof of the pudding", since the removal of the files allowed better functioning then their presence could be considered to have an effect on that function. Only at this point we don't know just what effect.

Stephen

:(
ID: 1886689 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1886693 - Posted: 28 Aug 2017, 22:34:52 UTC - in response to Message 1886679.  

This is the script I made for loading up for Tuesdays outages using the Petri/Laurent cpu2gpu.pl script

c2g.sh
# QOpt for Linux
# Run as ROOT

for (( ; ; ))
do

  clear
  # Stop BOINC
  echo  "  ======== Stop BOINC ======"
  /etc/init.d/boinc-client stop
  echo  "  ======== Sleep 10 ======="
  sleep 10
  #
  # Run CPU2GPU
  echo  "  ==== Running cpu2gpu ===="
  ./cpu2gpu.pl
  echo  "  ======== Sleep 2 ========"
  sleep 2
  #
  # Start BOINC
  echo  "  === Start BOINC Client =="
  /etc/init.d/boinc-client start
  echo  "  ========================="
  echo  "  === Sleep 15 Minutes ===="
  #
  date
  echo  "  ==== CTRL+C to Exit ====="
  sleep 15m
  #

done

You can remove the loop for a 'run once' option.

It's pretty simple, and gives a readable separation of the tasks being done.
ID: 1886693 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1886816 - Posted: 30 Aug 2017, 6:13:03 UTC - in response to Message 1886307.  


. . Hi again,

. . Moving in both directions is the biggy, but with Stubbles' script you can select the tasks to be moved by suspending them first. That would be great for me as I am used to that method. But I cannot speak for others ...

Stephen
:)


I have a new version which can :
1) move tasks from gpu to cpu (-r flag)
2) move only suspended tasks (-s flag).

If you want to have a try....
ID: 1886816 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1886825 - Posted: 30 Aug 2017, 11:58:20 UTC - in response to Message 1886816.  


I have a new version which can :
1) move tasks from gpu to cpu (-r flag)
2) move only suspended tasks (-s flag).

If you want to have a try....


. . 8^} Where do ya get it ??

Stephen

:)
ID: 1886825 · Report as offensive
Bruce
Volunteer tester

Send message
Joined: 15 Mar 02
Posts: 123
Credit: 124,955,234
RAC: 11
United States
Message 1886833 - Posted: 30 Aug 2017, 14:10:11 UTC - in response to Message 1886816.  

I have a new version which can :
1) move tasks from gpu to cpu (-r flag)
2) move only suspended tasks (-s flag).

If you want to have a try....



I'd be interested in getting it also.
Bruce
ID: 1886833 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1886914 - Posted: 30 Aug 2017, 19:30:27 UTC

Let me help you. You can try a private message or just wish. Whichever you think is more effective.
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: 1886914 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1886948 - Posted: 30 Aug 2017, 21:35:38 UTC - in response to Message 1886914.  

Let me help you. You can try a private message or just wish. Whichever you think is more effective.


. . Star light, star bright, first star I see tonight ...

Stephen

:)
ID: 1886948 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1887088 - Posted: 31 Aug 2017, 13:49:56 UTC - in response to Message 1886825.  
Last modified: 31 Aug 2017, 13:50:32 UTC


I have a new version which can :
1) move tasks from gpu to cpu (-r flag)
2) move only suspended tasks (-s flag).

If you want to have a try....


. . 8^} Where do ya get it ??

Stephen

:)


Sorry, you can get it at :
http://www.w3perl.com/seti/cpu2gpu.pl

beta software ! Let me know how to improve or send me bug report ! ;)
ID: 1887088 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1887615 - Posted: 3 Sep 2017, 0:51:39 UTC

. . @ Petri and TBar

. . Yay team! Thanks to you guys my current top rig has made Page 1! At a RAC of 103K Bertie has crawled into 20th spot. Now all I have to do is rebuild La-Bamba (only 2 x 1060-6G) and see where she climbs to. I don't think she will make page 1 but maybe page 2 :).

Stephen

:)
ID: 1887615 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1887977 - Posted: 5 Sep 2017, 10:46:38 UTC

When I try running the script in mint, I get this error, what do I do?

./cpu2gpu.pl -r

bash: ./cpu2gpu.pl: /usr/bin/perl^M: bad interpreter: No such file or directory
ID: 1887977 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1887979 - Posted: 5 Sep 2017, 11:43:19 UTC

I also has problems running the script in windows, first 2 times in was making ghost WUs, 3rd time I made a backup of BOINC before I tried, here it worked. So I tried again just now, suspended 50 WUs and ran the script:

cpu2gpu.pl -s
GPU version found : 821 710 - (class : opencl_nvidia_SoG cuda_opencl_100)
CPU version found : 800
AP version found : 703
Found GPU version used : 821 in D:/ProgramData/BOINC/client_state.xml (opencl_nvidia_SoG / windows_x86_64)
Found CPU Version used : 800 in D:/ProgramData/BOINC/client_state.xml
219 wu found
-----
Your current D:/ProgramData/BOINC/client_state.xml :
- CPU (MB8_win_x64_AVX_VS2010_r3330.exe) (800) : 50 wu
==> Converted wu from CPU to GPU : 50

But when I ran BOINC again, the 50 WUs dissarired into ghost state :(
ID: 1887979 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1887980 - Posted: 5 Sep 2017, 12:01:02 UTC - in response to Message 1887977.  

When I try running the script in mint, I get this error, what do I do?

./cpu2gpu.pl -r

bash: ./cpu2gpu.pl: /usr/bin/perl^M: bad interpreter: No such file or directory


. . It is a "perl" script and you need to have perl installed to run it.

Stephen

..
ID: 1887980 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1888029 - Posted: 6 Sep 2017, 1:56:10 UTC

. . @ TBar and/or Petri

. . I know this app is not intended to crunch Arecibo VLAR tasks but I have processed dozens of the blighters using it on both the GTX970s and the GTX1050ti. So when I needed work for the GPUs over the maintenance outage I felt OK moving a couple of dozen more across for the occasion. Those processed by the GTX970s went fine as previously, but the four that ended up on the GTX1050 ALL failed as computation errors. My suspicion is that it may be because the 970s and the 1050ti all have 4Gb ram while the 1050 only has 2GB, but it also has only 5 CUs so I thought it would not be an issue. I am including links to the results in case you want to have a look.

. . I do not understand why they have the error "Exit time limit exceeded" as they only ran for 8 mins each and the expected run time for these tasks on the 1050 is 10 mins.

http://setiathome.berkeley.edu/result.php?resultid=5989290320

http://setiathome.berkeley.edu/result.php?resultid=5989290326

http://setiathome.berkeley.edu/result.php?resultid=5989290352

http://setiathome.berkeley.edu/result.php?resultid=5989289944

Stephen

??
ID: 1888029 · Report as offensive
Previous · 1 . . . 49 · 50 · 51 · 52 · 53 · 54 · 55 . . . 83 · Next

Message boards : Number crunching : Linux CUDA 'Special' App finally available, featuring Low CPU use


 
©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.