Message boards :
Number crunching :
Setting up Linux to crunch CUDA90 and above for Windows users
Message board moderation
Previous · 1 . . . 133 · 134 · 135 · 136 · 137 · 138 · 139 . . . 162 · Next
Author | Message |
---|---|
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
@ TBar . . I am running your boinc client 7.14.2 on three Linux machines but I am getting slightly flaky results. While I did not change any settings on setup two machines are behaving very differently to the third. The odd machine is not displaying "ack" messages in the log but the other two are, it is also displaying the stats page normally while the other 2 are plotting all RACs from the date of the creation of that machine (after nearly 3 months it is getting very squeezy on that graph). Is there a way to change this behaviour? I know there is a setting to reduce messages in the log but I cannot remember where that is, but I have not heard of any setting for the stats tag in manager. Stephen ? ? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Client configuration Logging: <sched_op_debug>(including acks) Options: <save_stats_days>N</save_stats_days> |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13731 Credit: 208,696,464 RAC: 304 |
the other 2 are plotting all RACs from the date of the creation of that machineThe period of time for which the Manager displays the RAC can be set- i had all my RAC on one system (till a crash) from when i first joined the BOINC version of Seti. From memory the default is 30 days? <cc_config> <log_flags> </log_flags> <options> <save_stats_days>3665</save_stats_days> </options> </cc_config> Edit- Richard beat me to it. Grant Darwin NT |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
|
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
The period of time for which the Manager displays the RAC can be set- i had all my RAC on one system (till a crash) from when i first joined the BOINC version of Seti. From memory the default is 30 days? . . But thanks for the reply, it all helps .. Stephen :) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . Redundant response. .. |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Is there a Boinc Screen saver for Linux? Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Is there a Boinc Screen saver for Linux? No. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I understand that trying to balance multiple projects on one box is tricky. My goal is to run a mix of projects on my cpu threads. The gpu's are all Seti@Home, all the time. The cpus are currently Seti@Home 2 threads max. (resource 1000) (1 in cache, 1 running) Seti@Home Beta 1 thread max (resource 10) (44 in cache, 1 running) and World Community Grid 5 threads max (resource 100) (5 in cache, 5 running). WCG is already constrained so it doesn't bury my cpu tasks with unlimited tasks. The cpus are set to 90% which translates into 14 threads available. What I am missing is 2 S@H threads running. I have 1 running now and the log reports it didn't ask for any more S@H cpu work. Sun 24 Nov 2019 03:37:54 AM CST | SETI@home | [sched_op] CPU work request: 0.00 seconds; 0.00 devices If I set S@H Beta to NNT (no new tasks) and delete the bulk of the Beta tasks then S@H cpu tasks are downloaded and "all is good". Any ideas about how to get 2 S@H cpu threads to be fed constantly (besides dropping Beta and WCG :) Tom A proud member of the OFA (Old Farts Association). |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
Maybe by using <project_max_concurrent>xx</project_max_concurrent> Seems to work on my laptop, it's only running CPU tasks on SETI & WCG. I have set WCG to 3 project_max_concurrent. |
Frank Meade Send message Joined: 15 May 99 Posts: 6 Credit: 613,302,878 RAC: 139 |
I am currently investigating using Ubuntu to run Boinc/Seti@Home. I have one computer running Ubuntu 18.04 and it is doing well on SETI. One of the reasons for switching from Windows is that Ubuntu could be running Tasks faster. Early results suggest that there is no increase in speed - maybe some lose in speed. I am in need of the CUDA app some times referred to as the "Special Sauce." How do I get it? Can anybody help me? |
rob smith Send message Joined: 7 Mar 03 Posts: 22190 Credit: 416,307,556 RAC: 380 |
Using "stock" and the "normal" optimised applications it is normal for there to be very little difference between Windows and Linux (on the same hardware). However the special sauce is a huge leap in performance - others wil be around to give you the low-down on how to find it and run it. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I am currently investigating using Ubuntu to run Boinc/Seti@Home. I have one computer running Ubuntu 18.04 and it is doing well on SETI. One of the reasons for switching from Windows is that Ubuntu could be running Tasks faster. Early results suggest that there is no increase in speed - maybe some lose in speed. This is most useful for the Nvidia gpu "special sauce". I am uncertain if the cpu apps in it are better than stock. They might be as fast as the Lunatic's cpu app though. Try this link. It should download the archive: http://www.arkayn.us/lunatics/BOINC.7z If you don't have an un-archiver let us know. Tom ps. I think this is the system you are experimenting with? https://setiathome.berkeley.edu/show_host_detail.php?hostid=8852490 A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Ubuntu's Archive Manager doesn't natively understand 7zip compression. Add it to the Archive Manager with: sudo apt install p7zip Mint apparently does understand 7zip compression natively. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Since you are running the distro version of BOINC, you either need to uninstall the distro version and run the TBar version included in the package. Or copy the science applications and support files along with the app_info.xml file into your distro BOINC folder and restart to come up on the anonymous platform. You need to set NNT on your current tasks and run them out before switching or you will abandon your current work. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . Looking at Frank's results there it is so clear why running "stock" using Linux is a trap. You keep getting Cuda60 tasks which are like entering a foot race by cutting off one leg ... such a sllooowww app, why do they persevere with it? Sorry just a little peeve I have ... Stephen :( |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
. . Looking at Frank's results there it is so clear why running "stock" using Linux is a trap. You keep getting Cuda60 tasks which are like entering a foot race by cutting off one leg ... such a sllooowww app, why do they persevere with it? Sorry just a little peeve I have ... I agree. Only benefit is someone running older drivers and haven't updated and haven't installed any OpenCL drivers. At least they would get work for the CUDA60 tasks. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I agree. Only benefit is someone running older drivers and haven't updated and haven't installed any OpenCL drivers. At least they would get work for the CUDA60 tasks. . . True, some might not have drivers with OpenCL support ... but what a penalty ... :( Stephen :( |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
Or for those running GPU's that don't support the Special Sauce app.. . Looking at Frank's results there it is so clear why running "stock" using Linux is a trap. You keep getting Cuda60 tasks which are like entering a foot race by cutting off one leg ... such a sllooowww app, why do they persevere with it? Sorry just a little peeve I have ...I agree. Only benefit is someone running older drivers and haven't updated and haven't installed any OpenCL drivers. At least they would get work for the CUDA60 tasks. Cheers. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Or for those running GPU's that don't support the Special Sauce app.. . Looking at Frank's results there it is so clear why running "stock" using Linux is a trap. You keep getting Cuda60 tasks which are like entering a foot race by cutting off one leg ... such a sllooowww app, why do they persevere with it? Sorry just a little peeve I have ...I agree. Only benefit is someone running older drivers and haven't updated and haven't installed any OpenCL drivers. At least they would get work for the CUDA60 tasks. True. Anyone running Nvidia cards with less than 5.0 Compute Capability would still need the CUDA60 app. Unless they dug around the forums for the link to the zi3v special app which can run on CC 3.5 cards. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
©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.