Message boards :
Number crunching :
Setting up Linux to crunch CUDA90 and above for Windows users
Message board moderation
Previous · 1 . . . 124 · 125 · 126 · 127 · 128 · 129 · 130 . . . 162 · Next
Author | Message |
---|---|
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
Is there anything modified for linux for amd cards? |
Wiggo Send message Joined: 24 Jan 00 Posts: 34768 Credit: 261,360,520 RAC: 489 |
No. AMD cards have to rely on OpenCL apps. Nvidia cards can use the OpenCL apps, but do much better using Nvidia's propriety CUDA code. Cheers. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Is there anything modified for linux for amd cards? ? ? ? ? In what respect? Are you asking if there is an special app for AMD cards in Linux? Then the answer is no. AMD cards just have the stock SoG app for gpus. You can set a more aggressive parameter set for AMD just like you can for Nvidia. 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 |
@ TBar . . Hi, I am having problems with the version of BOINC in your AIO distro, ver 7.14.2 (for Linux 14.04). When I try and exit BOINC, I select 'stop runnning tasks' (which to me means terminating the client) but the client continues to crunch WUs. Then when I try to start BoincManager again ... nada ... zip. Nothing happens. To get back into Manager I have to reboot the computer. At that point the client has been terminated and I can then restart manager. I had noticed several 'compute errors' on CPU tasks and now I understand why, they are getting trashed by running when I think they are stopped. . . Oh well, what is life without some hassles? Stephen :( |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
You can always use the Task Manager to Kill the client if you are impatient enough to wait for the process to stop on its own. 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 |
You can always use the Task Manager to Kill the client if you are impatient enough to wait for the process to stop on its own. . . It's not a matter of patience, they simply do not stop at all, ever. They just keep going ... and going ... and going. . . There is some sort of bug in the process ... . . Stephen :( |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
@ TBar If the Kill task trick doesn't work, then your gpu's are hung. I have tried everything suggested and only a system restart cleared the issues under Linux. Try the "nvidia-smi" on Linux terminal. If it comes back with a short error message then one or more is hung. Tom A proud member of the OFA (Old Farts Association). |
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
I found some 280x-2 gpus and 680-2 gpus. What settings should I try? |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
You can always use the Task Manager to Kill the client if you are impatient enough to wait for the process to stop on its own.. . It's not a matter of patience, they simply do not stop at all, ever. They just keep going ... and going ... and going. This was there back in 2012 when I first started using Linux, and it's Still there; Linux x64 Tested on the current Ubuntu distribution; may work on others. I wanted it to work, so I chose to use Ubuntu... it seemed the Logical thing to do. Now 7 Years later, it's still the same and I know a great deal about Ubuntu. Tasks not stopping happened in earlier versions of Mint, it was cured by switching to Ubuntu. Tasks not stopping was also noted when using an older version of BOINC in a Newer Ubuntu, it was cured by using a Newer version of BOINC. Basically, if it won't work in what you are trying, you are probably trying the wrong combination. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
If the Kill task trick doesn't work, then your gpu's are hung. I have tried everything suggested and only a system restart cleared the issues under Linux. . . OK Tom, I will try that ... . . The interesting thing is that we are both running Lubuntu 18.04 ... Stephen ? ? |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
This was there back in 2012 when I first started using Linux, and it's Still there; . . The version of BOINC is the one packaged in your latest AIO, 7.14.2 (from the folder 'for 14.04'). . . The OS is Lubuntu 18.04 which is tantamount to the latest and as I understand it just a cut down version of Ubuntu 18.04, in fact it detects in BOINC as Ubuntu 18.04. . . Is it possible the problem is due to something missing in the Lubuntu package that is necessary for BOINC? I will try the good ol' Linux ldd trick and see if that shows anything. Stephen |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
I guess the next question is why would you use the client clearly labeled “for Ubuntu 14.04â€, when you do not have Ubuntu 14.04? You should be using the default client provided in the package. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I guess the next question is why would you use the client clearly labeled “for Ubuntu 14.04â€, when you do not have Ubuntu 14.04? . . A good question, I will have a look at the package and get back to you ... Stephen ? ? |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I guess the next question is why would you use the client clearly labeled “for Ubuntu 14.04â€, when you do not have Ubuntu 14.04? . . OK, I have reviewed my processes and arrived at several possible theories ... a) . . . I opened the wrong 'readme' file b) . . . I misread the destructions c) . . . While attempting to deal with several issues at once with insufficient sleep I stuffed up d) . . . A combination of any or all of the above ... <shrug> . . For some reason I thought I had to use the 5 critical files from one of those three sub-folders, so the lack of sleep thing seems like a high odds possibility. But I implemented the new platform on 3 rigs, streamlining it on each attempt. On the third and final installation I simply selected all the files in the extracted BOINC directory except the 3 extraneous sub-folders and copied them into the then empty working BOINC directory. So that installation should have had the correct setup from the get go. This problem is present on all 3 rigs. So to try and resolve the question of wrong BOINC version I suspended all tasks on that last machine, exited Boinc Manager 'stopping' the client as usual but again it remained open as did boincmgr in TaskManager though the window itself had closed. So I 'killed' both executables and copied the 5 .exes from the default extraction BOINC folder into the working BOINC folder. I restarted BOINC and it had the same result, exiting failed to end the programs. Just in case I had managed to corrupt or overwrite any of those files in the original extraction folder I repeated the extraction from the archive and copied these fresh versions into the working BOINC directory, same result. So, while I may have stuffed up on the original install on one machine this is not the cause of the issue ... :( . . Tom seems to have his installs sorted out but the fact he experienced similar issues under Lubuntu makes me suspect this could be central to the problem. The trouble is I really like the Lubuntu interface, it feels homey to me, so I would hate to have to abandon it to solve the issue. Stephen |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
This was there back in 2012 when I first started using Linux, and it's Still there; . . So since that version is from Feb 2014 I take it I should NOT be using that ... . . Hmmm :( Stephen ? ? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
. . So since that version is from Feb 2014 I take it I should NOT be using that ...BOINC would work, but be missing modern improvements. You might need to update the security certificate bundle. SETI would NOT work, except with stock apps, because v8 wasn't released until the very end of 2015. |
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
how to mod the boinc to get work,tuesday when seti goes down. I get no work. Some users showing 64gpus. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . So since that version is from Feb 2014 I take it I should NOT be using that ...BOINC would work, but be missing modern improvements. You might need to update the security certificate bundle. . . Thanks Richard ... Stephen . . |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
how to mod the boinc to get work,tuesday when seti goes down. I get no work. Some users showing 64gpus. . . They fiddle the books. Unless you are good at re-writing code for BOINC and then recompiling it you are probably not in the running to do that, which is why I cannot. Though at the rate modern machines can chew through WUs the current levels of work allowed are inadequate even with moderate outages. This weeks outage lasted almost 7 hours which is longer than the norm of late, but 3 of my machines ran through the allocated 100 tasks per GPU well before the outage was over. The minimum increase in reported number of GPUs seems to be a multiplier of 4, but I would be happy even with a multiplier of just 2, though I would probably settle for the 4 :) Stephen . . |
elec999 Send message Joined: 24 Nov 02 Posts: 375 Credit: 416,969,548 RAC: 141 |
how to mod the boinc to get work,tuesday when seti goes down. I get no work. Some users showing 64gpus. How do you multiple by 4 |
©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.