Message boards :
Multicore CPUs :
New QC app
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · Next
| Author | Message |
|---|---|
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
Each BOINC slot with a QC job is about 29 GB, so plan accordingly. I am running two now, which uses all eight cores and still leaves 29 GB free, so I think I am safe for now. |
|
Send message Joined: 9 Dec 08 Posts: 1006 Credit: 5,068,599 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() |
Current test WUs are limited to 30GB, but many are failing because the limit is hit. I will need to raise the limit (the production QC app is 60 GB). |
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
I will need to raise the limit (the production QC app is 60 GB). I will put in a 256 GB SSD, but I have a 500 GB sitting on the shelf too. Let us know what you need. |
|
Send message Joined: 8 Apr 18 Posts: 21 Credit: 1,309,700 RAC: 0 Level ![]() Scientific publications
|
Two tasks have "restarted" themself and started run from th beginning when they achieved about 40%. Just before it my CPU has heated suddenly to over 80 Celsius degrees, fortunately only for a moment. |
|
Send message Joined: 12 Jul 17 Posts: 404 Credit: 17,412,649,587 RAC: 32 Level ![]() Scientific publications ![]() ![]()
|
The default BOINC settings for disk memory are far too low for your experiment. The RAM requirements are all over the map. You're sending these jobs to Xeons with 24 to 44 logical CPUs that can only run one or two at a time. Why don't these WUs respect the client's resources??? E.g., if I have 16 GB RAM you'll start 5 or 6 WUs and consume all RAM and the computer freezes. If I have an app_config.xml set limiting QC or QC_beta to one WU (4C needs 16 GB RAM) then you want 60 GB of disk memory which is fine if I've set BOINC high enough. Did you think this through before you launched this mayhem??? You should be able to tell us what these requirements are before launching. 4 GB per CPU thread per WU = 16 GB RAM is 4 times higher than Rosetta and by far the highest I've seen. This is the first time in 5 hours I've been able to access this website. |
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
The default BOINC settings for disk memory are far too low for your experiment. Good point. I don't remember what the defaults are, but I always increase them anyway. They should probably create a sticky on how to set up BOINC for this project. Also, a sample app_config.xml file on how to limit the number of work units operating at a time would be helpful. Some projects allow you to limit the number of work units downloaded at at time (and even the number of cores to be used per work unit) on their preferences page. That would be even easier. I have trouble connecting here too. It seems to be a problem for everyone outside of Spain, but especially in the U.S. |
|
Send message Joined: 21 Mar 16 Posts: 513 Credit: 4,673,458,277 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
My windows machine is doing QC WUs, is this supposed to happen? |
|
Send message Joined: 9 May 13 Posts: 171 Credit: 4,739,796,466 RAC: 1,182 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
PappaLito said: My windows machine is doing QC WUs, is this supposed to happen? Could be, I also have 2 running on a Windows machine. |
|
Send message Joined: 26 Feb 14 Posts: 211 Credit: 4,496,324,562 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
So about half are failing the other half are finishing. They are using all of the threads on my CPUs. I had a max concurrent in my app_config but I think these are a different name so that limitation is being ignored. I should have tried a <project_max_concurrent> but didn't have time to play with it to see if it would restrict the number running. Guess that will have to wait until Saturday when I can re-access those machines.
|
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
So about half are failing the other half are finishing. They are using all of the threads on my CPUs. Even with a limit on the number of work units running, some of them will fail anyway with a "196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error. They need more disk memory than they are allowed at the moment, as Toni mentioned below. |
|
Send message Joined: 5 Mar 13 Posts: 348 Credit: 0 RAC: 0 Level ![]() Scientific publications ![]() |
Sorry to hijack again this thread. I might make a new one but since I made here the original post: the current WUs I sent to the QC app (not the beta) should run fine. They are the same molecules as before. I changed my mind about submitting the other monsters, I'll try to run them on our local cluster first. |
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
I changed my mind about submitting the other monsters, I'll try to run them on our local cluster first. Well I was just about to put in a 500 GB SSD, but I will wait until you give us the heads up. |
|
Send message Joined: 10 Sep 10 Posts: 164 Credit: 388,132 RAC: 0 Level ![]() Scientific publications
|
After some times i restart my linux box and.... 226 mb of upload of a "Toni" quantum chemistry wu!! Ouch |
|
Send message Joined: 12 Jul 17 Posts: 404 Credit: 17,412,649,587 RAC: 32 Level ![]() Scientific publications ![]() ![]()
|
The QC per core RAM load has dropped. What's today's requirement??? |
|
Send message Joined: 18 Sep 08 Posts: 368 Credit: 4,174,624,885 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
What are the requirements to receive Regular Non BETA QC Wu's ??? I have never received 1 even though I have been calling for them all along. I can get the BETA QC Wu's so one would think I should be able get the Regular QC Wu's. Running Windows 10 Pro with 64gb's Memory on a Intel i7 7770k CPU. When my Box calls for Wu's I get this: PBOYZTOYNP9873 13027 GPUGRID 1/12/2019 4:35:13 AM No tasks are available for Quantum Chemistry Server Status shows Quantum Chemistry 125,158 Unsent ... Thanks STE\/E |
|
Send message Joined: 12 Jul 17 Posts: 404 Credit: 17,412,649,587 RAC: 32 Level ![]() Scientific publications ![]() ![]()
|
STEVE, Have you gone to BOINCmgr/Options/Computing Preferences/Disk and Memory??? The combination of those 3 checkboxes under Disk must yield over 60 GB of free disk space. May be 60 GB per QC WU. In my case I was getting QCs but they seem to have stopped over night. No idea why. How many months will it take to clear 125,000 WUs??? |
|
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
The combination of those 3 checkboxes under Disk must yield over 60 GB of free disk space. May be 60 GB per QC WU. As I read Stefan's post, they have backed off from the large ones, and have gone back to the earlier ones. So 30 GB (per work unit) should do; that is what I have been getting. http://www.gpugrid.net/forum_thread.php?id=4785&nowrap=true#51272 It is entirely possible that they will go to the big ones (or even bigger) in the future, unless they can figure out how to break up the work into smaller pieces. |
|
Send message Joined: 18 Sep 08 Posts: 368 Credit: 4,174,624,885 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
STEVE, Have you gone to BOINCmgr/Options/Computing Preferences/Disk and Memory??? The combination of those 3 checkboxes under Disk must yield over 60 GB of free disk space. May be 60 GB per QC WU. I have a 1 Terabyte HD that I run Boinc on, there is 878gb of free space on it, there should be more than enough space to run the Wu's. Disk: use at most 750 GB Disk: leave free at least 1 GB Values smaller than 0.001 are ignored GB Disk: use at most 100 % of total Tasks checkpoint to disk at most every 60 seconds Swap space: use at most 95% of total Memory: when computer is in use, use at most 95% of total Memory: when computer is not in use, use at most 98% of total STE\/E |
|
Send message Joined: 9 May 13 Posts: 171 Credit: 4,739,796,466 RAC: 1,182 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
STEVE asked, What are the requirements to receive Regular Non BETA QC Wu's ??? As far as I know, regular non BETA QC WU's are still Linux only. I am getting plenty on my Linux box, none on a Windows box. |
|
Send message Joined: 18 Sep 08 Posts: 368 Credit: 4,174,624,885 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]()
|
STEVE asked, Yeah that's what I was afraid of, guess they don't need the Wu's run very fast if it's Linux only. Why would they test the BETA Wu's on Windows Machines but not run the Regular Wu's on them, weird ... STE\/E |
©2026 Universitat Pompeu Fabra