Posts by Steve

1) Message boards : News : Moved to New Server (Message 62279) Posted 26 Mar 2025 by Steve Post: Hello GPUGRID Community, We have recently moved to a new server! This transition will enhance the capabilities of the project. We are currently redeploying applications and addressing some bugs that have emerged. Your patience and support during this process are greatly appreciated. For quicker responses from the team, please check out our Discord channel: https://gpugrid.net/gpugrid/forum_thread.php?id=5435 . We also continue to monitor this forum, so both avenues of communication remain open. Thank you for your continued contributions to GPUGRID! Best regards, The GPUGRID Team
2) Message boards : Frequently Asked Questions (FAQ) : No tasks in quite a while. (Message 62242) Posted 6 Mar 2025 by Steve Post: My test WUs on the new server have revealed a new bug (somehow caused by the combination of the newer server code and importing a very old database schema) in the scheduler that is causing these stuck tasks. Next week I should have time to fix it. Until then we cannot send new tasks.
3) Message boards : Number crunching : new error requesting tasks (Message 62226) Posted 26 Feb 2025 by Steve Post: yes those 3 are old and stuck
4) Message boards : Number crunching : new error requesting tasks (Message 62221) Posted 25 Feb 2025 by Steve Post: Hi, yes the server is back but various things have changed, including some of these internal urls/paths. We are still testing. - If you re-attach to the project these should be fixed (they are set by the get_project_config.php) - I will look into setting up some redirects from the old paths to the new ones
5) Message boards : Frequently Asked Questions (FAQ) : Boinc 8.0.2. (64bit) (Message 62218) Posted 25 Feb 2025 by Steve Post: We are in the process of a server move and upgrade. The current URL is https://www.gpugrid.net/ the old http (no s) will no longer work. Maybe the .org version will be redirected to the .net but currently it is not the correct url.
6) Message boards : Number crunching : Quantum chemistry calculations on GPU (Message 61815) Posted 17 Sep 2024 by Steve Post: Yes that temp dir used by cupy should be located at $HOME/.cupy by default as I mentioned here: https://github.com/BOINC/boinc/discussions/5811#discussioncomment-10670615 And as shown by Ian’s path. I don’t quite know why it would be different for your setup. Maybe a side effect of the new client version you have installed combined with our older wrapper script. Although as mentioned by the BOINC developers it is better if this runtime folder is instead located in the boinc slot or project directory. I will add this change to the next app update.
7) Message boards : Number crunching : ATMML (Message 61800) Posted 12 Sep 2024 by Steve Post: I think a P5000 should work, given that it is the same generation as a 1080 which I have confirmed to work. It may be that your drivers are too old. I have a 1080 with driver version 536.
8) Message boards : Number crunching : ATMML (Message 61776) Posted 5 Sep 2024 by Steve Post: This is very interesting, thank you for the numbers. I still don't understand where the flops number for a machine comes from. does it use the data of your hardware? or is it purely based on maths done from the rsc_fpops_est number we have set and the time taken for WUs? I am also unsure how I would set this rsc_fpops_est number to be more accurate. given one of these WUs takes maybe an hour on a 4090: A 4090 is 80 TFLOPS. x1 hour = ~ 3x10^17 float point operations. Which is not actually that far off the estimated value of 1x10^18. of course the WUs will not be using all the Tflops of the 4090. And there is no sane way for me to calculate the number of floating point operations the program uses.
9) Message boards : Number crunching : Server is out of disk space (Message 61734) Posted 26 Aug 2024 by Steve Post: Thank you for reporting. It is now back and with more space.
10) Message boards : Number crunching : ATMML (Message 61732) Posted 26 Aug 2024 by Steve Post: Thank you for the explanation. The time limit exceeded error therefore happened because: - we had a bug in some circulating WUs where certain errors would not trigger a proper error code. The result would then be validated with short runtimes. - these fast runtime results then skewed the correction factors for the newly released windows app version. To fix the problem I 10x'ed the rsc_fpops_bound value while leaving the rsc_fpops_est unchanged. This appears to have worked and hosts that previously had the time limit exceeded errors now do not.
11) Message boards : Number crunching : ATMML (Message 61709) Posted 24 Aug 2024 by Steve Post: Hello. I apologise about the time limit exceed errors. I did not expect this. The jobs run for the same time as the linux ones that have all been working so I dont really understand what is happening. Unfortunately the way boinc deals with "runtime" is completely inadequate for gpu projects. In a WU we have to estimate the flop use, which is a difficult thing to do for a gpu app. The boinc client then somehow estimates the flops performance of your computer in a way I don't understand. I cannot simply put a runtime limit of x hours as would be typical. Does anyone know where the denominator comes from in this line?: <message> exceeded elapsed time limit 5454.20 (10000000000.00G/1712015.37G)</message> <stderr_txt> The numerator I believe is the fpops_bound that is set in the WU template which is controlled by us.
12) Message boards : Number crunching : ATMML (Message 61701) Posted 23 Aug 2024 by Steve Post: Hi, the windows host are working successfully. There are more errors than on linux as expected, but plenty are working well. Unfortunately some WUs with the very short run time but validated status bug are still in circulation. (each WU runs in a chain of 5 steps, when a step finishes it launches a new job with the same settings.) New WUs do not have this bug. This is the bug I am talking about: https://www.gpugrid.net/forum_thread.php?id=5468&nowrap=true#61682
13) Message boards : Number crunching : Strange host (Message 61682) Posted 20 Aug 2024 by Steve Post: Hello. We have identified the problem and it has been fixed in our code. The next round of WUs should not have this problem. This is not any sort of hack. It is just a case of some specific error types (that occur with old GPUs we do not have available locally to test on) not raising a proper error code and slipping through the validation.
14) Message boards : Number crunching : Strange host (Message 61679) Posted 19 Aug 2024 by Steve Post: Hello, we are looking into this! thanks
15) Message boards : Frequently Asked Questions (FAQ) : for FIVE years now GpuGrid violates EU regs about owners of the personal data being able to access, modify,it. (Message 61658) Posted 8 Aug 2024 by Steve Post: Hello. Yes profiles can no longer be created. This part of the server software is broken. If you wish to delete your profile can you please try this page of the website: https://gpugrid.net/delete_profile.php The BOINC website user initiated account delete method is not supported in the current server software version. If you wish to delete an account it will need to be done by a GPUGRID admin.
16) Message boards : News : ACEMD updated app (Message 61644) Posted 28 Jul 2024 by Steve Post: It is now fixed, one of the files was corrupted.
17) Message boards : News : ACEMD updated app (Message 61643) Posted 28 Jul 2024 by Steve Post: Hello, it does appear the the Windows app version 2.28 is now broken. 2.27 worked. We are investigating.
18) Message boards : Number crunching : ATMML (Message 61634) Posted 25 Jul 2024 by Steve Post: Hello. This app is actually built with cudatoolkit version 11.8 For reasons indicated here: https://docs.nvidia.com/deploy/cuda-compatibility/#application-considerations-for-minor-version-compatibility The minimum driver version is not 450.80.02 as stated in table 1 of that link. But most likely 520 that was released with 11.8: https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html#id5 This app used OpenMM which uses PTX code. Hence if you are using a too old driver you will see the error: CUDA_ERROR_UNSUPPORTED_PTX_VERSION (222) For reference my test machine is a GTX 1080 with driver version 545
19) Message boards : Number crunching : ATMML (Message 61585) Posted 6 Jul 2024 by Steve Post: This is an app in testing mode, it does not appear as one to select yet. You will only get the WUs if you have selected to run the test applications. It is a different version of the existing ATM app that includes machine learning based forcefields for the molecular dynamics.
20) Message boards : News : PYSCFbeta: Quantum chemistry calculations on GPU (Message 61245) Posted 12 Feb 2024 by Steve Post: Hello, Yes I would not expect the app to work on WSL. There are many linux specific libraries in the packaged python environent that is the "app". Thank you for the feedback regarding the faliure rate. As I mentioned different WUs require different memory use that is hard to check before they start crunching. From my viewpoint the failiure rates are low enough that all WUs seem to suceed with a few retries. This is still a "Beta" app. We definitely want a Windows app and it is in pipeline. However, as I mentioned before the development of this is time consuming. Several of the underlying code-bases are linux only at the moment so a windows app requires a windows port of some code.

Next 20