Message boards :
Number crunching :
SETI@home v8 v8.22 (opencl_nvidia_sah) x86_64-pc-linux-gnu Errors
Message board moderation
Author | Message |
---|---|
Chris904395093209d Send message Joined: 1 Jan 01 Posts: 112 Credit: 29,923,129 RAC: 6 |
Any idea why these tasks are erroring out? I noticed 1 PC had a bunch (30 or so tasks) going to waiting/postponed status and kept going from 1 task to the next. I manually aborted them before the weekly outage. But now that we are back live, I have noticed a couple other Ubuntu machines with the same GPU app having the same issues - but they went in to error status automatically. ~Chris |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Hmmm, GeForce GT 710 (2001MB). That's a pretty low end card, maybe that's the problem. I haven't heard of anyone having that trouble with the stock app; ERROR: OpenCL kernel/call 'clGetDeviceIDs (second call)' call failed (-32) in file ../../src/GPU_lock.cpp near line 1311. If you don't mind trying Anonymous platform there are a couple different Apps you can try, both come bundled with a SSSE3 CPU App that should work with your CPU as well. The OpenCL version nVidia_OpenCL_r3567.zip was compiled using the iGPU path and should work better on the Low End GPU. The CUDA 8 Special App Should work on a 710 as it's listed as a CC 3.5 GPU. I don't think anyone has tried the Special App on that GPU though. I would suggest testing them Offline in the Benchmark App before trying them on SETI Main. If it works on that GPU, the CUDA 8 Special App would probably work best. You will need to Download the CUDA 8.0 Libraries to use the CUDA App, and those files are Not included in the Download. The Apps are both here; http://www.arkayn.us/forum/index.php?topic=197.msg4497#msg4497 The Benchmark App is here; http://lunatics.kwsn.info/index.php?action=downloads;sa=view;down=360 |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
It seems you're receiving a similar Error with the Stock CUDA App; Cuda error 'Couldn't get cuda device count ' in file 'cuda/cudaAcceleration.cu' in line 138 : no CUDA-capable device is detected. It appears you might be suffering from the Startup Bug where BOINC looks for the Device Before the Driver has finished loading. The people with this problem usually install a Startup Script that delays BOINC from starting until the Driver has loaded. A Quick check would be to Stop BOINC and then Launch it again to see if the Device is found. If you Stop/Start BOINC after Bootup and the problem is solved that is probably the trouble. http://setiathome.berkeley.edu/forum_thread.php?id=76659&postid=1636763#1636763 |
petri33 Send message Joined: 6 Jun 02 Posts: 1668 Credit: 623,086,772 RAC: 156 |
Hi, Is it possible that you are running on default graphics driver? Try installing a driver from www.nvidia.com or use the driver manager on desktop and select a NVIDIA driver instead of the free public one. Just my thoughts. Petri 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 |
Chris904395093209d Send message Joined: 1 Jan 01 Posts: 112 Credit: 29,923,129 RAC: 6 |
I'll check my 3 Ubuntu PCs to make sure they are still using the NVIDIA driver. I'm pretty sure I switched them from the default driver to NVIDIA, but who knows what may have happened since these PCs just sit in a corner crunching away. BOINC could be starting too fast as well. I'll explore that option. One of my PCs is having issues downloading new GPU tasks after I rebooted it a few minutes ago so I'll check this one out by stopping BOINC and restarting. https://setiathome.berkeley.edu/results.php?hostid=8188521 This weekend I will look into switching over my Xeon workstation over to Anonymous platform since I'll get more bang for the buck out of this box. Thanks, Chris ~Chris |
tazzduke Send message Joined: 15 Sep 07 Posts: 190 Credit: 28,269,068 RAC: 5 |
Greetings A valuable lesson I learnt ages ago and that was to add boinc to the startup applications and then edit to add a time delay, my preference is 60. Running two Linux boxes One has 2 x GTX960's running the Petri Cuda Special App (Thankyou TBAR for the post about the Xorg.Conf file) One has 1 x GTX780 running the OpenCL app Both are on Mint Linux 18..1. Just my 2cents but Happy Crunching Mark |
Andrew Shill Send message Joined: 19 Aug 99 Posts: 1 Credit: 10,910 RAC: 0 |
I got tired of the problems and got rid of it. |
©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.