Message boards :
Number crunching :
NVIDIA Driver 388.13 Crashing/Recovering...
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
The most obvious thing to me would be to go back to crunching Seti and see if the problem shows up. Thank you for that. And, yes, Collatz crunching IS quite different than Einstein or SETI! MUCH more aggressive than ANY Project I've run in the past. The fact that the Driver Crash ONLY SEEMS to happen during the daytime hours, (Non-crunching Hours), leads me to believe that just switching Projects back to SETI won't reveal anything at this time. I'm beginning to believe that 388.13 itself may be a faulty Driver for my 1050s under Win7 Pro x64... Or, that the 1050 listed as Dev 0 in BOINC, (attached to my monitor), has somehow been damaged by Collatz AFTER Optimization with the Original Default of "sieve_size=30"... I had good luck with 353.30 on Old Prometheus. (GA-EP45-UD3P MOBO with Intel Quad Core Extreme - QX9650 at 3GHz...) [New Hardware List for New Prometheus:] Intel i7 7700K @ 4.2GHz - New CoolerMaster Hyper212 EVO - 4+ Years Old Gigabyte GA-Z270-HD3 - New 32GB Corsair Vengeance DDR4 RAM, (4x8GB Sticks), 2400MHz - New Lite-On ATAPI DVD Burner - 4+ Years Old SilverStone FS303B Hot Swap Bay - 2+ Years Old, went in when Hackintosh-Andromeda was created. WD Black SATA 1TB Hard Drive - 4+ Years Old TWO EVGA GTX-1050 2GB GDDR5 VRAM Cards - New Corsair HX750i Platinum Rated 750Watt PSU - New Antec Mid Tower ATX Case - 4+ Years Old TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
i mean try even LESS aggressive settings than you're using now.
even brand new hardware can be bad. I'm not saying that this is your issue right now, but if you check everything else off this list... at that point, you don't necessarily have to buy new hardware, but you you truly think a component is defective, your option is usually to: a) send it back to the manufacturer for warranty repair/replacement (you'll pay shipping b) deal with it until it fails completely in any case, i think you should try at least some of the easier checks first to rule things out and help narrow down the list of possible problems. and kittyman is absolutely right. different tasks will load up your hardware in different ways so a different app might expose issues that wont show up with another. you could run SETI all day long no problem, but run into issues trying to run an intense 3D application like FurMark, just for example. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
For those interested in comparing Original Collatz Optimization Specs: Original Collatz Optimizations: verbose=1 kernels_per_reduction=48 threads=8 lut_size=16 sieve_size=30 reduce_cpu=0 cache_sieve=1 sleep=1 These were running for about 4-5 Days before I noticed the Sluggishness in trying to use the System while crunching. STILL DIDN'T see Video Driver Crashing at this point. Started adjusting these settings a couple days ago, (before getting this Thread created), and found the "Sweet Spot" Settings Posted earlier. After getting the "Sweet Spot" Settings, THEN during Non-Crunching Hours started having Driver Crashes. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
i wouldn't immediately think the GPU was DAMAGED by running collatz. that's pretty unlikely unless the GPU was defective to begin with. new GPUs have lots of preventative measures built in that aim to specifically prevent damage to the core by the workload, usually in the form of throttling the card if it exceeds current, power, or thermal limits. are you planning to try to install the 353.30 drivers? when searching for the older drivers, the oldest ones i can find from nvidia's site for the 1050 is the 385 driver from 1 year ago. I'm not sure if they purge drivers older than a year or something. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
I do plan on implementing some of the Fixes you've Posted... Just need to take them one step at a time. As just recently state, YES, I agree that different Projects and Tasks run VERY differently than SETI. It seems that Collatz is VERY aggressive in what they do and how they do things... If a different Video Driver doesn't yield positive results, I will switch the GPU Device Positions and see what happens there. Temps do NOT seem to be an issue... Non-Crunching Temps seem to hover Less Than 40 C, and Crunching Temps Less Than 60 C for CPU... Non-Crunching GPU Temps are Less Than 30 C, and Crunching Temps are Less Than 70 C. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I am not a GPU expert. I have a GTX 1050 Ti installed in my Windows 10 PC with its A10-6700 AMD CPU, not overclocked. I have the following results; Einstein@home GPU tasks run peacefully, with a low error percentage. SETI@home GPU tasks run also, but cause frequent reboots, but the tasks complete and validate,no errors. GPUGRID GPU tasks run, but the GPU reaches 80 C and then the task crashes, without causing a reboot. So I am mostly running Einstein@home tasks on it, leaving SETI@home and GPUGRID GPU tasks to my Linux box with its GTX 750 Ti which never goes above 70 C. Hi, Did you ever install TThrottle on your Windows box? I remember you posting the exact same info above but no report on getting your temperature under control using robust temp control software. Respectfully, Tom A proud member of the OFA (Old Farts Association). |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
I am not a GPU expert. I have a GTX 1050 Ti installed in my Windows 10 PC with its A10-6700 AMD CPU, not overclocked. I have the following results; Einstein@home GPU tasks run peacefully, with a low error percentage. SETI@home GPU tasks run also, but cause frequent reboots, but the tasks complete and validate,no errors. GPUGRID GPU tasks run, but the GPU reaches 80 C and then the task crashes, without causing a reboot. So I am mostly running Einstein@home tasks on it, leaving SETI@home and GPUGRID GPU tasks to my Linux box with its GTX 750 Ti which never goes above 70 C. The Temps, with Hyper212, Non-Crunching are UNDER 40 C, and Crunching Under 60 C... NO NEED for TThrottle... I ROUTINELY monitor ALL System Temps and Specs through CPUID Hardware Monitor. TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
On the Linux box with GTX 750 Ti the SETI@home GPU tasks reach 54 C, the GPUGRID tasks 61 C. Einstein@home GPU tasks on the Windows 10 PC reach 64 C. What is the problem on it are the Windows 10 reboots I check on the Maintenance/Reliability Monitor and all reboots happen only when SETI@home GPU tasks are running. Tullio |
Cliff Harding Send message Joined: 18 Aug 99 Posts: 1432 Credit: 110,967,840 RAC: 67 |
I have one very important question -- why are you running GTX 1050s under 388.13, which is almost a year old (10 Oct, 2017)? I venture that if you upgrade your driver to at least 391.01 (26 Feb, 2018) you will have better success. I don't buy computers, I build them!! |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13732 Credit: 208,696,464 RAC: 304 |
I have one very important question -- why are you running GTX 1050s under 388.13, which is almost a year old (10 Oct, 2017)? I venture that if you upgrade your driver to at least 391.01 (26 Feb, 2018) you will have better success. And unless he's running games, or a particular programme that the new driver addresses an issue for, I doubt it would make any difference. My GTX 750Tis were running quite happily on the original supporting release drivers on my old Vista system. They had to use a more recent driver when they moved to new hardware & Win10. No change in processing times, or system reliability. Grant Darwin NT |
TimeLord04 Send message Joined: 9 Mar 06 Posts: 21140 Credit: 33,933,039 RAC: 23 |
I have one very important question -- why are you running GTX 1050s under 388.13, which is almost a year old (10 Oct, 2017)? I venture that if you upgrade your driver to at least 391.01 (26 Feb, 2018) you will have better success. Well, 388.13 is exactly what was given on the Install DVD contained in the EVGA Packaging. Keep in mind that I bought the system and GPUs piece by piece, and that I started this endeavor 5 Months ago! Also, as Grant said: "And unless he's running games, or a particular programme that the new driver addresses an issue for, I doubt it would make any difference." For games, I run Blizzard's Battle.net App and run StarCraft:Remastered, StarCraft II, HoTS, and Hearthstone... I also run GOG.com's Galaxy Launcher and therein run 7th Guest, 11th Hour, and Wing Commander III. NONE of the above games requires the latest Driver Set, and since BOINC DOES NOT require the Latest Cutting Edge Drivers to work... :p TL TimeLord04 Have TARDIS, will travel... Come along K-9! Join Calm Chaos |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
I am not a gamer. What I suspect is that the BOINC Screen saver is causing some crashes on my Windows 10 PC.While running SETI@home GPU tasks, GPU-Z gives 67 C temp, 45% fan, 98% GPU load. Tullio |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
@tullio I am not a gamer. What I suspect is that the BOINC Screen saver is causing some crashes on my Windows 10 PC.While running SETI@home GPU tasks, GPU-Z gives 67 C temp, 45% fan, 98% GPU load. I remember Windows 10 running the screen saver used to screwup the processing but I don't remember the people having that problem saying it was crashing things. This was the creator edition, a year ago last fall, I think. Does the problem go away if you disable the screen saver in Windows? Are you running any commandlines files or app_config.xml files in the Seti dir? If yes, could you post the contents? Thank you. Tom A proud member of the OFA (Old Farts Association). |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
I am a Windows novice, since I use mostly Linux after having used UNIX in my professional life. So all my Windows parameters are default. I cannot but accept the Windows upgrades, and get my nVidia drivers via Geforce. Tullio |
©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.