Folding Forum

hiigaran wrote:C: ill check them next time round. are there specific 'bad' units, or...how do i know?

Note the four numbers that identify a specific WU: Project, Run, Clone, and Gen (also known as PRCG). If you have repeated failurs with the same four numbers, ignore all but the first one. It might be a bad WU. If you have repeated failures with various values of PRCG, it's not the WUs.

If you do suspect a bad WU, first look for a report here: viewforum.php?f=19. If there are no reports, make one using the WU PRCG for a title. The Moderators will generally look your WU up for you.

what i have noticed, which could just be coincidence, is that the usually happen about 10-20 minutes after turning on my computer and opening the clients. not sure if that means anything though.

That probably is how long it takes to heat up the inside of your case, pointing to a air circulation problem.

At those temperatures, I'd be concerned. Either adjust your fans to move more air or underclock the GPU. Either should lower the temperatures. It will probably make everything stable. If it does, you'll know it was heat/power.

Is the GPU the reference design that takes two slots so it can exhaust the hot air out the back of the case or one of the one-slot boards that just recirculates all the heat inside the case?

alright, ill do that with the WU next time.

as for the heating, like i said, if it was a heating issue, i would have been getting EUEs at the higher temperatures that i had, considering that im running it at lower temps now. i guess i shouldnt rule out temperature completely, but the evidence i have doesnt point to heat. i underclocked in the first place because i thought it might have been temperature and/or stability.

as for the cooling, the GPU exhaust is in the front and back of the card, but because i have a custom designed case, this was taken in to account. my case is basically a long, thin one shaped like an aircraft wing, and doubles up as a shelf:



the GPU is parallel to the motherboard, rather than being perpendicular, because of the space issues, but even then, the intake fan has plenty of room, and ive got 7 fans placed strategically within the case, so that the exhaust of the PSU, chipset, and RAM doesnt meet with the GPU, and vice versa. also, i run my fans at full power all the time. youd be amazed at how well that round plastic covering blocks sound (the panels that span the length of the front). so yeah, because of the cooling, and the ducting, i can get some pretty high clocks on my hardware.

oh and yes, i should have taken a photo after fixing the cables, but that was done right after completion!

okay, regarding the unstable machine thing, it looks like its the exact same work unit, and i got two unstables in a row:

[03:51:13] Project: 7621 (Run 811, Clone 0, Gen 85)
[16:07:44] Project: 7621 (Run 811, Clone 0, Gen 85)

so what can i do about this? you said i can report this - where? also, is there a way to exclude these units?

also, if it helps, a few lines up, there are these:

[03:51:10] mdrun_gpu returned 52
[03:51:10] NANs detected on GPU

so could that mean that its still the GPUs fault?

hiigaran wrote:okay, regarding the unstable machine thing, it looks like its the exact same work unit, and i got two unstables in a row:

[03:51:13] Project: 7621 (Run 811, Clone 0, Gen 85)
[16:07:44] Project: 7621 (Run 811, Clone 0, Gen 85)

so what can i do about this? you said i can report this - where? also, is there a way to exclude these units?

At the subforum Bruce mentioned: viewforum.php?f=19 Make a new topic, make the title the name of the WU (like "Project: 7621 (Run 811, Clone 0, Gen 85)" see the the titles for other topics) and then describe your problem with it specifically from there. There's no way on your end to exclude specific WUs from your client, sorry.

okay made a post, and now im trying to get the client to download another unit instead, but it keeps giving me the exact same one. ive deleted the work folder, queue.dat, and even the FAHCore app. every time it downloads again, its the same one.