I think the wu's are fine, I haven't had an errored out one in a long time. Are you sure you cards or fine? Maybe to high on the shaders? I'm running a GTX260, 2 GTS250's, and an 8800gt, all have been working great for months.
Printable View
I think the wu's are fine, I haven't had an errored out one in a long time. Are you sure you cards or fine? Maybe to high on the shaders? I'm running a GTX260, 2 GTS250's, and an 8800gt, all have been working great for months.
4x4n Love your new avatar! That thing is sweet! :up:
I'm not having too many errors either. My 4870 stopped folding, but a reboot of the machine (which hadn't happened for a few months) fixed that.
Thanks coo-coo. My son did it for me, I suck at photoshop :D
still getting errors all the time on either card. Unstable machine NAN errors, cards just stop folding and i have to re-start the client.
[23:08:41] NANs detected on GPU
[23:08:41]
[23:08:41] Folding@home Core Shutdown: UNSTABLE_MACHINE
[23:08:44] CoreStatus = 7A (122)
[23:08:44] Sending work to server
[23:08:44] Project: 5765 (Run 7, Clone 214, Gen 1306)
[23:08:44] - Read packet limit of 540015616... Set to 524286976.
[23:08:44] - Error: Could not get length of results file work/wuresults_05.dat
[23:08:44] - Error: Could not read unit 05 file. Removing from queue.
[23:08:44] EUE limit exceeded. Pausing 24 hours.
not overclocked.
i have these flags Arguments: -gpu 1 -verbosity 9 -local
This is likely related to either a messed up client or your driver might also be borked, if you're really sure the cards are stable at the clocks you've set.
Steps to take:
1. Set up completely fresh work folders (no need to delete the old ones, this is just for testing), with freshly unzipped clients, you can copy the .cfg's
Still getting errors?
2. Check the card's temperatures. Are they fine (around 80°C is ok), nothing clogged up with dust bunnies?
Still errors even though everythings clean?
3. Drop the clocks until you're running without errors.
Still errors even at stock settings or close to that?
4. Uninstall drivers, reboot. Run driver cleaner. Reboot. Install new drivers. Reboot. Setup fresh clients, restart folding at previous stable clocks.
Still errors? There's only two chances now, either your cards degraded or your windows is borked. So either you work your way down until stable or you go the extra mile of excluding windows as source of error and then begin the whole shebang again. :shrug:
Edit: Just noticed that I missed that the card is not overclocked. Take what you need from this post then... :(
Are you turning up your fanspeed? This is something I get if I forget to set the fanspeed (I run 100%) ...try say 75%+ if 100 is too loud
Guys, anybody tried folding on a 5870 yet? Is it worth it?
I might try it.. though milkyway is just running so effordlessly, plus it's madly optimized to utilize all those nice stream shaders the ATIs have over the NVs...
Edit: Nevermind... it doesn't work yet. I got it to start up using the old forcegpu rv700 flag, but it shuts down immediately reporting a "serious error". Lol. Oh well, back to milkyway for now.
I'd be curious how it performs once Stanford gets the support out there for it.
I'm getting my last GTX260-216 that has been sitting around here for a few months up finally tonight! Yay :D
same speed as 4870 from what articles i have read. their gpgpu implementation dates back so far the newer gen inherits their old problems software wise. nvidia didnt have this problem. even the 4870 is cpu bound by 20% so the 5870 is even more bottlenecked. just wait for gpu3. it should bring 4x more ppd for ATi. the main reason they are slow is because the cache is not used as opposed to nvidia. if they had cache in r600 then they would be a lot better off. i am not sure how well molecular dynamics performs on ATi. i cant wait for gpugrid and gpu3 to have ATi cards fully utilized.
http://www.brightsideofnews.com/news...-as-4870!.aspx
i guess you could try forcdgpu ati_r700 like the article says.
There are ways of screwing with a bunch of settings to get the CPU usage down but it is a pain.
been folding for about a day and a half with no errors...maybe its fixed....temps are high 78C with fans at 80%.
how hot does your guys PSU's get? mines so hot on the back it nearly burns my hand to hold it there....and its making a buzzing noise...i sapose this is normal though.
I've run 5-6 cards/sides combinations of gx2 and 98gtx on both 1K SR and UCP 1100 and a single 295 with 4 gig x9650 on a 910 silencer and I'd say that they don't seem to even come out of Idle fan speed and certainly don't seem that hot. Is yours getting enough air, is the fan running up?
Uh, no.
What PSU you have? You really should not leave that thing alone if its doing that - you don't want
http://i189.photobucket.com/albums/z...ning-house.gif
yeah, are you sure the fan is working? my psu under full load is slightly warm and there is no noise coming from it. you dont want to fry your hardware.
My PSUs on my two crunchers get pretty warm. Not burning hot, but definitely heat up the top of the cases.
its a 1,000 watt dell xps 720 H2C PSU, modded to fit a ATX motherboard. The fan is def working, there is a alot of hot air blowing out the back. its working hard powering my GPU's which are folding. Sensors are saying the rpm is about 2500...hopefully there is no massive implosion/explosion in the near future...
If it dies, assuming it doesnt take my whole computer with it, ill just buy a real PSU, its about time anyway
everyone stand back in awe of my amazing PPD
Sparky cant touch dis
Damn fine day yesterday :up:
wow, broke thru the 20K PPD barrier.
http://i174.photobucket.com/albums/w...n222/Vader.jpg
does anyone know how extreme overclocking stats calculates its 24 hours point average? Is it a avg of the last 7 days?