| Author |
Message |
|---|
|
|
|
Hi,
Since two days ago, PPS-sieve (1.39, cuda32) tasks have been taking 3-4 times as much time as before to complete, while giving the same credit. You can see the increase in run time here:
http://www.primegrid.com/results.php?hostid=267690&state=3
Was there something like a 3-4x increase in the workload for PPS-sieve CUDA tasks or could something be wrong on my system slowing down CUDA tasks? This is with the same CUDA driver version (5.0.24) before and after this started happening, on GeForce GT 650M on OSX.
(Btw, this started happening after I upgraded to OSX 10.8.2, but not before completing 8-10 hours worth of tasks as quickly as before on that new OSX version. So, I don't think it's related to the OS upgrade).
Thanks. |
|
|
|
|
|
have you tried rebooting? There was a problem with previous windows drivers downclocking, no idea if the mac drivers suffered from the same problem. |
|
|
|
|
|
Make sure "Power Managment" setting in Nvidia Control Panel is still on "Prefer Maxium Performance" not Adaptive. Mine has mysteriously reset on more than 1 occasion. It has to be set in both of the tabs there. |
|
|
|
|
Make sure "Power Managment" setting in Nvidia Control Panel is still on "Prefer Maxium Performance" not Adaptive. Mine has mysteriously reset on more than 1 occasion. It has to be set in both of the tabs there.
I'm using the gfxCardStatus program (in the "discrete only" mode) to make sure it's always at the "Higher performance" setting. So, that doesn't seem to be the issue. But thanks for the suggestion. |
|
|
|
|
have you tried rebooting? There was a problem with previous windows drivers downclocking, no idea if the mac drivers suffered from the same problem.
Yes, I've tried rebooting (and reinstalling the CUDA driver). Didn't work :) |
|
|
|
|
|
It looks like CUDA here is now overall 3-4x slower than before. I've tried collatz, which is 3x slower than before, and CUDA-Z. GT 650M has 321 Gflops single-precision performance on CUDA-Z with OSX 10.8.1 according to this: http://www.barefeats.com/rogue02.html, while I'm getting 95 Gflops single-precision performance on CUDA-Z with GT 650M & OSX 10.8.2. I'd say it's the OSX 10.8.2 update, but a few PPS-sieve tasks were completed as quickly as before, after updating. This is weird. |
|
|
|
|
|
Are you testing the times with no cpu tasks running? Last week I noticed my pps tasks we running 25% slower than they should and it turned out to be because they wanted a whole core to themselves.
Also have you confirmed that the card's clock speed is actually what it should be? |
|
|
|
|
|
Can you guys post how long in minutes it's now taking on your card.
My 5770 ati is taking 80 to 90 minutes a task.
____________
|
|
|
|
|
|
GTX 570, stock clocks (732), driver 304.43, lubuntu 12.04, IB i7. 16.5 minutes.
--Gary |
|
|
|
|
|
Overclocked (shader at 1860) gtx 460 - 1666 seconds (faster by close too 200 seconds from a few days ago at same o/c)
Overclocked (shader at 1956) gtx 570 - 916 seconds - about the same for a few weeks now.
both systems also running single pps llr core
____________
@AggieThePew
 |
|
|
|
|
|
This is one of mine, again this is a stock ATI 5770, latest drivers. PPS LLR running on 3 cores.
Completed and validated: Runt time: 4,460.23 CPU time: 2,117.15
Do those times look right? They look terrible to me if I'm comparing them to your GTX cards.
I think I'm going to go back to wwwwcl for now.
____________
|
|
|
|
|
This is one of mine, again this is a stock ATI 5770, latest drivers. PPS LLR running on 3 cores.
Completed and validated: Runt time: 4,460.23 CPU time: 2,117.15
Do those times look right? They look terrible to me if I'm comparing them to your GTX cards.
I think I'm going to go back to wwwwcl for now.
Hopefully someone with an ATI card can clarify if your times are good. Unfortunately I have no experience with the ati at all and have only read about their "limited" potential here. (also as my cards are o/c'd very high the results can't be compared apples to apples) |
|
|
|
|
|
To give an update, my CUDA performance on OS X 10.8.2 seems to be back to normal for no apparent reason. Now I'm seeing a single precision performance around 320 Gflop/s on CUDA-Z, compared to 95 Gflop/s earlier. (Btw, yes, I was running the CUDA tasks with no CPU tasks running, and without anything else resource intensive). I'm not sure what caused this large temporary drop in CUDA performance. |
|
|
|
|
To give an update, my CUDA performance on OS X 10.8.2 seems to be back to normal for no apparent reason. Now I'm seeing a single precision performance around 320 Gflop/s on CUDA-Z, compared to 95 Gflop/s earlier. (Btw, yes, I was running the CUDA tasks with no CPU tasks running, and without anything else resource intensive). I'm not sure what caused this large temporary drop in CUDA performance.
Hi OzgurDusunceli, do you remember what you did to make CUDA back to work?
Like did something? Adjusted system preference? Or did nothing just let the Macbook sitting there being idle?
There are many people experiencing this (including me). I just want to get my CUDA back to work, :( |
|
|
|
|
|
Didn't want to start a new thread but are the PPS Sieves now larger than before? Seems like they are taking just a tad longer to complete.
____________
@AggieThePew
|
|
|
|
|
Didn't want to start a new thread but are the PPS Sieves now larger than before? Seems like they are taking just a tad longer to complete.
No significant change during last month on this side.
Did you change drivers? Moving from 2xx to 3xx might cause changes on oc'ed gpus.
____________
676754^262144+1 is prime |
|
|
|
|
Didn't want to start a new thread but are the PPS Sieves now larger than before? Seems like they are taking just a tad longer to complete.
No significant change during last month on this side.
Did you change drivers? Moving from 2xx to 3xx might cause changes on oc'ed gpus.
Nope. Both systems are the same as before but both are running the tasks a little longer so I just wondered. Heck I even followed my own advice and rebooted :)
____________
@AggieThePew
|
|
|
|
|
Can you guys post how long in minutes it's now taking on your card.
My 5770 ati is taking 80 to 90 minutes a task.
====================================================
I did not see any gpu tasks when I checked your computer for any problems?
but as requested I am running 6870's they run almost twice as fast as 5770's
in proth prime sieve I run between 40 and 50 min. roughly twice what you should be able to do.
looks good :)
____________
|
|
|
|
|
|
Hi
580s run about 14.30 mins
590s run about 18.40 mins
570s run about 16.00 mins
660TI run about 11.00 mins
680 run about 10 to 10.30 mins
7970 about 22.50mins
this is all on 3930ks running @ 3.8 with no other cpu work W7 64 16 GIG ram
there has been no increase in time taken, in some cases less with new drivers.
hope this helps
Ross*
____________
|
|
|
|
|
|
the 7970 has 2048 streaming cores
the 5770 only 800
the 5970 3200
the number of shading core is roughly the speed.
my 6870's with 1120 cores will run proth prime sieve is about 40 min.
the 5770 will take about 60 ,min.
____________
|
|
|
 Davina  
Send message
Joined: 13 Feb 12
Posts: 3435
ID: 130544
Credit: 2,773,241,770
RAC: 532,000
                                
|
|
Does BOINC version make a difference? |
|
|
