• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

nVidia Project 5900

A

APOLLO

[H]ard|DCer of the Month - March 2009
Joined
Sep 17, 2000
Messages
9,089
This new nVidia WU is greatly affecting the performance of my system. Has anyone else noticed this? It's like a throwback to the initial WUs that caused a lot of video and other slowdowns, especially with XP. I also noticed that a new core was downloaded to process it and likely the reason for the issues.
 
what drivers are you using? because i get video lag with anything that uses cuda with 180.xx or newer drivers.. ever since 178.24 ive yet to find a driver set that hasnt caused video lag with cuda..
 
I'm using driver version 177.89, so that would be older than 178.24. You know what? I'm working on two 5900s at this writing and the lags have gone as far as I can tell. Maybe its was just the first crop of 5900s?

Since were talking about drivers, which version out of all of them so far released is producing the highest PPD? I remember there was mention of v178.24 in another thread, is this still true?
 
FYI, this is the new projects using FahCore_14 which include a different way to do scientifical computation and checks. More informations can be found at http://foldingforum.org/viewtopic.php?f=52&t=8734 and here's the summary :

ihaque said:
We've just released project 5900 to advmethods. It uses a new core, FahCore_14.

Please note, the behavior of this core is different from what you may expect. The following behaviors are NORMAL, and the points credit takes it into consideration:

Varying time per frame
Low GPU temperatures
Fluctuating CPU usage (somewhat higher than GPU core 11).

Preferred deadline: 2 days
Final deadline: 3 days
Stats credit: 420 points/WU
Click to expand...
 
So more importantly - what ppd are you guys seeing with these? :D
 
First thing I noticed was my GPU blowers were changing their volume levels. That's what got me to take a look at what was going on. They used to hum loudly, almost a whine; running around 80-82C on both GPU's. Now the blowers go up and down instead of a constant hum, so these new WU's do run cooler. I don't know the current temps, I had a bad experience with SpeedFan locking my system; is there a temp monitoring app that anyone could suggest?

Some WU timings from my log files:

##########
##########

5758 -gpu1
21:43:09 ---> 00:28:50 ( 9941s )

5759 -gpu1
16:11:11 ---> 18:55:01 ( 9830s )

5761 -gpu1
18:57:17 ---> 21:40:48 ( 9811s )

5766 -gpu1
08:51:08 ---> 10:59:49 ( 7721s )

5900 -gpu1
17:11:48 ---> 20:28:59 ( 11771s )
20:30:50 ---> 23:46:41 ( 11751s )
23:48:40 ---> 03:04:49 ( 11769s )

##########
##########

5751 -gpu0
12:32:49 ---> 17:38:04 ( 18315s ) (Doubled checked the timings on this one, was this a longer WU in general?)

5761 -gpu0
17:40:22 ---> 20:23:59 ( 9817s )

5769 -gpu0
10:20:23 ---> 12:29:11 ( 7728s )

5900 -gpu0
16:40:35 ---> 19:58:42 ( 11887s )
20:00:33 ---> 23:19:09 ( 11916s )
23:21:06 ---> 02:39:29 ( 11903s )

##########
##########

I checked the CPU Time in taskmgr, it doesn't seem to be any higher with Core_14. I haven't been keeping close attention to the CPU time, but the 4 Core_a1's are sitting at a little over 20h each atm, Core_14's are sitting at 6s and 10s. RAM usage is sitting around 66MB to 67MB with Core_14.
Running with 182.06 drivers.


_
 
OK, I'm still getting problems with the P5900s but it's not video related. The mouse just stops responding and overall system responsiveness slows down. I'm forced to Ctrl+Alt+Del more often than I care to count, whenever I work on my computer and one of these are being processed. So, is there any way to prevent receiving them?
 
its the drivers.. its been an issue since the 180.xx drivers.. its why i no longer do anything thats cuda related.. i quit using seti, quit folding.. ive pretty much lost all hope for nvidia's cuda drivers..
 
Gonna bump this. Anyone seeing the 5902 WU worth 1680 pts? Currently working on a few of them. Running 4-4.5 minute frames on 280GTX stock clocks.

 
SatTech said:
Gonna bump this. Anyone seeing the 5902 WU worth 1680 pts? Currently working on a few of them. Running 4-4.5 minute frames on 280GTX stock clocks.
Click to expand...
Yeah, I have all my nVidia clients working on them. Very slow, but worth a substantial amount. They seem less problematic with my system's performance and hope this trend continues.
 
Got em running on all 11 of my GPU clients. Nice PPD, that's for sure.
 
I've got 1 running on a 280GTX for 7000+ PPD. The other 280 GTX is running one ar 5000+ PPD. They really do vary wildly depending on R,C & G. I haven't looked too closely at the 8 running on my 9800GX2's, but they show a pretty wide spectrum of PPD.
 
JesterOI said:
First thing I noticed was my GPU blowers were changing their volume levels. That's what got me to take a look at what was going on. They used to hum loudly, almost a whine; running around 80-82C on both GPU's. Now the blowers go up and down instead of a constant hum, so these new WU's do run cooler. I don't know the current temps, I had a bad experience with SpeedFan locking my system; is there a temp monitoring app that anyone could suggest?

Some WU timings from my log files:

##########
##########

5758 -gpu1
21:43:09 ---> 00:28:50 ( 9941s )

5759 -gpu1
16:11:11 ---> 18:55:01 ( 9830s )

5761 -gpu1
18:57:17 ---> 21:40:48 ( 9811s )

5766 -gpu1
08:51:08 ---> 10:59:49 ( 7721s )

5900 -gpu1
17:11:48 ---> 20:28:59 ( 11771s )
20:30:50 ---> 23:46:41 ( 11751s )
23:48:40 ---> 03:04:49 ( 11769s )

##########
##########

5751 -gpu0
12:32:49 ---> 17:38:04 ( 18315s ) (Doubled checked the timings on this one, was this a longer WU in general?)

5761 -gpu0
17:40:22 ---> 20:23:59 ( 9817s )

5769 -gpu0
10:20:23 ---> 12:29:11 ( 7728s )

5900 -gpu0
16:40:35 ---> 19:58:42 ( 11887s )
20:00:33 ---> 23:19:09 ( 11916s )
23:21:06 ---> 02:39:29 ( 11903s )

##########
##########

I checked the CPU Time in taskmgr, it doesn't seem to be any higher with Core_14. I haven't been keeping close attention to the CPU time, but the 4 Core_a1's are sitting at a little over 20h each atm, Core_14's are sitting at 6s and 10s. RAM usage is sitting around 66MB to 67MB with Core_14.
Running with 182.06 drivers.


_
Click to expand...

http://www.cpuid.com/hwmonitor.php
 
I'm seeing...and hearing very fluctuating load on my GTX 280 with the 59XX WUs. The card seems to cycle between being for about 2-3 seconds or so, then gets full load for about 1-2 seconds. I can see that in RivaTuner by looking at the GPU temps bouncing up and down and I can hear the card squeal when it's put under full load. Kind of weird, but the card's cranking out 7-8k PPD, depending on the WU, so works for me!

Oh, and no desktop performance issues for me. I'm running the 181.22 drivers in Vista 64.
 
Thanks osrk!

The 5900 was taking ~2min/frame for me; the 5902 is taking ~8min/frame. So the increase from 420 to 1680 looks to be the same ppd.
Actual ppd might be a sliver higher due to more sequential computing time per WU. Faster WU's are uploaded and downloaded more often and cycles are lost during this downtime between WU's.

The rise and fall of the fans during the 59XX WU's is normal. I didn't really notice a difference with my GPU blowers. I thought so at first, but it was actually the PSU fan spinning up and spinning down. But, I downloaded the Precision Tool from EVGA and manually set my GPU blowers to run at 100% anyway. No whining from the blowers, just a louder airflow sound.
The rest of the system is quiet enough considering it's running at 100% all the time. Hurray for 120mm fans. If only the PSU came with a 120 instead of an 80...

I'm using the 182.06 drivers atm, I notice small stuttering when watching videos, but nothing during normal browsing/messaging.

If anyone is having severe issues, do the normal driver/firmware updates again and make sure all bios settings are good. Run msconfig and disable all non-essential crap in the Startup tab. Find a good Services guide and optimize your list of services. Disable antivirus programs. Disable Indexing and System Restore. Disable UAC.

A lean system is a happy system! ;)

Might also be from the videocard's physical build, which one can't fix without getting a new card. The load on the card is constantly going up and down with the 59XX WU's.
I have EVGA Superclocked cards that came with lifetime warranties, I assume the components used in cards with lifetime warranties are a little more resilient. Might explain why I am not getting any EUE's or noticeable performance issues.
 
Has anyone experienced EUEs with the 5900s? I've been getting some today and was wondering if there was a way to reduce the number of 5900s received? The past week I'm downloading only these WUs, and think I may need to change my client settings because other people are still receiving the core_11 WUs and I'm not. Thanks.
 
All I've been seeing is these. Almost exclusively 5902 on all 11 GPU clients. Maybe change from big to med or med to small WU? Or remove -advmethods?



 
I'm not using -advmethods, and I'm stuck with one of the worst 5902 WUs yet, I think - 1250 ppd on a 9600GSO OCed to 1770 on the shaders. Downloaded 7 hours ago, Just over 1 day left to finish, running between 12 and 18 minutes per frame (and the 4% mark taking 25 minutes) with lots of fluxuation! So far today I've seen this:

Code: 
[09:16:39] Project: 5902 (Run 14, Clone 820, Gen 0)
[09:16:39] 
[09:16:39] Assembly optimizations on if available.
[09:16:39] Entering M.D.
[09:16:45] Tpr hash work/wudata_04.tpr:  3570896404 203445983 2763229024 4273123948 4270361139
[09:16:47] Working on Protein
[09:16:49] Client config found, loading data.
[09:16:50] Starting GUI Server
[09:29:17] Completed 1%
[09:46:19] Completed 2%
[10:02:52] Completed 3%
[10:27:08] Completed 4%
[10:50:38] Completed 5%
[11:07:40] Completed 6%
[11:22:01] Completed 7%
[11:41:07] Completed 8%
[12:03:27] Completed 9%
[12:15:04] + Working...
[12:23:01] Completed 10%
[12:41:26] Completed 11%
[12:58:02] Completed 12%
[13:20:07] Completed 13%
[13:35:27] Completed 14%
[13:57:07] Completed 15%
[14:11:59] Completed 16%
[14:30:50] Completed 17%
[14:49:47] Completed 18%
[15:05:24] Completed 19%
[15:23:35] Completed 20%
[15:41:59] Completed 21%
[16:00:30] Completed 22%
[16:21:43] Completed 23%
At this rate, I'll barely hit my Million pernts... Any idea if the ATI cards with all the extra shaders are handling these any better? If so, might be time for moar "D"...
 
I don't have the -advmethods flag either, so it's either in the client setup or Stanford is sending only 5900s. Yet, some people in other threads claim that they don't see these often. :confused:
 
I just got one.

Didn't know we where talking about the 5902s in here so I started a new thread.

My GTX 260 seems to be doing ok, running about 6k+ ppd.
 
You must log in or register to reply here.
Back
Top