• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Manual Download v1.19

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
P

Psykoikonov

Member
​
Joined
Dec 16, 2003
Location
ON, CA
Last edited:
I guess if you eliminate some optimizations it's going to be slower, but I went ahead and installed it on this machine. Going from 1.15 to 1.18 made a big difference, but I'm still seeing an EUE here and there. I'll give up 5% or even 10% of my production if it means I won't see any more units end early. I'll wait to see if it does end the problem before I install it on my other machine.
 
Tried it and lost about 8% PPD. Since I'm getting only the occasional EUE, I'm staying with 1.18 for the moment.
 
Installing on all my clients... heck, maybe this one will be v1.09 reborn... the version number sure is appropriate. ;)
 
Bump...

This appears to be a winner folks... I'd recommend installing it right now!!!

Since installing on my 6 clients I've been EUE/UM free... only ~12 hours worth of Folding, but I think that's enough to say that this core is waaaayyyyy more stable than any v1.13 -> v1.18 core.

It's a touch faster than v1.09... so they did improve a bit even when removing a lot of the optimizations present in v1.13 -> v1.18. I think that's key for Stanford/nVidia here... don't go for the entire motherload of optimizations all at once, makes it too hard to debug when something goes wrong. I would think stability would be more important to them than slightly faster production... especially since the instability is slowing the project by a much greater margin than the small speed up gave them in the first place.
 
Okay, what does NANS detected on GPU mean? I had one card put out three or four units after switching to 1.19 and then EUE out of every single work unit after that unit it automaticallay paused folding on that GPU for 24 hours. Is that card fried now?
 
What series card? I think 8600 series are still having problems...

However, I had that same experience with v1.13 - v1.18. Once the card went into the death spiral, as you're seeing, I had to reboot the machine to get it back Folding. Nothing else seemed to work... lowered clocks, deleted the current WU, nothing.

The thread link'd to at the top of this thread also has folks with mixed results... some are EUE/UM free, others still having issues. So far :crosses fingers: I'm EUE free. :)

I don't think your card is dead... I just think the machine needs to clear its head. Give it a reboot and see if that does the trick.
 
I've theorized that the constant UM loop has got something to do with the nVidia drivers and/or the CUDA layer getting hosed over time by the FAH core, taking a restart to get things cleared out of the pipeline... but that's just my guess. If you're now running EUE/UM free... I doubt you'll experience that situation again.

I've also seen recommendations to run the latest 178.24 drivers if you're not already.

24 hours strong here... no EUE/UM on v1.19. :beer:
 
harlam357 said:
24 hours strong here... no EUE/UM on v1.19. :beer:
Click to expand...

Well, after thinking about it, I went back to the 1.19s, and I'm over 24 hours with no EUE/UM either. And my PPD is up, so I think that any PPD 'loss' is being more than made up by not having EUEs/UMs.

(Of course I've also swapped out a 8800 GTS 640 for a 9800 GX2, but I think the difference is more than that.)

I still have 1 machine to switch over, and after that I'll see what my PPD really is.
 
I don't know why, but I have two 8800GTs in this machine and after the switch one is stable, not a single EUE, the other is having more problems than it did under either 1.15 or 1.18. I suppose the card could have just went bad at about the time I switched, but that would be quite a coincidence.

I went ahead and underclocked both cards and I'll see how it goes over the next day. Both are factory overclocked cards, EVGA 8800GT SC.

Oh, and editing to ask a quick question, how do I know which card is which if I decide I need to RMA the one that is having problems?
 
Before you do that... did you install the display drivers with both cards already installed?

Or... did you install the second card after drivers were already installed?

The latter could possibly cause problems... just something to try before going with an RMA.
 
I've had at least three instances on my daily office machine where one or the other cards and sometimes both went into a UM loop. Un and reinstalling the drivers fixed it every time.
 
A result that lends itself to my theory that the FAH core is indeed hosing something in the drivers....

A simple reboot didn't fix it for you ChasR?
 
I had one card installed and folding, and installed the second one a month or two later. The thing is that I haven't changed anything in my setup since I started GPU folding. I looked a few days ago and could see at least a two week period with over 400 work units completed without a single EUE.

I've been planning to move these two cards to a dedicated folding system, so I guess I'll have to make the time to do that as soon as I can.
 
You must log in or register to reply here.

Similar threads

c627627
How to install Windows XP on a 2016 Skylake Asus Z170-A motherboard
2 3 4 5
Replies
99
Views
61K
Zeldaboy14
Z
Johan45
Maximus Formula to Rampage conversion. Plus Vdrop mod.
Replies
13
Views
4K
Aleslammer
Aleslammer
c627627
How I installed CyanogenMod 13 on Galaxy S5 to get Android 6.0.1 Marshmallow
Replies
6
Views
8K
c627627
c627627
'Cuda340
MSI Afterburner 4.1.1 Final Download
Replies
2
Views
5K
Johan45
Johan45
WhitehawkEQ
End of Beta cores?
2
Replies
21
Views
2K
WhitehawkEQ
WhitehawkEQ
Back