Linux driver 331.20 released

It seems that a lot of GPU problems revolve around specific versions of drivers. Though NVidia has their own support structure, you can often learn from information reported by others who fold.

Moderators: Site Moderators, FAHC Science Team

ChristianVirtual
Posts: 1596
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Linux driver 331.20 released

Post by ChristianVirtual »

http://www.nvidia.com/Download/driverRe ... 9372/en-us

An interesting function foul be this one:
Added support for multiple NVIDIA kernel modules. This feature allows users to assign different GPUs in the system to different NVIDIA kernel modules, potentially reducing the software overhead of coordinating access to multiple GPUs.
Will try on weekend if the PPD get hit like the last beta; hope a bug free folding run can be done with 780/660TI couple for more then 36 hours with high PPD.
ImageImage
Please contribute your logs to http://ppd.fahmm.net
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Linux driver 331.20 released

Post by bruce »

Interesting, indeed.

Will this require a change to any of the FahCores?

I wonder what they mean by overhead. Will this increase or decrease the amount of CPU time needed to run FAH on multiple GPUs? (or leave it unchanged?)
ChristianVirtual
Posts: 1596
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Linux driver 331.20 released

Post by ChristianVirtual »

Still not looking good for the "old" 660TI,
A frame on 319.x driver which took around 2:38 now takes 4:01; not as bad as with the beta driver but still bad

As for 780 and the first frames it looks good.

So the question: what should I do with my 660TI ?

I hope that the new function as mentioned above might allow us to use 331.20 on 780 card while 319.x on 660
Last edited by ChristianVirtual on Fri Nov 08, 2013 1:59 am, edited 1 time in total.
ImageImage
Please contribute your logs to http://ppd.fahmm.net
jimerickson
Posts: 533
Joined: Tue May 27, 2008 11:56 pm
Hardware configuration: Parts:
Asus H370 Mining Master motherboard (X2)
Patriot Viper DDR4 memory 16gb stick (X4)
Nvidia GeForce GTX 1080 gpu (X16)
Intel Core i7 8700 cpu (X2)
Silverstone 1000 watt psu (X4)
Veddha 8 gpu miner case (X2)
Thermaltake hsf (X2)
Ubit riser card (X16)
Location: ames, iowa

Re: Linux driver 331.20 released

Post by jimerickson »

@ChristianVirtual: running gtx780's here. i am on 331.17, waitingfor xorg-edgers ppa to update. what project is your gtx780 on and what do ppd look like?
ChristianVirtual
Posts: 1596
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Linux driver 331.20 released

Post by ChristianVirtual »

Here the graph from all three cards

Image
ImageImage
Please contribute your logs to http://ppd.fahmm.net
jimerickson
Posts: 533
Joined: Tue May 27, 2008 11:56 pm
Hardware configuration: Parts:
Asus H370 Mining Master motherboard (X2)
Patriot Viper DDR4 memory 16gb stick (X4)
Nvidia GeForce GTX 1080 gpu (X16)
Intel Core i7 8700 cpu (X2)
Silverstone 1000 watt psu (X4)
Veddha 8 gpu miner case (X2)
Thermaltake hsf (X2)
Ubit riser card (X16)
Location: ames, iowa

Re: Linux driver 331.20 released

Post by jimerickson »

FAHMM, got to love it! use it all the time myself. thanks for the info, looks good.
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Linux driver 331.20 released

Post by bruce »

Using all the time itself is good. Potential disadvantages: Heat is worse. Screen lag may or may not be worse.

Anything to report?
jimerickson
Posts: 533
Joined: Tue May 27, 2008 11:56 pm
Hardware configuration: Parts:
Asus H370 Mining Master motherboard (X2)
Patriot Viper DDR4 memory 16gb stick (X4)
Nvidia GeForce GTX 1080 gpu (X16)
Intel Core i7 8700 cpu (X2)
Silverstone 1000 watt psu (X4)
Veddha 8 gpu miner case (X2)
Thermaltake hsf (X2)
Ubit riser card (X16)
Location: ames, iowa

Re: Linux driver 331.20 released

Post by jimerickson »

not seeing any screen lag at all. i really like the graphing feature. helps me understand whats going on with my gpus. heat maybe a little more than usual but nothing to be overly concerned about. thanks for all your work!
ChristianVirtual
Posts: 1596
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Linux driver 331.20 released

Post by ChristianVirtual »

660 TI finished ok, but took longer with less points (first line 331.20; second line the 319.x)

Code: Select all

Project	Run	Clone	Gen	TPF	Credit
7810	0	8	408	4:28	9856
7810	0	153	385	2:38	12837
Now I stop the 660 TI and let the system run to see if at least the 36 hours are ok ...
ImageImage
Please contribute your logs to http://ppd.fahmm.net
powerarmour
Posts: 34
Joined: Wed Oct 29, 2008 1:00 am
Location: UK

Re: Linux driver 331.20 released

Post by powerarmour »

Yep still on the 319's myself, I was hoping Nvidia would have resolved this by now... :(

I've left my Fermi GTX 560 Ti running the 331.20's as the hit isn't too bad/almost negligible, but it still seems like half speed on the Kepler GTX 760.
Image
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Linux driver 331.20 released

Post by bruce »

At one time, installing Silverlight changed FAH's productivity. I've heard no recent reports either way, but it might be worth a try.

I'm trying out Win8.1 this week (ugh) and notice that it's an optional install.
art_l_j_PlanetAMD64
Posts: 472
Joined: Sun May 30, 2010 2:28 pm

Re: Linux driver 331.20 released

Post by art_l_j_PlanetAMD64 »

Now, this is for a Windows 7 system, not for a Linux system, but the newest driver that I tried out, v331.65, showed a 50%-60% drop in PPD values, when compared to the "best" older driver that I have used (v327.23 WHQL).

Please see a more detailed description about this, at this link.

Cheers,
Art
art_l_j_PlanetAMD64
Over 1.04 Billion Total Points
Over 185,000 Work Units
Over 3,800,000 PPD
Overall rank (if points are combined) 20 of 1721690
In memory of my Mother May 12th 1923 - February 10th 2012
powerarmour
Posts: 34
Joined: Wed Oct 29, 2008 1:00 am
Location: UK

Re: Linux driver 331.20 released

Post by powerarmour »

Yeah it's definitely the same issue though on the 331's, and although the Linux driver isn't quite in sync numerically, I'd have to assume all 331's will have this bug until resolved.
Image
art_l_j_PlanetAMD64
Posts: 472
Joined: Sun May 30, 2010 2:28 pm

Re: Linux driver 331.20 released

Post by art_l_j_PlanetAMD64 »

bruce wrote:At one time, installing Silverlight changed FAH's productivity. I've heard no recent reports either way, but it might be worth a try.

I'm trying out Win8.1 this week (ugh) and notice that it's an optional install.
That's interesting, I hadn't heard about that before (that installing Silverlight could change FAH's productivity).

I do know that I stopped installing Silverlight on my (Windows-only) machines, and removed it from all (Windows) systems where it had been installed, for some definite reason. My poor 60-year-old brain cannot remember the reason why I did this, but I do remember that there was a definite reason, why I did this (eliminated Silverlight from all 24 of my Windows folding systems).

Cheers,
Art
art_l_j_PlanetAMD64
Over 1.04 Billion Total Points
Over 185,000 Work Units
Over 3,800,000 PPD
Overall rank (if points are combined) 20 of 1721690
In memory of my Mother May 12th 1923 - February 10th 2012
ChristianVirtual
Posts: 1596
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Linux driver 331.20 released

Post by ChristianVirtual »

The good news: system still working after 36 hours; I'm now in 39th hour after last reboot. That's good news for two GTX 780.

The bit strange thing is that after 36hours the already bad performing GTX 660TI gets even a tick worst and hit by another one or two second per frame. Not really important but indicating there is still something going on with the 36hours ...
Image
ImageImage
Please contribute your logs to http://ppd.fahmm.net
Post Reply