Nvidia still trying to fix their ****. 331.93

Update: Notebook drivers

NVIDIA DRIVERS GeForce 331.93 Driver (BETA) BETA32 bit


NVIDIA DRIVERS GeForce 331.93 Driver (BETA) BETA64 bit

I still think most of the fermi users with problems have trashed cards, NV even asked for people to send them their cards. Why? So they could examine burned ones and figure out where they messed up! - not to get that model for sure! People leaving NV in droves right now since NV refuse to even comment.

We shall see.....

Edit: https://forums.geforce.com/default/...ay-driver-feedback-thread-released-11-27-13-/

 

Updated: links now working...

 

Thanks for sharing! +1

 

I can confirm that 331,65 was burning fermi's (like mine).
Its stupid of them (and alienware who broke more than they fix)..

Did someone already confirm exactly why fermi's break down on 314,22++ drivers?

 

Nah, probably never going to know - unless NV come up with some c0ck and bull story that fits the circumstances.

All I can say is:

At 320.18 a whole bunch of cards failed.
All 3xx drivers were overvolting the cards I saw (580/675).
This overvolt was fixed with the beta prior to the last WHQL.
Most common problems are black screen lockup, and TDR related BSOD's

Something after 314.22 put the cards through too much stress and along with the extra voltage they fried (my theory). I have used both the last drivers on a replacement card without issue.

This latest one will have to wait for a couple of days for me. I want to download the latest patch to BMAO so |I can benchmark it on the WHQL for comparative data

 

This driver is working well for me, no issues so far. Identical performance in game benchmarks & 3DMark11 in comparison to previous whql (331.82), even though this driver contains a newer version of PhysX. Tested game benchmarks: Metro Last Light / Batman Arkham Origins / F1 2012.

 

Just gave it a quick run through, had something strange happen though.

BMAO +3fps so basically the same. F12012 +20fps! Big jump (around 45%, something not right there?). Dirt Showdown +6 (around 15%).

When I was base marking BMAO the v-sync cut in after my first test. Couldn't get rid of it, even disabled in the driver. Even more annoying was it capped at 120hz in the menus but caped at 60hz in the game. Not a problem in gaming really but messes up benchmarking. I'm wondering if the problem was across the board and what gave me messed up F1 2012 numbers

A full restart did not help but as soon as I updated to this driver v sync is off. I have read some people having issues with v-sync so probably fixed in this version.

 

Just a quick update on this version

Working great and it is fixing a lot of problems for older card users. Around 50% are still not able to move up from 314.22 though

The DPC latency is the lowest I have ever seen - this should hike up benchmarks I would have thought?



Doesn't even show the NV driver as the top hitter!

 

@Micky Hard pagefault count doesn't look good , check under Stats and Processes (besides latmon.exe) it's not Nvidia related.
A11 BIOS .NET 4.5 installed, try updating IRST version 12.8.0.1016 Driver Details | Dell US

 

Thanks man, I didn't turn off the windows essentials that crabs that number

 

.... there's no need to disable defender, open w. defender -> Settings -> Excluded files and locations - select C:\Program Files\LatencyMon -> Excluded processes - both C:\Program Files\LatencyMon\LatMon.exe / IDLT.exe -> Done

 

Thanks, where is IDLT.exe? When I turn off realtime I get zero hardpage faults. I excluded latmon.exe but it made no difference

Edit: I added latencymon to the excluded processes as well as 'files and locations'. All good now, Thanks

 

C:\Program Files\LatencyMon\IDLT.exe is probably part of v6, exclude AW OSD too :thumbsup:

 

One just for the laptops!

Quote:

"For those requesting a notebook driver with desktop stability fix to test if you have the same issue fixed with this driver, I have uploaded drivers that applies this fix for all notebook GPUs. Driver is based on 331.58. The links are in the other desktop stability thread on page 79:

Our software team has created a notebook package version that applies the fix for all notebook GPUs for those who wish to see if their notebook stability issue is related to this problem. If your issues continue, you are likely to have a different issue. Like the previous test version, this is also based on driver 331.58 driver:

Notebook
Windows 7/Windows 8 32-bit:
http://geforceforums.s3.amazonaws.com/NB-32-bit.zip

Windows 7/Windows 8 64-bit:
http://geforceforums.s3.amazonaws.com/NB-64-bit.zip


https://forums.geforce.com/default/...read-released-11-27-13-/post/4045201/#4045201"

 

What is it with nvidia burning fermi chips

 

IMO it was mulitple rushed fixes and then backpeddling on the real issue! At the 3.xx release there was a major change and the GPU boost feature was triggering a brief voltage boost at card initialisation on cards that do not support 'boost'. Didn't seem to cause a problem and we know that up to 314.22 the fermi cards were OK. Then NV had to rush out a driver for their latest card that was just about to be released. So, even though there was two beta's released before 320 WHQL that had the lockups and BSOD's they went ahead and WHQL'd it.

Then it really hit the fan. So many cards that either could not run the later drivers or the card just gave up altogether (like mine did). I wondered why NV had asked users to send them their cards saying that they would like to test then. Worst affected was factory OC'd cards, I assume they already had a voltage hike applied and adding more killed them. I think they simply wanted to find out exactly what on the card was getting damaged.

The separate driver line for the fermi cards was introduced attempting to lower the stress on the cards to the level of 314.22 but still give some performance improvements. Partly as P R as well, they are not going to admit that the damage to cards was their fault. Users expect a new driver for new games when in reality just a game profile for the existing driver would suffice. Over time most people suffering the problems have either gone over to the other team or are just waiting for AMD's newest flagship.

Most of the fixes posted (not drivers) are simply a workaround trying to lower the latency by messing with windows. NV seem to have fixed the latency issue with the last two drivers along with the voltage boost no longer happening. For some people this was all that was needed to get them back working, for others the damage is too much and they will simply never work correctly.

My take on it anyway

 

Doing it on purpose so people have to upgrade to Kepler for an extortionate amount of money? lol

 

Never ascribe to malice what can be adequately explained by incompetence is a favourite quote of mine.

 

That's a good one... have to remember that. We can apply it to Haswell as well (pun intended) and Windows 8 Modern UI, LOL.

 

I upgraded to the new GPU drivers,

my 3d Mark scores are

ice storm 86002
cloud gate 16179
fire strike 4003

I am running stock clocks on my 680M. This is too low right ??

 

Perfect!

Its only too slow if you're laptop lags so much it becomes unusable.

Benchmarks dont mean in the real world.

 

4000 is about normal for a stock 680M

Benchmarks are good for comparing the same card model to ensure it is performing normally.