Nvidia GeForce Hotfix Driver 350.05

I think I owe this driver an apology.

I reported earlier that I was getting a BF4.exe app crash with it, turns out I did with 347.88 and 347.90 aswell.

Seems it may be punkbuster related as my event viewer is full of this

errrm,
Carry on.

 

It seems I have some fundamental problem with openCL part of this driver.


well the problem is only with this test alone, the rest seems correct

 

Once I've had this driver spontaniously shut down the HDMI output on my video card, as it locked up and nothing else did.

Since a reboot its been fine.

 

Woot. Glad to hear

 

Target is by Windows 10 RTM.

 

Not working in windowed for me.

 

Let´s see if I can play nfs rivals without the intermitent directx error, same engine...

 

On my system this driver and MSI AB do not play at all well together.
First time i installed it [the driver] everything seemed fine for awhile, i played MWO for a few hours then i was in chrome browsing when the screen went pink and locked up,restarted in safe mode,used DDU and installed 347.52 that got me into windows again but GPU was slow when i checked in device manager it had a code 42 thought the GPU had died,did a quick install of 347.52 shut down system completely and restarted, everything was good reapplied OC and used the system for next 3 days no worries.
Today i thought i would give 350.05 another go installed it no worries and than checked AB my OC was not applied so i clicked the profile number to apply it all the sliders went to the far right and the system locked again.
now back at 347.52. Sorry for long post

 

Naah, same directx error device hung again on nfs rivals c:

 

I find these to have better img quality too.

 

You can always check texture filtering quality for D3D10/11 with 3DCenter Filter Tester:
http://www.3dcenter.org/dateien/dateien/3DCenterFilterTesterSetup 1.1.exe
Load ground2 texture of textures subdir and do exact screenshot or video comparisons.
I'm very sure that there is no difference with HQ on Maxwell 2.0 between this and the previous driver.
However, HQ is not optimal with Kepler (and maybe even Fermi) and Q without trilinear optimization should be used, as long as there is no negative LOD bias.
However, you could check if there's still the (stupid) difference between HQ and w.o. tril. optm.

 

Did you have a point with this? You contribute 0 value.

 

Then you might contribute -1?
Stop spreading FUD and deliver facts, ffs?

 

I did actually provide a well thought of contribution, by confirming what someone else had already said. You provided some subpar way of proving you right, why don't you use it. Lazy, rude and poor at measuring is not what I call a good way of contributing.

 

This filter tester is very useful to me, will store screenshots of every driver update to see if there's any difference, thanks! :thumbup:

 

I probably already did more image quality comparisons between several drivers, several GPUs and several vendors than you ever will in your entire life.
As a result, Nvidia fixed broken HQ filtering with Maxwell GPUs (did a bugreport which showed the issue in Watch Dogs).
Me with others also started a petition to ask for texture LOD control in D3D10+, which Nvidia allowed. As a result, also negative LOD bias with D3D10+ got fixed in terms of quality.

You better think twice before you call other people lazy etc. I've compared NV's HQ filtering of this driver and GTX 980 to ALU reference implementation and couldn't find any difference (they still look absolutely compareable, as long as you don't color different pixels).
I don't have to prove what is not, you get that?

 

I had a strange issue with this driver.
When my computer would goto shut off the monitor on idle, the screen would say on and lock, you can't click on anything, CTRL+ALT+DEL opens task manager everything just starts works again (No hard lock).
Rolled back to 347.88, issue gone.

 

But did you compare it now??


I've been on nv scene for quite some time too, since 6600GT days, 8800gts, 250gts, 570gtx, now 780gtx.. And I've seen AF altering few times in the past, mostly driver forced, not ingame AF.. And I played AVP2010 last time and then switched to this one and saw, idk if 1% less DSR smoothing would make such a noticeable difference though..


If he saw it too then I didn't get delusional either, because it really looked a little crispier texture with forced AF, instead of ingame AF.


But then again why do you care so much if they're "crispier"? Its like you want to defend ****ty filtering or something :grin:

 

Some reason I'm not setting any difference in AF settings between High performance + High quality other than, well performance once I set AF to 16x. Just Cause 2 with transparency AA on there's a huge performance difference, but no quality difference that I can tell.

Any idea's?

 

JC2 has disabled TrAA in driver profile, since it has some conflicts.


Use ingame trilinear and force driver AF, there will be a difference in filtering for sure, that's why I force driver AF in AVP2010 because it has sharper filtering then ingame one.. Same thing in older TombRaider one before TR2013, or Sniper Elite Zombies (same engine as AVP2010) and few more games.