Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by stasio, Jun 18, 2012.
Great drivers everything is smooth and adaptive actually works!
It works like that, but not worth it. Rather use older drivers
No performance problems with this driver, but i had an invisible cursor in Supreme Commander 2 with them, so i had to roll back to 301.42 whql. :/
Not working with World of Wacraft, stuttering like hell.
All other games seem to work good .. strange.
Does it help if you disable Aero just for CivilizationV_DX11.exe ?
Might be also helping with the Deus EX issue
now this is something interesting, though keep in mind this will fluctuate with the P-States so if you measure it try to stay @ the same P-State for all also that you get lower throughput with PCIE 1.1 is pretty normal
when playing heroes of newerth on adaptive mode, the clocks are in 324, however they shoot up to 1000 when exiting the game. i know hon isnt the most graphics intensive game, but comon, 2d clocks when playing it? previous driver versions didnt do this
Yes, that is me. But if you read the post I made I was asking others to check if their game also crashed; I never reported it as a bona fida issue. I would not have been asking anyone else to check if I thought it was, e.g. I reported the Bully Scholarship Edition black torso/arm and F1 2010 black 'mass' glitches in the early R300 drivers and both issues have now been fixed.
Anyway, after extensive testing last night at 3.6 GHz and 3.8 GHz (using much higher voltages than should be needed) I've come to the conclusion that this game randomly crashes or rather freezes my system during the benchmarks and, yes, that includes both Test B and Test A unfortunately at 3.6 GHz and 3.8 GHz. If it isn't the driver then it must be some conflict with some other software running on my system. I've tested with my GTX 680s at stock speeds and the benchmarks still freeze my system... sometimes.
The reason I can make that assumption is because every other benchmark, application and game that I've played since I overclocked my PC to 3.6 GHz a year ago has been 100% stable and blue screen crashes and total freeze ups of this kind are extremely rare (if they've occurred then it has been traced back to a software bug and eventually fixed). Of course, I now have SLI so it may be that which has introduced the instability.
If I have time then I'll reinstall the v301.42 drivers and run both demos ten times each. That should prove once and for all whether it is the driver or game itself. I ran Test B at least seven times last night and it froze my PC fives times. I've ran Test A twice and it froze my PC once (at 3.6 GHz as well). I've ran these benchmarks several times before with earlier drivers at 3.8 GHz and they were both stable (I could have just been lucky though). Prior to this recent issue, Lost Planet 2 has always been 100% stable and has never crashed while I've been playing it.
I've just ran three loops of Metro 2033's Frontline benchmark on these settings:
Resolution: 1920 x 1200; DirectX: DirectX 11; Quality: Very High; Antialiasing: MSAA 4X; Texture filtering: AF 16X; Advanced PhysX: Enabled; Tesselation: Enabled; DOF: Enabled
It ran fine and I scored 57.33 fps (that's with DOF on!). Heaven v3.0, 3DMark11 Extreme, Batman: Arkham City DX11 and Resident 5 DX10 benchmarks all work perfectly too. I refuse to believe that this one game, Lost Planet 2 DX11, which was previously 100% stable remember, is crashing because my overclock is unstable (especially when 3.6 GHz is now unstable with these v304.48 drivers). It's not impossible, granted, but IMO it is highly unlikely. There must be some other cause.
I'll play the game tonight (v-synced) and see if it still crashes. It has never crashed before and I usually play it for an hour or two (the benchmarks randomly freeze my system after a few minutes so it should be long enough).
Still at 3.8 GHz, I enabled v-sync and dropped AA down from CSAA8xQ to MSAAx4 and ran Test B THREE times. It passed all three times without a single freeze or crash. This is the first time I've achieved that at 3.8 GHz although it only worked once out of two attempts at 3.6 GHz with CSAA8xQ using the v302.48 drivers. And Test A crashed at 3.6 GHz with CSAA32x.
I know what you're thinking, v-synced 60 fps is less CPU-intensive so that might be why it doesn't freeze. So leaving MSAA at 4x I ran the benchmark TWICE again with v-sync OFF. It ran perfectly both times giving me a Ranking of A with 84.7 fps.
It's probably not absolutely conclusive but I'm wondering if it's either the higher levels of AA or even CSAA that is causing instability in Lost Planet 2. I did Google "Lost Planet 2 DX11 freezes" last night and came across an old thread where SLI users had issues with the game freezing unless they used less than 8xMSAA. I'll see if I can dig out the link and post it here.
Do post that link if you can.
Maybe it's just a bug with the game and SLI.
Have you thought about trying it with just a single card installed?
Have you tried DX9 with 8XQ to see if that crashes too ?
If that works, i'd just chalk it up to a DX11+SLI problem. I don't know why you'd want to play in DX11, or at least I know I wouldn't. Because I get this Motion blur artifact in DX11 on the left or right depending on which direction i'm moving the camera
If i do this the game takes about 20 minutes to load a level, reverted back to last drivers and all is fine.
No it does not, i always play aero disabled..
guys what patch you are using in deus ex????
@ MrBonk - Here's one link to posts I found from typing in "lost planet 2 benchmark freezes" into Google:
http://hardforum.com/showthread.php?t=1542655 (posts #4 and #5 both had issues with AA at 8x)
If I recall correctly the standalone benchmark that was released ahead of the game had issues running under DX11 on my HD 5870 CFX until it was fixed in a later driver (annoyingly it came out after the game was released from memory but that's AMD for ya!).
I haven't tried running it on a single card but I'll certainly try that by disabling SLI as well as running DX9 SLI with 8xMSAA or higher. To be I'm getting a bit bored of running Test B though as it is dreadfully dull and has terrible art design (the boss creature looks absolutely ridiculous IMO). The game seems to run OK with 4xMSAA so I'll play the game for an hour tonight on those settings to test overall stability and I may just leave it at that.
I'll wait and see if anyone else says they have issues with this game before I report it on the NVIDIA forum as a bug. Of course, no-one may come across the issue if the use a lower 4xMSAA setting with SLI so it may well be one of those annoying issues that never gets fixed simply because no-one else has it!
I'm not suffering from a 1fps bug in Deus Ex HR, however it has crashed on me twice. May or may not be related to the driver ..
The invisible cursor also occurs for the newly released SpellForce 2 Faith in Destiny, it's only visible when in motion.
LP2 is one of my favorite games. The PC version is actually missing weather in that scene for some reason
Post again on your testing to see if anything changes
Yeah, I reported it in the Nvidia thread for these drivers, do you by chance run in the DX11 API too?
I haven't tried maximised windowed mode, but I assume it runs at 60frames because windowed mode runs perfect. I just stopped playing my DXHR playthrough for now as this driver makes other games much smoother I decided I'd just play them and wait for the next beta...
The same problem mentioned there in Aion, the cursor it's invisible and flicks when move the mouse.
Yes, I am aware of that, but my GPU runs either pcie 2.0 (win 7) or 1.1 (win8) - it has only one performance level (P0).
Lower PCIE 1.1 means ~2500 MiB/s (half of 2.0) and thats what I get with other driver versions, but not this one.
Interesting did you tried other bandwith tests same results ? 290 was Cuda 4.x riight so we have a measurable difference here in applications that where written for pre CUDA 5 ?
Though i would rather ask myself why does it fallback to PCIE 1.1 on Win 8 anyways ?
What does the WHQL show for a result on both Win 7/8 ?
PS: I wonder also if this might be caused by the massive change from 290 to 300 with the whole new manipulation layer (Adaptive-Vsync,FXAA, AO), will do some tests myself.