**THE NVIDIA DRIVERS THREAD**

[Kaapstad]

Will the 384 bit bus be able to take advantage of 6GB in 2 card set ups?

Yes

That is Gregsters current setup.

 

[whyscotty]

The R9-295x2 is faster than gtx 780ti sli in almost everything in this review
http://www.pcper.com/reviews/Graphi...-295X2-8GB-Graphics-Card-Review/Battlefield-4

Shouldn't believe everything you read

 

[setter]

Shouldn't believe everything you read

Very true, like a recent bench from a german site that had a 780ti averaging 55fps at 1080 on max settings on bf4's shanghai, (91 for an r9 290x under mantle). I get way better averages on a bog standard rev 1 wf 780. Problem with paid benches imho, is who's doing the paying. Hence id rather hear results from real users on here etc.

 

[Rroff]

My 780M (a fully fledged Kepler GK104) is showing improvements so a 6xx series card should as well.

Seeing some improvements on my laptop GTX675m - actually fermi based - setup my nephew on it to play test drive unlimited 2 which is fairly old now and it used to be around 40-60fps with max settings but was pretty much 55-60+fps with very few drops with these drivers.

 

[humbug]

Very true, like a recent bench from a german site that had a 780ti averaging 55fps at 1080 on max settings on bf4's shanghai, (91 for an r9 290x under mantle). I get way better averages on a bog standard rev 1 wf 780. Problem with paid benches imho, is who's doing the paying. Hence id rather hear results from real users on here etc.

Just like this one. Its because they were using Mantle, this BTW is on Nvidia's new Drivers.

Are [H] also lying?

 

[Rroff]

Just like this one. Its because they were using Mantle, this BTW is on Nvidia's new Drivers.

Are [H] also lying?

Regardless of whether thats accurate or not and quite possibly is - H have a habit of cherry picking results to use the ones that show the story they want to tell - maybe not lying but it wouldn't be that far off the mark to say they can be selective with the truth.

 

[oweneades]

Seeing some improvements on my laptop GTX675m - actually fermi based - setup my nephew on it to play test drive unlimited 2 which is fairly old now and it used to be around 40-60fps with max settings but was pretty much 55-60+fps with very few drops with these drivers.

Nice! That is a decent gain.

 

[setter]

Just like this one. Its because they were using Mantle, this BTW is on Nvidia's new Drivers.

Are [H] also lying?

The one i was referring to was at 1920x1080. Im on 1920x1200 myself.

How many time you going to spam that chart, wore out the BF4 sweetspot one?

Regarding that spot on shanghai, ive seen an increase of 20fps ono on that bit using the new drivers. Tested at both lowest possible and max possible graphics settings. Res scale at 100%. Stock gtx 780.

 

[Rroff]

Regarding that spot on shanghai, ive seen an increase of 20fps ono on that bit using the new drivers. Tested at both lowest possible and max possible graphics settings. Res scale at 100%. Stock gtx 780.

Had a quick play with some older drivers including .86s and for some reason in that spot I can't get over 108fps tops which is even worse than what you get with the 86s but putting the 337s on jumps to 170fps which is about right compared to your results given my extra clock speed.

 

[setter]

Had a quick play with some older drivers including .86s and for some reason in that spot I can't get over 108fps tops which is even worse than what you get with the 86s but putting the 337s on jumps to 170fps which is about right compared to your results given my extra clock speed.

334.89 were the drivers i had been running for a good while now. Swapped over to the new betas yesterday. Could probably eek a few more fps out by clocking this card, not that it's much of a clocker though from what little i have tried.

 

[Gripen90]

Just installed the Forceware 337.50 driver - I haven't really noticed any differences over the 334.89 as to smoothness etc, but in the games that should get a boost did get it.
Benchmark results as for Heaven, Valley, 3Dmark11/FireStrike haven't changed as such - I still get better bench results with forceware 320.49.

Well the are 337.50 are keepers but I'm hoping for more performance to 3-way SLi systems

Shouldn't believe everything you read

I don't I just wondered because I see review going either way ! Only one I trust for real are user reviews and user comparisons

 

[Ian Evey]

The good is in BF4 i have 20 fps average higher than with the old drivers on any 64 player map, which is very good.

The Meh is i did sniper elite , sleeping dogs, and heaven have seen 1 to 5 fps so meh. on a 780 and 3570k.

 

[James J]

The good is in BF4 i have 20 fps average higher than with the old drivers on any 64 player map, which is very good.

The Meh is i did sniper elite , sleeping dogs, and heaven have seen 1 to 5 fps so meh. on a 780 and 3570k.

My dads geforce experience has defaulted to the ultra preset now in BF4 on his 760 GTX card (was a mixture of high and ultra before) and runs silky smooth. Hes also gained some FPS in titanfall on very high settings with 4x AA, yet I haven't measured it yet! Definitely smoother though.

I've personally not noticed any load time delays with the shader cache setting enabled, but its business as usual as far as my 2x evga superclock 780 ti's are concerned. No stutter, no lag, no nonsense.

 

[PhysicsMan94]

I saw a post from Gregster that said he speculated if MS and Nvidia worked together on these performance optimizations drivers and it got me thinking. Its a possibility i think, Mantle annoyed microsoft. It wouldn't surprise me if MS helped them do this to help negate some of the performance benefits of Mantle.

Makes sense, it's what I would do in Microsoft's position. MSFT has every reason to be peed off with AMD in my opinion, given that they went from helping with many generations of DirectX development (along with Nvidia) to suddenly releasing a competing product.

 

[Silent_Scone]

Yes Matthew, it's called competition. Congratulations you've unearthed a previously unknown business incentive.

Back OT

I mentioned in the improvement thread. AC4 Tri SLi performance is much improved for me. Far smoother as I think there was definitely a CPU bottleneck there. Doesn't really translate in fps as rhe engine is locked, though.

 

[Rroff]

334.89 were the drivers i had been running for a good while now. Swapped over to the new betas yesterday. Could probably eek a few more fps out by clocking this card, not that it's much of a clocker though from what little i have tried.

Meant 89s lol bit tired from work and not too hot on specifics atm :S

 

[James J]

Might have a bug with these beta drivers btw, which I need other people to test for me. Recorded some shadowplay footage last night and the in game audio only works on around half of the clips. I've not changed any settings in between these games as it was a 5-6 game (basically about an hour straight of titanfall) worth of gameplay clips. Can someone who also uses shadowplay test this out?

 

[Rroff]

Still really no one really posting anything interesting about how the driver is in their system. Performance increases, bugs etc. Single, SLi, 3way SLi results/improvements if any... the AMD vs nVidia ranting is really boring to read.

They seem to be great for me on both systems I've tried them on so far - to be fair my laptop has languished on fairly old drivers for awhile so the gains there might just be the normal incrimental updates just experienced in one go but seeing ~10% average gains in most of the stuff its used for gaming wise not just DX11 stuff and decent stabalisation of minimum fps. At the very least so far seem like a nice stable driver release with some good gains in general over older drivers.

The difference on my gaming PC with BF4 is nice too - ultra settings only seeing about 20% gains overall but with a mix of low to ultra settings that I use to try and hold 125fps while still looking reasonably good sees a huge increase especially in minimum fps.

 

[JGStonedRaider]

Now then don't kill me or Matt because I just wanna ask him a question that relates to the new Nvidia driver..IMO It's a similar perf boost to the driver AMD released during BF3 that made the 7970 and other cards much faster. Why you may ask, well I feel the current Nvidia driver is similar in the way it has improved BF4 performance and would use the two drivers as a comparison rather than the whole Mantle / 337 argument.

So Matt (or others) what was that AMD driver?

Also, the new driver / shadowplay is very nice. FPS hasn't dropped under 80FPS for me since installing.

 

[Rroff]

Might have a bug with these beta drivers btw, which I need other people to test for me. Recorded some shadowplay footage last night and the in game audio only works on around half of the clips. I've not changed any settings in between these games as it was a 5-6 game (basically about an hour straight of titanfall) worth of gameplay clips. Can someone who also uses shadowplay test this out?

I recorded a whole load of BF4 with shadowplay last night with them and didn't notice any audio issues in the 2-3 clips I played back, not tried any other games though.