**THE NVIDIA DRIVERS THREAD**

[spankingtexan]

Feeling nervous? Probably not. They are probably find it funny people keep on trying to praise Nvidia, despite their graphic cards have been under-performing for such a long time due to these dx11 limitation and they didn't bother to address and tackle the issue until AMD brought out Mantle.

Also, people are hyping up this "theoretical" performance increase far too much. Wouldn't it be better let these "miracle drivers" making their way in to actual application first before hyping it? Let the performance do the talking, not the typical PR and marketing mambo-jumbo telling what the audience wants to hear.

Says the 290x owner

 

[drunkenmaster]

Again, the miracle performance increase that people seem to be talking about here is, 4 fps more in Thief. Also again this seems to be apparently taking it to Thief/Mantle beating performance..... by Nvidia lying about the average performance of a 290x in Thief. So where is the magical performance increase precisely?

One game has been shown to raise performance by under 10%, and it CAN'T beat AMD under DX11 by all the 290p/290x results I've seen. Mantle is raising average fps for some people, for others not, it's drastically increasing minimums and making the game much more smooth for everyone though.... is smoothness or minimums on any of these graphs?


So ultimately the only thing Nvidia is claiming that matters, is a sub 10% gain in Thief, where anyone can look in the Thief thread and see that Nvidia are flat out lying about the 290x performance. With this 10% improvement and 58fps, a 780ti will still fail to beat "real" (as in, not run by Nvidia) 290x DX results, and will almost certainly not match the 25-40% minimums increase people are seeing from Mantle and for most people a 10-20% increase in average as well.

Since when is 10-20% fps gains in newish titles via a driver update unheard of or even unexpected? Nice graphs of 400% more performance under some DX call..... but it's 4fps more in a game in which they have to lie about 290x performance to make it seem like a massive deal.

 

[Marine-RX179]

Says the 290x owner

Yea? The difference is most "AMD owners" are quite critical of the shortcomings as well and don't try to defend of justify it. Take the black screen issues, most 290/290x owners were quite vocal on the matter and give fair warnings to people that that are considering buying them, and GTX780 is may be a better alternative if they didn't want to take them chances (when GTX780 was dropped to the same price range as the 290, but it has not bounced back up the the price of £400+).

"Nvidia owners" (at the very least the few high profile ones around here) would always try to justify what Nvidia does, good or bad, right or wrong

You tried to turn this into a AMD vs Nvidia, but the issue is standalone matter of why it took Nvidia so long to act on it, and it ISN'T about AMD. I may be using an AMD card right now, but I have my fair share of Nvidia cards in the pass as well. I mean...it's Nvidia for god sake! If they have the capability optimise dx11 to improve to the extend as they claim, what the bloody hell have they been doing for the pass years!? They should be ahead of AMD taking the proactive approach, not reacting to things as they come and deal with it only when they feel the need to.

 

[spankingtexan]

Again, the miracle performance increase that people seem to be talking about here is, 4 fps more in Thief. Also again this seems to be apparently taking it to Thief/Mantle beating performance..... by Nvidia lying about the average performance of a 290x in Thief. So where is the magical performance increase precisely?

One game has been shown to raise performance by under 10%, and it CAN'T beat AMD under DX11 by all the 290p/290x results I've seen. Mantle is raising average fps for some people, for others not, it's drastically increasing minimums and making the game much more smooth for everyone though.... is smoothness or minimums on any of these graphs?


So ultimately the only thing Nvidia is claiming that matters, is a sub 10% gain in Thief, where anyone can look in the Thief thread and see that Nvidia are flat out lying about the 290x performance. With this 10% improvement and 58fps, a 780ti will still fail to beat "real" (as in, not run by Nvidia) 290x DX results, and will almost certainly not match the 25-40% minimums increase people are seeing from Mantle and for most people a 10-20% increase in average as well.

Since when is 10-20% fps gains in newish titles via a driver update unheard of or even unexpected? Nice graphs of 400% more performance under some DX call..... but it's 4fps more in a game in which they have to lie about 290x performance to make it seem like a massive deal.

Well excuse Nvidia owners for looking forward to free performance gains, I haven't a clue what these new drivers really have to offer but it's very welcome whatever the gains and will tide us over till DX12.

 

[ubersonic]

AMD crew? Oh dear. I cringe whenever i see someone use that word.

Would you prefer POSSE

Big up da red side massiv!

 

[meldarthx]

If there is more low hanging fruit for performance......sweet I welcome our green brothers into free performance......

I just don't think its going to see any close to the gains we've seen with mantle. I say though gaming is getting exciting again.....

 

[gpuerrilla]

Mantle is raising average fps for some people, for others not, it's drastically increasing minimums and making the game much more smooth for everyone though.... is smoothness or minimums on any of these graphs?

Quoted for precision.

Yea? The difference is most "AMD owners" are quite critical of the shortcomings as well and don't try to defend of justify it. Take the black screen issues, most 290/290x owners were quite vocal on the matter and give fair warnings to people that that are considering buying them

You tried to turn this into a AMD vs Nvidia, but the issue is standalone matter of why it took Nvidia so long to act on it, and it ISN'T about AMD.

Agree. The 780ti is a great card but it's still the only card to rival 290's costing lots more. Blinkers again I think.

 

[markymoo]

Had my current card for quote a few years and so many new cards. Are some of these just a rebranding code or am I falling down the pecking line.

I've got a Sapphire PC 7950 that I run at 1050/1450 and recently put my i5 2500k to 4.3ghz

Is this going to last me a while with talks of all the new cards etc. I am gaming at 1080p but see a lot are gaming at higher resolutions.

Everything looks good up to now, just wondering how far behind "great" I am and if to wait longer before upgrading this area!?

 

[shankly1985]

Can I ask why there mantle bench on 290x gets 55fps but mine on a single 290 gets 68ffps? Avg
There amd results seem very low?

 

[LtMatt]

Can I ask why there mantle bench on 290x gets 55fps but mine on a single 290 gets 68ffps? Avg
There and results seem very low?

They used 14.2 and i think 14.3 had Thief optimizations. Even my 290 on DX scores 61fps at stock.

 

[shankly1985]

They used 14.2 and i think 14.3 had Thief optimizations. Even my 290 on DX scores 61fps at stock.

So they test a beta driver an old driver may i add and claim better than mantle performance? What happens if amd improve mantle more in next driver. lol

 

[LtMatt]

So they test a beta driver an old driver may i add and claim better than mantle performance? What happens if amd improve mantle more in next driver. lol

Its corporate weasel benchmarks and this is (roughly) how it works.

No vendor shows slides of false outcomes. Those can be tested. Instead, a vendor picks truthy outcomes that support an argument and are mostly verifiable.

Examples include:

1). Choosing highest-quality texture filtering if your competition takes a higher hit than you do in that mode.

2). Activating certain functions (DoF, AO, MSAA) if they favor you, avoiding them if they don't.

3). Opting for specific game titles.

4). Testing in thermally constrained environments.

5). Preferring a "low noise" environment.

I can tell how confident AMD or NV feel based on the specificity of their tests and the amount of data they provide. If a company thinks it has a winner, they approach the situation from a "hit me with anything!" If the situation is more nuanced, they come in with qualifications. That's just a few examples.

 

[Silent_Scone]

Um...

LOL

 

[ubersonic]

Its corporate weasel benchmarks and this is (roughly) how it works.

No vendor shows slides of false outcomes. Those can be tested. Instead, a vendor picks truthy outcomes that support an argument and are mostly verifiable.

Examples include:

1). Choosing highest-quality texture filtering if your competition takes a higher hit than you do in that mode.

2). Activating certain functions (DoF, AO, MSAA) if they favor you, avoiding them if they don't.

3). Opting for specific game titles.

4). Testing in thermally constrained environments.

5). Preferring a "low noise" environment.

I can tell how confident AMD or NV feel based on the specificity of their tests and the amount of data they provide. If a company thinks it has a winner, they approach the situation from a "hit me with anything!" If the situation is more nuanced, they come in with qualifications. That's just a few examples.

Hey your forgetting Nvidia's legendary Geforce FX trick, decrementing your texture quality in the drivers (so Very High is the same as High used to be, High is what Medium used ot be, etc) so you can beat ATi on the same setting by having it mean lower quality for your customers, genius

 

[LtMatt]

Hey your forgetting Nvidia's legendary Geforce FX trick, decrementing your texture quality in the drivers (so Very High is the same as High used to be, High is what Medium used ot be, etc) so you can beat ATi on the same setting by having it mean lower quality for your customers, genius

Lol you're opening a can of worms so at this point ill bow out.

 

[ubersonic]

Can I ask why there mantle bench on 290x gets 55fps but mine on a single 290 gets 68ffps? Avg
There amd results seem very low?

Your non GPU system is faster than theirs, but I doubt it's by enough to make a 13 FPS difference so they prob used older drivers as people have said.

Lol you're opening a can of worms so at this point ill bow out.

It was actually quite good for me as I had a rubbish card so it made games run better on lowest settings lol.

On a side note if this performance boost is a repeat of that history: /shudder

 

[mmj_uk]

Hey your forgetting Nvidia's legendary Geforce FX trick, decrementing your texture quality in the drivers (so Very High is the same as High used to be, High is what Medium used ot be, etc) so you can beat ATi on the same setting by having it mean lower quality for your customers, genius

You mean that trick AMD are using TODAY with high levels of tessellation because their GPU's are only good at more basic tessellation levels.

 

[shankly1985]

Hey your forgetting Nvidia's legendary Geforce FX trick, decrementing your texture quality in the drivers (so Very High is the same as High used to be, High is what Medium used ot be, etc) so you can beat ATi on the same setting by having it mean lower quality for your customers, genius

Maybe that explains why Gregs Textures look so bad on Full Ultra compared to mine?

Spoiler

mine

Gregs

 

[~>Dg<~]

boom !

all people need to remember is its sales talk !

AMD will tell you same thing as NVIDIA. whoever you believe is who you buy



this thread is amusing. amd mantle we are great now nvidia doing same ting and its all rage up brappp !

 

[ubersonic]

You mean that trick AMD are using TODAY with high levels of tessellation because their GPU's are only good at more basic tessellation levels.

Maybe that explains why Gregs Textures look so bad on Full Ultra compared to mine?

My god what have I done