The Bottom Line - Nvidia Drivers Vs AMD Drivers Over Time

[maonayze]

It looks like this one can finally be put to rest.

So who's drivers currently are better and do they gain any extra performance over time.

Hard[OCP] (Well renowned by many for Nvidia Bias) tell us their findings...this is The Bottom Line.....

It seems compared to AMD, that NVIDIA driver improvements are more about bug fixes, and "Game Ready" day one game support, rather than performance improvements over time. On the other hand AMD focuses on performance improvements alongside bug fixes and new game support over time. This may very well be a question of resources for each company and access to game developers.

In this review, and the previous one, we have seen a clear pattern of AMD having a more consistent progression of performance on its GPUs from driver to driver over the course of the video card's lifetime compared to NVIDIA. The 6 month old AMD Radeon RX 480 has achieved the same performance advantages as the AMD Radeon R9 Fury X has in 18 months is quite telling. On the other hand, the 8 month old GeForce GTX 1080 has been flat in terms of performance improvements. The performance you got out of it at launch is the same performance you get out of it today.

So that begs the questions... Does AMD launch video cards with performance left on the table in terms of drivers? Does NVIDIA launch video cards that are optimized to the utmost out of the gate?

Or?

Does AMD keep its driver engineers' noses to the grindstone eking out every bit of performance that it can find as time passes? Does NVIDIA let performance optimizations go undiscovered over time?

Full article here

http://www.hardocp.com/article/2017/02/08/nvidia_video_card_driver_performance_review/1

 

[Panos]

Nothing new. I can see it clearly with my Nano for more than a year I have it, the perf goes up clock for clock. Even that the last drivers allow me to overclock it to 1150!!!!!!!
And can be seen with the RX480 pitted against the GTX1060 6GB from July 16 to January 17.

Or more prominent if you compare the GTX780Ti & Titan Black with the 290X!!!!!!!

And yes Fallout 4 perf jump by more than 20% on the FuryX/Nano between 15.7 and 17.1 is trully impressive!
If you consider is an Nvidia gameworks title!!!!

 

[Aretak]

It doesn't really answer anything of course. The Nvidia supporters will say that it's because AMD's drivers are garbage at launch, whilst Nvidia's are perfect day one and can't be improved upon. The AMD supporters will say that Nvidia are happy to leave performance on the table for existing cards because they only care about selling you new ones, whilst AMD work hard to bring extra value to existing owners.

I know which version I believe of course.

AMD have also had the benefit of working with a largely set architecture for five years now, with incremental iterations of GCN resulting in a multi-series product stack that can all be improved at the same time for little extra effort or cost. That obviously isn't possible for Nvidia, who've used three different architectures in the same timeframe. Still, that doesn't do anything to explain the improvement versus lack of improvement when comparing the 480 and 1080 over a similar timeframe... meaning you're back to the initial argument anyway.

 

[Blackjack Davy]

I can honestly say I had less issues with AMD's drivers than with Nvidia's lately, not that I expect anyone to believe a word of it.

 

[Lokken86]

The crimson drivers themselves gave quite a good little boost when they came out and they still managed to get more since then.

 

[Kaapstad]

Too much bloatware with both vendors drivers.

 

[Rroff]

Still, that doesn't do anything to explain the improvement versus lack of improvement when comparing the 480 and 1080 over a similar timeframe... meaning you're back to the initial argument anyway.

One thing they seem to have not touched on in testing - there was a bugfix soon after launch with the 480 that boosted performance 4-5% - it launched down on its intended performance.

AMD also only recently introduced DX11 optimisations that nVidia did at the end of 2013/start of 2014 which skews the picture a bit.

 

[Nasher]

Nothing new. I can see it clearly with my Nano for more than a year I have it, the perf goes up clock for clock. Even that the last drivers allow me to overclock it to 1150!!!!!!!
And can be seen with the RX480 pitted against the GTX1060 6GB from July 16 to January 17.

Or more prominent if you compare the GTX780Ti & Titan Black with the 290X!!!!!!!

And yes Fallout 4 perf jump by more than 20% on the FuryX/Nano between 15.7 and 17.1 is trully impressive!
If you consider is an Nvidia gameworks title!!!!

Witcher 3 also performs noticeably better on my rx480 than it did on my 970. Despite being a "gameworks" game. Even with all the options enabled like hairworks.

 

[Gregster]

Not sure how you can say that "This can finally be put to rest" when nothing is put to rest in the article.

So that begs the questions... Does AMD launch video cards with performance left on the table in terms of drivers? Does NVIDIA launch video cards that are optimized to the utmost out of the gate?

So if it was put to rest, that would be answered. Is AMD slow to get the best performance? Does NVidia release day one drivers and don't bother getting anymore performance?

 

[tommybhoy]

You just know AMD's going to get faster over time, unless it's faster in the first place-their 10% off the competitors pace is going to be at least 10% if not more going forwards.

Performance from tech savvy folks is abundantly clear@launch on both of them, unless you are a complete numpty and can't comprehend performance reviews(that an individual puts trust in) and not what's enforced on you by others with cherry picking results in the likes of OcUK forum.



But that could all be about to change with Vega's new arch.

Is AMD slow to get the best performance?

Imo-yes.

Does NVidia release day one drivers and don't bother getting anymore performance?

Imo-yes.

It was put to bed from my pov years ago.

 

[hardkore]

I buy the fastest card out at the time for what i can afford so i have ZERO brand loyalty.

For me, regular driver updates are about on par with each other.

However, When AMD introduced crimson drivers it kept under volting my vapor x 290 and even at stock frequencies it would crash on boot nearly rendering it useless.

I had to beg Sapphire in the end for a revised bios which increased the Vcore so it would remain stable.

**** poor and the feedback from AMD was less than helpful. I will think long and hard before returning to the red team after that incident. The 290 was a cracking card.

 

[TNA]

You just know AMD's going to get faster over time, unless it's faster in the first place-their 10% off the competitors pace is going to be at least 10% if not more going forwards.

Performance from tech savvy folks is abundantly clear@launch on both of them, unless you are a complete numpty and can't comprehend performance reviews(that an individual puts trust in) and not what's enforced on you by others with cherry picking results in the likes of OcUK forum.



But that could all be about to change with Vega's new arch.



Imo-yes.



Imo-yes.

It was put to bed from my pov years ago.

I am struggling to understand this point of view. Maybe you can clarify.

When you buy the card you buy it for what it is at that moment. Driver improvement is simply a bonus is it not? It may or may not come.

How can that be flipped and seen as "Nvidia offers the cards full performance day one"? Logic, does not compute

 

[Nasher]

Nvidia performance sometimes gets WORSE as time goes on :/

 

[tommybhoy]

When you buy the card you buy it for what it is at that moment.

Yes because hopefully you looked at performance figures at launch and you didn't dive in blindly.

Driver improvement is simply a bonus is it not?

Yes and one vendor is far more likely than giving it out than the other, part reason why one vendor is minted and released their performance last year.

It may or may not come.

The older I get, the more I tell her that.

 

[TNA]

Yes because hopefully you looked at performance figures at launch and you didn't dive in blindly.



Yes and one vendor is far more likely than giving it out than the other, part reason why one vendor is minted and released their performance last year.

Nice straw man there

The older I get, the more I tell her that.

Haha

 

[arc@css]

From those results, I can see more performance gains over time from AMD, but it's not as much as I was expecting. Overall, there didn't seem to be that much in it, really.

 

[tommybhoy]

Nice straw man there

From those results, I can see more performance gains over time from AMD, but it's not as much as I was expecting. Overall, there didn't seem to be that much in it, really.

I imagine if older gen was included like 7970 v 680, or 290 v Titan for comparison...

 

[Rroff]

Would be interesting to see some results including release era drivers and more recent from someone who benchmarked Kepler properly.

(That isn't to say I expect Kepler to do particularly well against the Hawaii cards with more recent drivers and some more recent games - just so many sites seem to have Kepler results way down on what actual users will see).

 

[arc@css]

I imagine if older gen was included like 7970 v 680, or 290 v Titan for comparison...

Hmmm.....maybe somewhere in the middle.

 

[Gregster]

Nvidia performance sometimes gets WORSE as time goes on :/

Interesting. What makes you say this?