Nvidia Pascal Reviews [1080XP, 1080ti, 1080, 1070ti, 1070, 1060, 1050, and 1030]

[Orion]

Not really strange they are just waiting to see what AMD has to offer then they will set the clocks.

I wonder if the 980 ti rumored perf of polaris is in general or just dx12. If it is in general that's around fury x, which is within 10% 1080 in some benchmarks... so if amd can boost clocks a bit, it could hypothetically match 1080 in some benchmarks going by rumors.

 

[pjbliverpool]

Seems to perform in line with what was expected, which is to say, very well. I'm more looking forward to the custom AIB variants though, especially the water cooled ones which are rumoured to clock as high at 2.5Ghz. I can't wait to see those benchmarks!

Also, a bit disappointed to have not seen any Quantum Break benchmarks yet.

 

[Voxilla]

Not really strange they are just waiting to see what AMD has to offer then they will set the clocks.

On the other hand Nvidia already stated the 1070 does 6.5TFLOP/s, so it's just a matter of disabling a certain amount of SMs and adjusting the clock.
Maybe they don't yet have a big enough batch to determine the optimal configuration.

 

[3dilettante]

I forget, weren't there some salvage SKUs in the past that had different clock/SM combinations yet kept the same product number?

 

[Kaotik]

Did a comparison of handful of sites, ~29% faster than 980 Ti, 32% faster than Fury X (1440p + 4K)

 

[gamervivek]

980Ti's great overclockability is a problem with 1080's average overclocks so far.

 

[iroboto]

What is the difference between tier1 and tier 2 CR?

From what I understand the higher the tier level the more additional cases are handled by the hardware.

 

[Orion]

Did a comparison of handful of sites, ~29% faster than 980 Ti, 32% faster than Fury X (1440p + 4K)

On average, but on a few dx12 benchmarks the gap can get quite smaller with fury x.

question is, is this mere amd tuning or is this indicative of future dx12 titles with async compute performance?

 

[Razor1]

Seems to perform in line with what was expected, which is to say, very well. I'm more looking forward to the custom AIB variants though, especially the water cooled ones which are rumoured to clock as high at 2.5Ghz. I can't wait to see those benchmarks!

Also, a bit disappointed to have not seen any Quantum Break benchmarks yet.

http://www.hardwarecanucks.com/foru.../72619-nvidia-geforce-gtx-1080-review-19.html

http://www.hardwarecanucks.com/foru.../72619-nvidia-geforce-gtx-1080-review-21.html

 

[DavidGraham]

question is, is this mere amd tuning or is this indicative of future dx12 titles with async compute performance?

I think its most probably just the usual vebndor optimizations, the engines of Ashes and HITMAN usually performed better on AMD hardware even in their previous incarnations. And they continue to do the same here. Yet If you look at ROTTR, Fable benchmark and Gears of Wars for example, they show the opposite, and run better on NVIDIA hardware.

 

[Orion]

I think its most probably just the usual vebndor optimizations, the engines of Ashes and HITMAN usually performed better on AMD hardware even in their previous incarnations. And they continue to do the same here. Yet If you look at ROTTR, Fable benchmark and Gears of Wars for example, they show the opposite, and run better on NVIDIA hardware.

I've heard gears was abysmal port. And Rise mysteriously shows drop in performance going from dx11 to dx12 in some benchmarks on fury, and in other boards there's been comments that enabling or disabling async does nothing on amd performance, which if true seems suspicious.

Don't know about fable benchmark.

 

[Silent_Buddha]

Anandtech Preview has landed: http://www.anandtech.com/show/10326/the-nvidia-geforce-gtx-1080-preview/

So looking at that. At 4k it's anywhere from 24% to 38% faster than 980ti depending on the game. A very respectable performance increase especially considering the sizes of the relative chips.

The gap to the Fury X is generally much larger in Dx11. However, if the game was built from the ground up for Dx12 the gap shrinks significantly to only about 6% to 12% faster than Fury X. Of course, that's a sample size of a whole 2 games. We'll have to wait and see if Nvidia can increase their performance in Dx12 and/or if AMD continues to do extremely well in Dx12.

Regards,
SB

 

[DavidGraham]

Gears port was indeed abysmal, still after patches abd fixes AMD performance suffers compared to NVIDIA. And Enabling/Disabling Async usually results in minimal differences on all hardware and in all DX12 games, maybe like 10%.

 

[Razor1]

I've heard gears was abysmal port. And Rise mysteriously shows drop in performance going from dx11 to dx12 in some benchmarks on fury, and in other boards there's been comments that enabling or disabling async does nothing on amd performance, which if true seems suspicious.

Don't know about fable benchmark.

Gears performance was fixed for AMD hardware within a week or two after release.

And async needs to be tailored for different hardware, each hardware gen needs different paths.

 

[Silent_Buddha]

I think its most probably just the usual vebndor optimizations, the engines of Ashes and HITMAN usually performed better on AMD hardware even in their previous incarnations. And they continue to do the same here. Yet If you look at ROTTR, Fable benchmark and Gears of Wars for example, they show the opposite, and run better on NVIDIA hardware.

ROTTR was a Dx11 game with Dx12 tacked on. Fable uses UE4 which is also primarily a Dx11 engine with Dx12 added on, AFAIK. GeOW uses UE4 also doesn't it? AOTR and Hitman were designed with Dx12 from the ground up.

Many games will likely go UE4, so Nvidia should maintain their engine advantage there. It'll be interesting to see what DICE end up doing with Frostbite for Dx12.

Regards,
SB

 

[Razor1]

GOW uses the UE3 engine. Hitman's engine wasn't built from ground up with DX12, they talked about it in GDC, the developer stated that if they could do it from ground up for DX12 they probably would have had better results for DX12 features.

 

[Raqia]

So looking at that. At 4k it's anywhere from 24% to 38% faster than 980ti depending on the game. A very respectable performance increase especially considering the sizes of the relative chips.

The gap to the Fury X is generally much larger in Dx11. However, if the game was built from the ground up for Dx12 the gap shrinks significantly to only about 6% to 12% faster than Fury X. Of course, that's a sample size of a whole 2 games. We'll have to wait and see if Nvidia can increase their performance in Dx12 and/or if AMD continues to do extremely well in Dx12.

Regards,
SB

I hope AMD pulls forward Vega as rumored and fuels a war at the upper echelons of performance parts. Presumably NVidia has room to breathe as it probably already has a working Titan version of this thing waiting in the wings based on a ~600mm^2 die based off the GP100 w/ HBM2. They'll also presumably cripple DP FP throughput and up memory speeds.

 

[DavidGraham]

Also there are other dx12 games that show the signs of target vendor optimizations, for example Forza 6 Apex runs far better on NVIDIA, Quantum Break runs better on AMD. While Killer instinct is in netural ground I guess. (I need more research on it)

I Don't even know why people get so worked up with Ashes. Its not a pretty game. In fact it's downright ugly and it's not popular either. Other DX12 games are much more popular and prettier. Like Tomb Raider, Hitman and Quantum Break.

 

[Deleted member 13524]

AFAIK, Return of the Tomb Raider is the only game whose DX12 brought a performance deficit to both AMD and nvidia cards (but with AMD losing more performance, or it wouldn't be a gameworks title), while also losing support for Crossfire and bringing only VXAO which is never enabled in these large-scale comparisons.
The devs themselves admitted that the only optimization feature they could take out of the DX12 was better multi-core utilization, which is worthless because this was probably the most threaded and least cpu-dependant game of all time in its original DX11 version.


In all fairness, benching ROTR in DX12 mode is just ridiculous.

 

[Razor1]

Well thats the issue, both IHV"s have games that just work better on each others hardware even in DX12, so we can always just eliminate those outliers, and we end up with the same thing we are at now with DX11 games. Some games will work better on nV some games will work better on AMD.