Doom Vulkan vs Open GL performance

sqwerty · 16 Jul 2016 at 13:11

TheRealDeal said:
I hope you are right. It would be nice to get more leaps in performance like we used to.

We already have if you compare GTX 970 and 980 to 1070 and 1080, it's just Nvidia is an extortionist.

Rroff · 16 Jul 2016 at 13:17

sqwerty said:
We already have if you compare GTX 970 and 980 to 1070 and 1080, it's just Nvidia is an extortionist.

1070/1080 performance leaps are pathetic compared to the past.

sqwerty · 16 Jul 2016 at 13:21

Rroff said:
1070/1080 performance leaps are pathetic compared to the past.

That's because they started selling their mid range GPU as top end AKA GTX 680/780/980 and now 1080 milk it dry edition.

The real top end were the Titan/Ti hidden away till the opportunity arose i.e competition.

Lokken86 · 16 Jul 2016 at 13:24

So now that results are heavily in my favour I can say I told you the Fury would beat the 980ti in 4k. Even overclocked 980ti vs stock fury pro. Its Iron Clad now.

Gregster · 16 Jul 2016 at 13:55

Lokken86 said:
So now that results are heavily in my favour I can say I told you the Fury would beat the 980ti in 4k. Even overclocked 980ti vs stock fury pro. Its Iron Clad now.

But where is it Ironclad? They tested the Fury X against the 1070 (not the Fury against the 980Ti). The 980Ti overclocked is pretty close to my 1080 and I am sure you watched my vid of the same areas and you can see from a quick comparison watching that the 1080 beats out the Fury X at 4K (as it should) and the 1080 isn't running with Async as of yet and when it does, that gap will get wider.

By all means though, notch it as a win if it lets you feel good

Rroff · 16 Jul 2016 at 13:57

Lokken86 said:
So now that results are heavily in my favour I can say I told you the Fury would beat the 980ti in 4k. Even overclocked 980ti vs stock fury pro. Its Iron Clad now.

Make the most of it

supposedly a driver update coming for nVidia with updated Vulkan support - though whether that will have much impact on the performance or not is another matter.

Lokken86 · 16 Jul 2016 at 14:01

Gregster said:
But where is it Ironclad? They tested the Fury X against the 1070 (not the Fury against the 980Ti). The 980Ti overclocked is pretty close to my 1080 and I am sure you watched my vid of the same areas and you can see from a quick comparison watching that the 1080 beats out the Fury X at 4K (as it should) and the 1080 isn't running with Async as of yet and when it does, that gap will get wider.

By all means though, notch it as a win if it lets you feel good

It was a joke. Comon lighten up a little.

Oh and cheers for the vid. Your 1080 is about 5%-10% faster than my result I would say.

Oh and Matt what software did you use to record I cant get anything to work.

Doom112 · 16 Jul 2016 at 14:20

Lokken86 said:
It was a joke. Comon lighten up a little.

Oh and cheers for the vid. Your 1080 is about 5%-10% faster than my result I would say.

Oh and Matt what software did you use to record I cant get anything to work.

You can just post the screenshot. I would love to compare with you and If you do then please just post the screenshot of first level UAC.

Orangey · 16 Jul 2016 at 14:22

Rroff said:
1070/1080 performance leaps are pathetic compared to the past.

I hope it's mere sloth and avarice, not that they've hit a brick wall just yet.

Back in the day new arch + shrink with mooned clocks would've been something VERY special, no matter if it's a mid-range part or not.

melmac · 16 Jul 2016 at 14:26

Lokken86 said:
It was a joke. Comon lighten up a little.

It's only a joke when it's against AMD, don't you know.

You see the same guys that jumped on you for saying that you didn't agree with a benchmark result favouring Nvidia, they didn't jump on the Doom112 guy when he disagreed with a benchmark that favoured AMD, where this whole comparing benchmarks started.

And they wanted you to be the better man. :rolleyes:

Seanspeed · 16 Jul 2016 at 14:33

melmac said:
It's only a joke when it's against AMD, don't you know.

You see the same guys that jumped on you for saying that you didn't agree with a benchmark result favouring Nvidia, they didn't jump on the Doom112 guy when he disagreed with a benchmark that favoured AMD, where this whole comparing benchmarks started.

And they wanted you to be the better man.

Please stop with the us vs them stuff. It's insanely hypocritical and it's tiring. So very tiring.

Seanspeed · 16 Jul 2016 at 14:38

Rroff said:
I like that he had a good balanced take on the end towards the various cards and how it is early days yet - commenting on the vsync thing, etc.

EDIT: He might be on the money there for why Kepler isn't doing well - the numbers suspiciously match up with vsync multiples - it is like the game is constantly bouncing off 2 down vsync multipliers for the potential FPS even with vsync disabled.

My guess with Kepler is simply that it's not well suited to many of the new low level features. A lot of the improvements with Maxwell were DX12-focused and while Kepler was still DX12-capable, it lacks many of these improvements.

Obviously Vulkan isn't DX12, but they're different sides of the same coin rather than anything completely different.

People also gotta remember that low level means a thin driver. Meaning that Nvidia/AMD play a far smaller role in how well something runs.

Rroff · 16 Jul 2016 at 14:41

Seanspeed said:
My guess with Kepler is simply that it's not well suited to many of the new low level features. A lot of the improvements with Maxwell were DX12-focused and while Kepler was still DX12-capable, it lacks many of these improvements.

Obviously Vulkan isn't DX12, but they're different sides of the same coin rather than anything completely different.

People also gotta remember that low level means a thin driver. Meaning that Nvidia/AMD play a far smaller role in how well something runs.

Not really expecting much in the way of results for Kepler - but currently Kepler cards aren't working right at all - given the way the framerate is dropping out it looks like even with V-Sync fully disabled Vulkan is adhering to some kind of V-Sync functionality resulting in artificial performance drop down the V-Sync multiplier ladder.

d_brennen · 16 Jul 2016 at 14:48

Anyone noticed a weird stuttery feeling despite frame rate being good? Not sure if it's mouse smoothing or something, but it don't seem right to me

shankly1985 · 16 Jul 2016 at 14:50

Nope soon smooth here. Been getting into the online

1440p ultra settings close to 100fps

melmac · 16 Jul 2016 at 15:23

Seanspeed said:
Please stop with the us vs them stuff. It's insanely hypocritical and it's tiring. So very tiring.

Go away Sean, with your complete lack of understanding, and your sanctimonious crap. My post wasn't about them or us, my post is pointing out the very hypocrisy that you are talking about.

Griffildur · 16 Jul 2016 at 15:24

so now that we had our fun and benchmarks, is there any reason why one with an Nvidia card would use Vulkan over OpenGL? Personally I don't see one, have to say.

DmitryD · 16 Jul 2016 at 15:26

sqwerty said:
We already have if you compare GTX 970 and 980 to 1070 and 1080, it's just Nvidia is an extortionist.

It's the price of new performance gains.

sqwerty · 16 Jul 2016 at 15:26

DmitryD said:
It's the price of new performance gains.

It's the price of zero competition.

Ian Evey · 16 Jul 2016 at 15:30

Griffildur said:
so now that we had our fun and benchmarks, is there any reason why one with an Nvidia card would use Vulkan over OpenGL? Personally I don't see one, have to say.

Right now there is no reason to use Vulcan on nvidia.

Imagine if a game is designed only for Vulcan :eek:

might aswell use OpenGL instead as that's a bigger market for devs.