computerbaseDota 2 Vulkan benchmarks

[csbin]

http://www.computerbase.de/2016-05/...gramm-dota-2-reborn-1920-1080-i7-6700k-replay

 

[frozentundra123456]

Frametimes are amazingly low for the 980Ti, not so good for Fury.

 

[TypoFairy©]

As a person who doesn't own this game, I would just like to ask if Open gl in this game is equal to direct x 11 in Graphical quality? I'm surprised there's such a huge performance delta between dx 11 and open gl on the 980 ti.

 

[ShintaiDK]

Those Fury Frame times are awful compared. But it seems to be a random toss up for what IHV they focus mostly on.

 

[antihelten]

That's a bit weird, Fury X sees a performance improvement on the high end CPU (13.2%) and performance degradation on the low end CPU (-10.2%).

980 Ti see improvement on both CPUs, but a significantly bigger one on the high end CPU (23.5%) than the low end CPU (12.6%).

One would think that it should be the other way around, given that any CPU bottlenecks are obviously much bigger with the low end CPU.

 

[therealnickdanger]

These results seem about in line with Phoronix - Vulkan is currently slower overall than OpenGL, even with CPU usage being way down. I'm sure the next wave of driver or game updates will resolve it. They have a Linux benchmark utility for Dota 2 at Phoronix.

Is there a Windows equivalent timedemo? I'd like to run it on a couple systems.

 

[TheELF]

That's a bit weird, Fury X sees a performance improvement on the high end CPU (13.2%) and performance degradation on the low end CPU (-10.2%).

980 Ti see improvement on both CPUs, but a significantly bigger one on the high end CPU (23.5%) than the low end CPU (12.6%).

One would think that it should be the other way around, given that any CPU bottlenecks are obviously much bigger with the low end CPU.

Focus on the word replay in those benches...

 

[antihelten]

Focus on the word replay in those benches...

Why would that matter, unless they used two different replays, that differed wildly in CPU load (although I very much doubt they did that)

 

[Denithor]

Both frametime charts are run on the same CPU, right? IE no funny business of Fury on X4 while 980 Ti is on i7? Just asking because CPU is not specified in the header bar.

Wonder why no 1080 in the lineup. None available I would assume.

 

[PhonakV30]

on the GTX 980Ti , Vulkan is slower than OpenGl , seems legit.

Edit : Why would nvidia go Vulkan when Nvidia doesn't get much benefit from Vulkan ?

 

[csbin]

on the GTX 980Ti , Vulkan is slower than OpenGl , seems legit.

Edit : Why would nvidia go Vulkan when Nvidia doesn't get much benefit from Vulkan ?

Vulkan≈DirectX 12

 

[moonbogg]

The take away here is that OpenGL is badass. Its always been the best and fastest.

 

[IllogicalGlory]

The take away here is that OpenGL is badass. Its always been the best and fastest.

Why is it so far behind in The Talos Principle?

 

[moonbogg]

Why is it so far behind in The Talos Principle?

That game is irrelevant. Try to find some where OpenGL does better. Then we can talk.

 

[Bacon1]

That game is irrelevant. Try to find some where OpenGL does better. Then we can talk.

Its the only other DX and OpenGL game with Vulkan at the moment.

Saying "only show results where OpenGL is better" to discuss "is opengl better" is mind boggling.

Even Doom isn't valid because it doesn't have a DX pipeline to compare against. Sure you can compare OpenGL vs Vulkan, but not all 3 to show "which is best of the best".

 

[IllogicalGlory]

The real answer is that neither one is better. They seem to support mostly the same features and both can be used to create good-looking, well performing games, but there's a reason developers prefer D3D, and it isn't the money. It's simply that D3D is easier to work with, and isn't inferior.

It took iD Tech 3 games to get their engine right. Wolfenstein and Rage both had lots of issues. Never played Rage, but I know most PC users didn't like it, while Wolfenstein was an improvement, we had to contend with a 60 FPS cap as well as texture pop-in, along with no multi GPU support. Compare that to DICE games and there's simply no contest.

 

[monstercameron]

Its the only other DX and OpenGL game with Vulkan at the moment.

Saying "only show results where OpenGL is better" to discuss "is opengl better" is mind boggling.

Even Doom isn't valid because it doesn't have a DX pipeline to compare against. Sure you can compare OpenGL vs Vulkan, but not all 3 to show "which is best of the best".

Obvious sarcasm is obvious...

 

[Bacon1]

Obvious sarcasm is obvious...

His post wasn't sarcasm look at his posts from the Doom thread.

 

[Phynaz]

Its the only other DX and OpenGL game with Vulkan at the moment.

Saying "only show results where OpenGL is better" to discuss "is opengl better" is mind boggling.

Even Doom isn't valid because it doesn't have a DX pipeline to compare against. Sure you can compare OpenGL vs Vulkan, but not all 3 to show "which is best of the best".

So you are saying a sample size of 1 is relevant.

 

[TheELF]

Why would that matter, unless they used two different replays, that differed wildly in CPU load (although I very much doubt they did that)

In replays the system already knows what comes next so you can channel all your resources into rendering frames,in real time gaming the card has to wait for input from all players to know what to do next.
At least that's my thinking,if someone with lots of cores could tell us CPU utilization difference between real play and replay we could make sure.

Yay,works as promised,no driver thread=the other threads get to share the freed up resources=moar FPS.
https://www.youtube.com/watch?v=-cDAWW8nkGU

 

[Bacon1]

So you are saying a sample size of 1 is relevant.

Well there are two games currently that support DX11, OpenGL and Vulkan.

How exactly can we get more than 2 sample games?

Are you saying that the single OP sample is better than having two?

 

[Raising]

What's the point of higher fps if the game stutters like crazy while panning the camera ?

Only dx11 has g-sync functional, I can play the game without any stuttering or input lag.

 

[railven]

Well those results weren't what I expected.

More so the performance lost on a weaker CPU considering all the hoolah about how it's suppose to help slower CPUs.

 

[MajinCry]

What's the point of higher fps if the game stutters like crazy while panning the camera ?

Only dx11 has g-sync functional, I can play the game without any stuttering or input lag.

Then limit your fps. You're going to have framerates all over the place, no matter what CPU you have, and no matter what API the game is using.

 

[sirmo]

Well those results weren't what I expected.

More so the performance lost on a weaker CPU considering all the hoolah about how it's suppose to help slower CPUs.

Vulkan is relatively new. And we don't know to what degree Vulkan's capabilities are being leveraged by Dota 2's engine.

The developers get way more control, but it's entirely up to them to utilize that new found freedom. Judging by how much the weaker CPU impacts the performance, I'd guess they aren't really leveraging the power of compute shaders to its full potential.