RX 480 performs ~30% faster with Vulkan!...

[TrigrH]

...compared to OpenGL in DOOM.

Source: http://www.gamersnexus.net/game-bench/2510-doom-vulkan-vs-opengl-benchmark-rx-480-gtx-1080

Video: 

 

Pretty amazing gains, its interesting to see that as the resolution increases the scaling with the new nvidia graphics cards vanishes.

 

[vanished]

Interesting.  Seems to be benefiting AMD more than nvidia right now.  Can't say I'm surprised.

[Rifian1]

Can some one do benchmark with GtX 950?

[manikyath]

i'd be interested to see how the vareous cards fare in DX11 vs openGL, maybe the RX480 is just not a good openGL card?

[ivan134]

5 minutes ago, manikyath said:

i'd be interested to see how the vareous cards fare in DX11 vs openGL, maybe the RX480 is just not a good openGL card?

None of the AMD cards are good OpenGL cards. The 480 isnt the only one benefiting from this patch

[manikyath]

1 minute ago, ivan134 said:

None of the AMD cards are good OpenGL cards. The 480 isnt the only one benefiting from this patch

i meant out the 3 selected cards.

 

thats kinda why i'd like to see DX vs OGL benchmarks. because people will most likely be using this news as "oh the RX480 is 30% faster now that vulkan is a thing" whereas it may also lose 20% going from DX to OGL, basicly putting it back "on par"

[gilang01]

[ivan134]

2 minutes ago, manikyath said:

i meant out the 3 selected cards.

 

thats kinda why i'd like to see DX vs OGL benchmarks. because people will most likely be using this news as "oh the RX480 is 30% faster now that vulkan is a thing" whereas it may also lose 20% going from DX to OGL, basicly putting it back "on par"

It's Vulkan performance is faster than the 970's OpenGL performance and Nvidia has really good OpenGL performance. Doom doesn't have dx 11.

[manikyath]

2 minutes ago, ivan134 said:

It's Vulkan performance is faster than the 970's OpenGL performance and Nvidia has really good OpenGL performance. Doom doesn't have dx 11.

didnt necessarily say it had to be done in doom, i know doom doesnt have DX

 

what i'm trying to get at here is that that "30% bump" may be less spectacular than this test portrays if the RX480 is the faster card from the start, and just kneecapped by openCL.

[zMeul]

27 minutes ago, Ryan_Vickers said:

Interesting.  Seems to be benefiting AMD more than nvidia right now.  Can't say I'm surprised.

because AMD has so shitty OpenGL drivers

it's hilarious a brand new 14nm GPU architecture needs a new API to catch up to a 28nm arch 

 

one other thing to note, different media outlets have revealed different results 

[Khvarrioiren]

7 minutes ago, zMeul said:

because AMD has so shitty OpenGL drivers

it's hilarious a brand new 14nm GPU architecture needs a new API to catch up to a 28nm arch 

 

one other thing to note, different media outlets have revealed different results 

The performance variaton seems ridicuslously inconsistent. Guess it's down to hardware differences between test paltforms or something that causes the discrepancy.

[Curufinwe_wins]

7 minutes ago, zMeul said:

because AMD has so shitty OpenGL drivers

it's hilarious a brand new 14nm GPU architecture needs a new API to catch up to a 28nm arch 

 

one other thing to note, different media outlets have revealed different results 

Probably just like aots, and other early dx12 titles they are pretty poorly optimized implementations. 

[ivan134]

It's even faster once asynchronous compute is used. As of right now you need to use TSSAA to make use of async compute. The devs said they'll add wider support later.

 

[Sakkura]

I'm not one to shit on AMD, but this is just a reflection of their poor OpenGL performance. Vulkan brings their cards up where they should be. Plus maybe a couple percent for the 480. 

[Spectrez]

going to be interesting too see if Nvidia's GPUs will benefit from Async computing

[nobelharvards]

.

[TrigrH]

3 minutes ago, nobelharvards said:

The first two graphs in the OP are the same thing.

edit fixed thanks!, please edit your post so its not quoting the entire article..

[AlTech]

1 hour ago, manikyath said:

i'd be interested to see how the vareous cards fare in DX11 vs openGL, maybe the RX480 is just not a good openGL card?

if anything I'd say Nvidia cards aren't that good at OpenGL compared to AMD.

 

This is just one game. Wait for some more to make a judgement.

[manikyath]

1 minute ago, AluminiumTech said:

if anything I'd say Nvidia cards aren't that good at OpenGL compared to AMD.

 

This is just one game. Wait for some more to make a judgement.

well sure, but seeing some DX/OGL benchmarks (in multiple other games) would paint more of a picture already.

[BakerT]

Do remember that this is still a "beta". Might be why the sites get different results.

 

[zMeul]

1 hour ago, Khvarrioiren said:

The performance variaton seems ridicuslously inconsistent. Guess it's down to hardware differences between test paltforms or something that causes the discrepancy.

I think it's the settings, for example: not using TXAA TSSAA disables async on AMD's cards

Edited July 12, 2016 by zMeul

[Khvarrioiren]

9 minutes ago, zMeul said:

I think it's the settings, for example: not using TXAA disables async on AMD's cards

It's actually TSSAA but yeah, might be that.

[AlTech]

9 minutes ago, zMeul said:

I think it's the settings, for example: not using TXAA disables async on AMD's cards

Why is that?

[Khvarrioiren]

2 minutes ago, AluminiumTech said:

Why is that?

They haven't gotten around to implementing it to other AA modes yet.

[AlTech]

2 minutes ago, Khvarrioiren said:

They haven't gotten around to implementing it to other AA modes yet.

Why does Async need to be tied to AA though?