Yep, good 30%+ from Vulkan alone, and another 10-20% from async compute support.
103 -> 136 -> 153
78 -> 113 -> 125
Can you tell me what's wrong with AMD's OpenGL for Doom?
http://www.guru3d.com/articles_pages/amd_radeon_r9_rx_480_8gb_review,14.html
Compare the 290, 290X, 390, 390X vs 970 and 980. Around that ballpark.
Certainly not as horrible as Project Cars deficit.
RX 480 gains up to 50% with Vulkan for example with AC enabled. So does Fury Nano (PCGH claims +50%).
https://www.youtube.com/watch?v=GlzKPBIjZPo
Btw, a 50% faster Nano is actually GTX 1080 or faster.
Without sources or further explanation, I don't see why anyone would take that seriously. Talk about FUD. Maybe you'd like to explain why console devs, dealing with GCN chips that have fewer than 1024 shaders to begin with in the case of the Xbone, benefit from 'what AMD calls asynchronous compute'.
http://forums.anandtech.com/showthread.php?p=38350234&highlight=#post38350234
http://forums.anandtech.com/showpost.php?p=38350354&postcount=76
Not going to repeat myself for every single time people try to FUD around the issue.
And i'm not going to rehash AMD/ATi's entire sordid history with OpenGL. It's common knowledge. Even if you may not possess it.
I can't seem to take a screenshot in DOOM while running the Vulkan API. BUT. I am running every setting at Ultra and my minimum FPS is now 60. In OpenGL it was 35-40. Max is 125, before it was about 80. Average seems to be about 110ish.
Core i7-4720HQ, GTX970m @ 1.3ghz core/3ghz ram. I will update the post tomorrow with a picture of the stats set to Nightmare so you can see the frametimes - but they were locked to 16ms, IIRC.
As soon as people talk about why Async Compute benefits AMD GCN because it's not fully using it's shaders, they already lost because they are repeating the same propaganda lies.
It's Multi-Engine. Targeting ROPs & DMAs with queues in parallel while work is running on Shaders. All GPUs have these Rasterizers or DMAs, yet in DX11 or hardware incapable of Multi-Engine, the queues are going in serial, when Shaders are running, ROPs/DMAs are idling and vice versa.
See, the smart people figured that Rasterizers or DMAs should be able to handle a lot of the rendering tasks in parallel to shaders, so they made next-gen APIs to take advantage of this. Somebody forgot to inform NV of this transition ahead of time, and that's all it is, NV makes hardware that's great for the DX11 era.
But we're told, again, this time from id Software, that NV will work on their Async Compute drivers...
Well something's not right. Anyone else getting stutter on their 970 with Vulkan :?
Getting really inconsistent frame times on my 980ti. Looks and feels noticeably less smooth than OpenGL even though the framerate is higher. Maybe it's something they'll sort in a patch
Same thing is happening with my 980 Ti. My FPS improved 20% or so according to the overlay, but it looks like it's running closer to 30 FPS instead of the 110-130 FPS it shows. Meanwhile OpenGL is running 70-90 FPS, and feels like it.
I tried this with latest drivers on my GTX 780 with open gl vs vulkan and I didnt notice any difference at 1440p with mostly high settings. Framerate was roughly the same too, if not 1-2 FPS worse on vulkan.... Maybe Im doing something wrong?. Im on latest Nvidia driver
Yeh I saw no difference with my 780. So much for getting more performance out of older cards...
Im curious. What are some of the things that the ROP can do in parallel given that their fundamental functionality is to grab the pixel/texel data (which again comes from the front-end of the graphics pipeline), blending (maths) them into the final pixel value and writing them back into the memory for display?
DMAs? direct memory access for what graphics operations?
Here's a second opinion for the unbelieving:
Looks like AMD performance is right where it should be, maybe very slightly lower than some other games. Still, the 980 appears much faster than usual.
Just got back home and installed the new patch, and holy shit, my Fury gains around 50% in certain scene, general perf increase is around 40%. Again, holy shit.
http://imgur.com/a/Ah02x
Edit: I forgot to turn off Vsync. Without Vsync, my Fury gains another 7fps, awesome. Here's the new screenshots: http://imgur.com/a/yGBpD
It's funny how the propaganda falls apart when probed, right?
Im curious. What are some of the things that the ROP can do in parallel given that their fundamental functionality is to grab the pixel/texel data (which again comes from the front-end of the graphics pipeline), blending (maths) them into the final pixel value and writing them back into the memory for display?
DMAs? direct memory access for what graphics operations?
Cerny is convinced that in the coming years, developers will want to use the GPU for more than pushing graphics -- and believes he has determined a flexible and powerful solution to giving that to them. "The vision is using the GPU for graphics and compute simultaneously," he said. "Our belief is that by the middle of the PlayStation 4 console lifetime, asynchronous compute is a very large and important part of games technology."
Cerny expects developers to run middleware -- such as physics, for example -- on the GPU. Using the system he describes above, you can run at peak efficiency, he said.
"If you look at the portion of the GPU available to compute throughout the frame, it varies dramatically from instant to instant. For example, something like opaque shadow map rendering doesn't even use a pixel shader, it’s entirely done by vertex shaders and the rasterization hardware -- so graphics aren't using most of the 1.8 teraflops of ALU available in the CUs. Times like that during the game frame are an opportunity to say, 'Okay, all that compute you wanted to do, turn it up to 11 now.'"
The team, to put it mildly, had to think ahead. "The time frame when we were designing these features was 2009, 2010. And the timeframe in which people will use these features fully is 2015? 2017?" said Cerny.
It's funny how the propaganda falls apart when probed, right?
Im curious. What are some of the things that the ROP can do in parallel given that their fundamental functionality is to grab the pixel/texel data (which again comes from the front-end of the graphics pipeline), blending (maths) them into the final pixel value and writing them back into the memory for display?
DMAs? direct memory access for what graphics operations?
It's not even propaganda. He doesn't really know what he's talking about. Perhaps he thinks modern games spend half of the frame time blitting stuff around.
I'm curious as well. Did it not occur to you that the better question would be "What are some of the things that the ROPs and Shaders can do in parallel?". Not trying to attack you, just pointing out how easy it is to misquote people and be an unintentional flamebait (as what happened).Im curious. What are some of the things that the ROP can do in parallel ...
Just got back home and installed the new patch, and holy shit, my Fury gains around 50% in certain scene, general perf increase is around 40%. Again, holy shit.
http://imgur.com/a/Ah02x
Edit: I forgot to turn off Vsync. Without Vsync, my Fury gains another 7fps, awesome. Here's the new screenshots: http://imgur.com/a/yGBpD