Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Performance will improve as drivers mature.
Wait a minute, how come 1080 is not supported by 368.19 if 1070 was tested on those drivers? I mean, make no mistake, the whole thing is a disaster, but that's suspect. What is also suspect is that i can't find 368.19 release notes.
I'm excited to see benchmarks, unfortunately the games I care about most are very biased toward nvidia in terms of performance, to the point where two cards that are dead even in most games will have a ~20% nvidia advantage. Still this seems to remind me about way back when I ran dual 4850s in crossfire. That was a great budget setup for it's time and I could see doing something like that again with a secondary rig.
I have a few friends who are using motherboards with only a 16x/4x configuration, which means SLI will not work. Will crossfire still work on a 16x/4x motherboard?
Idk if agree with this regardless of what's new.....I been impressed with amds drivers over last 2 years...but there always room for improvement yes?No, GCN architecture it's already mature since HD7000 Series, there's nothing to improve there more than the physical updates to the die of GCN 4. (1.4). you are taking polaris as brand new arch when that's not true.
You are ignoring facts. True it's not a brand new architecture. But they have upgraded he whole front end. Please read up and it's not a lie.No, GCN architecture it's already mature since HD7000 Series, there's nothing to improve there more than the physical updates to the die of GCN 4. (1.4). you are taking polaris as brand new arch when that's not true.
Idk if agree with this regardless of what's new.....I been impressed with amds drivers over last 2 years...but there always room for improvement yes?
You are ignoring facts. True it's not a brand new architecture. But they have upgraded he whole front end. Please read up and it's not a lie.
I don't know if this was meant towards me or not. You didn't qoute me but did post right after me.There will always be people trying to shit on this good news. If you're a pc gamer this should excite you. If it doesn't then somethings up.
By some benchmarks you mean only benchmark of 3D mark 11 where it's beating the 980? Scratching my head here please enlighten me what other benchmarks you have seen where it's just on par with 970? Of course there are no real benchmarks but then you say from what you seen its on par with 970 and then let's wait for benchmarks. Lol
Out of curiosity, what card are you currently using? If you ask me thats a big factor on how each of us build an opinion of these cards. Me for example its a 60% increase in performance over what im using now. (Not a bad jump for me) Those already using a 290/ 970 are not impressed of course. (and i can understand why)I don't know if this was meant towards me or not. You didn't qoute me but did post right after me.
To me the PC/overclocking market has been so stagnant the last 4-5 years, I've moved on to other things about a year ago. I check in on new stuff still though, and I'm sorry but a card that continues with the same generational leap we have been getting for the past 4-5 generations doesn't excite me. Like I said the best thing about this card is Its priced about $50 less then it normally would be, which makes it a good value. But it's not a revelation that many are acting like it is.
So again nothing bad about the card and getting x90 level performance would be fully expected for a new x80 card. The price starting at $50 less is the only thing that stands out.
I don't care if I run AMD or Nvidia, I go where the value is. So I may end up with this card, but I'll wait to see the 1060 first.
I don't know if this was meant towards me or not. You didn't qoute me but did post right after me.
To me the PC/overclocking market has been so stagnant the last 4-5 years, I've moved on to other things about a year ago. I check in on new stuff still though, and I'm sorry but a card that continues with the same generational leap we have been getting for the past 4-5 generations doesn't excite me. Like I said the best thing about this card is Its priced about $50 less then it normally would be, which makes it a good value. But it's not a revelation that many are acting like it is.
So again nothing bad about the card and getting x90 level performance would be fully expected for a new x80 card. The price starting at $50 less is the only thing that stands out.
I don't care if I run AMD or Nvidia, I go where the value is. So I may end up with this card, but I'll wait to see the 1060 first.
I had a 970 but sold it to a friend and went back to the 670 I had before it. Needed some quick cash. Honestly the 670 still plays what I play fine. But requirements are increasing lately, especially for the Win Store Xbox games I like.Out of curiosity, what card are you currently using? If you ask me thats a big factor on how each of us build an opinion of these cards. Me for example its a 60% increase in performance over what im using now. (Not a bad jump for me) Those already using a 290/ 970 are not impressed of course. (and i can understand why)
Agreed. $200-300 cards always offer best bang for buck for higher end gaming, but for those that want that extra mile have to pay more then 50% more for the premium experience. Nothing new, and I've done that before but always finding my self back in the $200 card range because I don't play that much anymore.I can certainly understand AMD's reasons for aiming at Mainstream, as that is where most cards are sold. It doesn"t suit my needs as I would like something a fair bit faster. I'm going to hold out and see whet the $300 Polaris brings. That I hope will bring performance around the 1070 mark. At least that is what I am hoping for.
I read that they got the phrasing wrong and that for whatever reason, scaling was only 51% instead of the normal amount you would expect to see from two cards. Not sure the root cause, but pre-release drivers are likely to be part of the issue.
No the phrasing wasn't wrong, they confirmed that 51% is the %time gpu bound for normal batches. Actual xfire scaling is 83%
So the performance of one card is 33.9fps, at 1080p crazy
A 390x does ~38fps in 1440p crazy
A fury does around 41-42 fps in 1440p crazy
So if this card is performing worse, at 1080p, than the 390x is at 1440p, explain to me how this is fury level performance exactly
Boy you are reading way way way too much in to that 51%. It was marketing bull shit. I don't know what the fuck he meant by that, but you seem to be blowing it out of proportion. SMH. You are making numbers up at this point.
I think at this point you are pulling on straws to prove your point that this card is slower. Common relax lets wait for the reviews. This 51% crap is getting annoying cuz I have no idea what the fuck raja was saying. It was to make 2 cards look good thats all.
What the hell are you saying ? I'M making up numbers? 51% is the "percent GPU bound" for normal batches.
The card has 1.83 CF scaling as confirmed by AMD. That means 34fps for a single card. At 1080p crazy.
A 390x does around 45 fps at those settings.
Before you accuse me of making numbers up, use your brain.
The 51% thing is annoying because you don't understand it ? A great many things must annoy you then. In fact you must find reading about hardware very annoying.
Actually Hallock stated that the CF scaling was " Scaling is 151% of a single card."What the hell are you saying ? I'M making up numbers? 51% is the "percent GPU bound" for normal batches.
The card has 1.83 CF scaling as confirmed by AMD. That means 34fps for a single card. At 1080p crazy.
A 390x does around 45 fps at those settings.
Before you accuse me of making numbers up, use your brain.
The 51% thing is annoying because you don't understand it ? A great many things must annoy you then. In fact you must find reading about hardware very annoying.
Are you saying it was scaling at 1.83% in that benchmark? I never heard that anywhere or seen it. Yea I think people are reading way too much in to it. I am sure when the reviews come out watch this card be faster than 390x. If its clocked at 1266 it already makes up the shader difference from 390x and with front end improvements I expect it to be faster. I have absolutely no reason to believe it will be slower than a 390x when the specs absolutely speak otherwise.
No he didn't.Actually Hallock stated that the CF scaling was " Scaling is 151% of a single card."
AMD_RobertTechnical Marketing[S] 185 points 16 hours ago*
Scaling is 151% of a single card.
//EDIT: To clarify this, the scaling from 1->2 GPUs in the dual RX 480 test we assembled is 1.83x. The OP was looking only at the lowest draw call rates when asking about the 51%. The single batch GPU utilization is 51% (CPU-bound), medium is 71.9% utilization (less CPU-bound) and heavy batch utilization is 92.3% (not CPU-bound). All together for the entire test, there is 1.83X the performance of a single GPU in what users saw on YouTube. The mGPU subsystem of AOTS is very robust.
YES HE DID.No he didn't.
51% is the %gpu bound for normal batches
83% is the CF scaling
YES HE DID.
Concerning the AOTS image quality controversy
it was posted today a convo he had after landing from flight.
All together for the entire test, there is 1.83X the performance of a single GPU in what users saw on YouTube.
there is 1.83X the performance of a single GPU in what users saw on YouTube.
1.83X the performance of a single GPU
I am not going to bother dissecting this completely, but I am going to point out the most obvious. The 480 crossfire benchmark was 1440p crazy, not 1080p crazy.
Then dont buy their cards.....Why are you trolling this thread? Be gone! lol34fps for a single card. At 1080p
If this is correct then this card is a piece of junk. Another AMD flop and they should be out of business by this time next year.
HERE FOR YOU since you seem to miss the whole thing spelled out in the thread I link on the convo:Quoting you so I can shame you in the future.
READING COMPREHENSION. IMPROVE IT.
ask Robert Hallock some questions before he leaves the thread.
Some notable quotes:
Hi. Now that I'm off of my 10-hour airplane ride to Oz, and I have reliable internet, I can share some insight.
System specs:
- CPU: i7 5930K
- RAM: 32GB DDR4-2400Mhz
- Motherboard: Asrock X99M Killer
- GPU config 1: 2x Radeon RX 480 @ PCIE 3.0 x16 for each GPU
- GPU config 2: Founders Edition GTX 1080
- OS: Win 10 64bit
- AMD Driver: 16.30-160525n-230356E
- NV Driver: 368.19
- In Game Settings for both configs: Crazy Settings | 1080P | 8x MSAA | VSYNC OFF
Ashes Game Version: v1.12.19928
Benchmark results:
2x Radeon RX 480 - 62.5 fps | Single Batch GPU Util: 51% | Med Batch GPU Util: 71.9 | Heavy Batch GPU Util: 92.3% GTX 1080 – 58.7 fps | Single Batch GPU Util: 98.7%| Med Batch GPU Util: 97.9% | Heavy Batch GPU Util: 98.7%
The elephant in the room:
Ashes uses procedural generation based on a randomized seed at launch. The benchmark does look slightly different every time it is run. But that, many have noted, does not fully explain the quality difference people noticed.
At present the GTX 1080 is incorrectly executing the terrain shaders responsible for populating the environment with the appropriate amount of snow. The GTX 1080 is doing less work to render AOTS than it otherwise would if the shader were being run properly. Snow is somewhat flat and boring in color compared to shiny rocks, which gives the illusion that less is being rendered, but this is an incorrect interpretation of how the terrain shaders are functioning in this title.
The content being rendered by the RX 480--the one with greater snow coverage in the side-by-side (the left in these images)--is the correct execution of the terrain shaders.
So, even with fudgy image quality on the GTX 1080 that could improve their performance a few percent, dual RX 480 still came out ahead.
As a parting note, I will mention we ran this test 10x prior to going on-stage to confirm the performance delta was accurate. Moving up to 1440p at the same settings maintains the same performance delta within +/-1%.
Ok I got some questions, if you're not using vsync or locked fps why is the GPU usage so low and not at 100% pushing maximum FPS?
DX12 uses Explicit Multi-Adapter. The scaling depends on how mGPU is implemented into the engine, and future patches could boost scaling more for any vendor or any GPU combination that works. Besides that, migrating to full production-grade drivers would help. But as you can image, the drivers are still beta. I'm not promising earth-shattering improvements, here, but there are many variables in play that wouldn't be present with GPUs that have been released for 12+ months.
DX12 explicit multi-adapter is not CrossFire. We do not control how mGPU functions in DX12 or Vulkan.
BlitzWulf[email protected] XFX R9 390X@1180/1650 18 points 2 hours ago*
Thanks for the post, may I ask one question though? When it's claimed that there is only 51% GPU utilization does that mean 51% of each GPU is being utilized or that the performance scaling is equivalent to 151% of a single card?
- permalink
- AMD_RobertTechnical Marketing[S] 24 points an hour ago
Scaling is 151% of a single card.
Thank you very much for clearing that up ,If I could trouble you once more I have another question.
There has been footage/photos of DOOM running on an RX 480, some people have claimed that this demo was at 1080p resolution I am under the impression that all DOOM demos run on the RX 480 at Computex were using 1440p VSR on 1080p monitors,am I mistaken?
AMD_RobertTechnical Marketing[S] 11 points 55 minutes ago
1080p monitor running at 1440p with VSR.
34fps for a single card. At 1080p
If this is correct then this card is a piece of junk. Another AMD flop and they should be out of business by this time next year.
183% scaling yes, that is what AMD STATED on reddit.
This comes from AMD.
So I'm not pulling at straws, and I'm not making numbers up. 34 fps 1080p crazy settings. 5.8 tflops vs 5.9 on a 390x, yet 25% slower than it.
HERE FOR YOU since you seem to miss the whole thing spelled out in the thread I link on the convo:
AMD_RobertTechnical Marketing[S] 185 points 16 hours ago*
Scaling is 151% of a single card.
//EDIT: To clarify this, the scaling from 1->2 GPUs in the dual RX 480 test we assembled is 1.83x. The OP was looking only at the lowest draw call rates when asking about the 51%. The single batch GPU utilization is 51% (CPU-bound), medium is 71.9% utilization (less CPU-bound) and heavy batch utilization is 92.3% (not CPU-bound). All together for the entire test, there is 1.83X the performance of a single GPU in what users saw on YouTube. The mGPU subsystem of AOTS is very robust.
The OP was looking only at the lowest draw call rates when asking about the 51%.
The OP was looking only at the lowest draw call rates when asking about the 51%.
So the guy that made the entire thread and knew exactly what specs the systems were and fully explained it is lying about 151%? LOL talk about picking what you like and spinning it your way to support your argument.
If you look at the thread I linked there is NO edit it only says 151%. I even copied and you still claim I knew of the edit which obviously I did not. You are far too hostile and I would rather you and the rest of the Nvidia loving ilk stay out of the AMD subforum so those of us that are interested in the 480 can get info and enjoy the release.Jesus fucking christ which part of this do you not understand
One of two possible scenarios JustReason:
A) You're a troll and you're just spewing crap you know is false
B) You have serious trouble with reading comprehension
If you look at the thread I linked there is NO edit it only says 151%. I even copied and you still claim I knew of the edit which obviously I did not. You are far too hostile and I would rather you and the rest of the Nvidia loving ilk stay out of the AMD subforum so those of us that are interested in the 480 can get info and enjoy the release.
One question. What does "Scaling is 151% of a single card" mean? I thought in typical CF scaling that "100%" was absolute double performance of a single card or in other words "best case possible" ..I wont lie 151% confuses me
AMD_RobertTechnical Marketing[S] 185 points 16 hours ago*
SThe OP was looking only at the lowest draw call rates when asking about the 51%. The single batch GPU utilization is 51% (CPU-bound), medium is 71.9% utilization (less CPU-bound) and heavy batch utilization is 92.3% (not CPU-bound). All together for the entire test, there is 1.83X the performance of a single GPU
And where did I lie? Based on the info I saw in that thread that is what I gave you. You then being hostile claimed I was lying when in fact it was because I wasn't aware of the edit which you didn't mention there being one even after I posted the post where I got my info.Shut up. You weird ... I don't know what it is you're doing. I don't know why it is you're lying about something I can go check myself. it's beyond pathetic really