Friday, March 6th 2020

AMD RDNA2 Graphics Architecture Detailed, Offers +50% Perf-per-Watt over RDNA

With its 7 nm RDNA architecture that debuted in July 2019, AMD achieved a nearly 50% gain in performance/Watt over the previous "Vega" architecture. At its 2020 Financial Analyst Day event, AMD made a big disclosure: that its upcoming RDNA2 architecture will offer a similar 50% performance/Watt jump over RDNA. The new RDNA2 graphics architecture is expected to leverage 7 nm+ (7 nm EUV), which offers up to 18% transistor-density increase over 7 nm DUV, among other process-level improvements. AMD could tap into this to increase price-performance by serving up more compute units at existing price-points, running at higher clock speeds.

AMD has two key design goals with RDNA2 that helps it close the feature-set gap with NVIDIA: real-time ray-tracing, and variable-rate shading, both of which have been standardized by Microsoft under DirectX 12 DXR and VRS APIs. AMD announced that RDNA2 will feature dedicated ray-tracing hardware on die. On the software side, the hardware will leverage industry-standard DXR 1.1 API. The company is supplying RDNA2 to next-generation game console manufacturers such as Sony and Microsoft, so it's highly likely that AMD's approach to standardized ray-tracing will have more takers than NVIDIA's RTX ecosystem that tops up DXR feature-sets with its own RTX feature-set.
AMD GPU Architecture Roadmap RDNA2 RDNA3 AMD RDNA2 Efficiency Roadmap AMD RDNA2 Performance per Watt AMD RDNA2 Raytracing
Variable-rate shading is another key feature that has been missing on AMD GPUs. The feature allows a graphics application to apply different rates of shading detail to different areas of the 3D scene being rendered, to conserve system resources. NVIDIA and Intel already implement VRS tier-1 standardized by Microsoft, and NVIDIA "Turing" goes a step further in supporting even VRS tier-2. AMD didn't detail its VRS tier support.

AMD hopes to deploy RDNA2 on everything from desktop discrete client graphics, to professional graphics for creators, to mobile (notebook/tablet) graphics, and lastly cloud graphics (for cloud-based gaming platforms such as Stadia). Its biggest takers, however, will be the next-generation Xbox and PlayStation game consoles, who will also shepherd game developers toward standardized ray-tracing and VRS implementations.

AMD also briefly touched upon the next-generation RDNA3 graphics architecture without revealing any features. All we know about RDNA3 for now, is that it will leverage a process node more advanced than 7 nm (likely 6 nm or 5 nm, AMD won't say); and that it will come out some time between 2021 and 2022. RDNA2 will extensively power AMD client graphics products over the next 5-6 calendar quarters, at least.
Add your own comment

242 Comments on AMD RDNA2 Graphics Architecture Detailed, Offers +50% Perf-per-Watt over RDNA

#1
medi01
@btarunr
1.5 not 2, the second slide.


btarunr
and NVIDIA "Turing" goes a step further in supporting even VRS tier-2. AMD didn't detail its VRS tier support.
AMD didn't detail it, yet we "know" NV "did step further".

How does this crap get into articles please?
Are people paid for sneaking BS like that in, or is it something happening unconsciously?
Posted on Reply
#2
R0H1T
I'm still struggling to see where it says 2x perf/W over RDNA in the slides or indeed any time in their conference?
Posted on Reply
#3
oxrufiioxo
medi01
@btarunr
1.5 not 2, the second slide.
I think people are confused because they called it Navi 2x even though the slide clearly shows 1.5.
Posted on Reply
#4
ratirt
oxrufiioxo
I think people are confused because they called it Navi 2x even though the slide clearly shows 1.5.
Maybe they meant 2x over GCN?
Posted on Reply
#5
oxrufiioxo
ratirt
Maybe they meant 2x over GCN?
Here's the slide I'm talking about

Posted on Reply
#6
ratirt
oxrufiioxo
Here's the slide I'm talking about


You are right. huh. This does not make much sense. Maybe the x1 or x2 and x3 doesn't represent performance uplifts but a moniker instead RDNA2 or 3 it is x2 or x3 as for generation of the chip?
Posted on Reply
#7
oxrufiioxo
ratirt
You are right. huh. This does not make much sense. Maybe the x1 or x2 and x3 doesn't represent performance uplifts but a moniker instead RDNA2 or 3 it is x2 or x3 as for generation of the chip?
Yeah, I figured this slide would confuse people even though I don't think that was the intention..... They clearly stated 50% more performance per watt in the live stream.
Posted on Reply
#8
btarunr
Editor & Senior Moderator
Fixed title, sorry for the confusion.
medi01
AMD didn't detail it, yet we "know" NV "did step further".
NVIDIA went a step further than Intel in supporting not just tier-1, but also tier-2. The complete sentence was:

"NVIDIA and Intel already implement VRS tier-1 standardized by Microsoft, and NVIDIA "Turing" goes a step further in supporting even VRS tier-2 ."
Posted on Reply
#9
medi01
Thread title is corrected now.

AMD didn't try to mislead anyone, as perf/w improvements are called out explictly:

Posted on Reply
#10
ratirt
oxrufiioxo
Yeah, I figured this slide would confuse people even though I don't think that was the intention..... They clearly stated 50% more performance per watt in the live stream.
Yeah. The x2 performance uplift for the RDNA2 is in comparison to GCN. This can be confusing for some people.

EDIT: Performance/Watt to be exact.
Posted on Reply
#11
medi01
ratirt
Yeah. The x2 performance uplift for the RDNA2 is in comparison to GCN. This can be confusing for some people.

EDIT: Performance/Watt to be exact.
Performance of what?
They have called perf/watt improvements to be 1.5 (it's 2.25 over CGN).
I have read they also hinted at 18TF $999 RDNA2 chip this year, which would be about twice of 5700XT perf.
Posted on Reply
#12
ratirt
medi01
Performance of what?
They have called perf/watt improvements to be 1.5 (it's 2.25 over CGN).
I have read they also hinted at 18TF $999 RDNA2 chip this year, which would be about twice of 5700XT perf.
I mentioned performance/watt. Of what? These are graphics chips so graphics?
I haven't read that article I'm just looking at the slides.
Posted on Reply
#13
lynx29
Going to be an interesting match up this year. It will be a combination of price, driver stability, and availability that wins my buy. I could give a crap less about RTX or Physx. So if Nvidia can only match the raw performance, but charge a $200-300 premium just for RTX, then I will roll AMD, assuming AMD has better drivers this time.

Time will tell.
Posted on Reply
#14
oxrufiioxo
medi01
They have called perf/watt improvements to be 1.5 (it's 2.25 over CGN).
I have read they also hinted at 18TF $999 RDNA2 chip this year, which would be about twice of 5700XT perf.
Considering the Xbox Series X is 12TF RDNA2 I would hope so.... Would be sorta odd for them not to have a discrete gpu that was about 50% better spec wise.
Posted on Reply
#15
medi01
ratirt
I haven't read that article I'm just looking at the slides.
I can't find confirmation of "$999 18TF chip" anywhere...
Posted on Reply
#16
R0H1T
ratirt
Yeah. The x2 performance uplift for the RDNA2 is in comparison to GCN. This can be confusing for some people.
It's still not 2x over GCN, 50% (perf/W) over GCN & then the same over RDNA. Looks like 125% efficiency over GCN but that's not saying much because I'm pretty sure that'd be the best case scenario.
Posted on Reply
#17
ratirt
medi01
I can't find confirmation of "$999 18TF chip" anywhere...
Honestly. If that is true 18TF chip, then good for everyone.

R0H1T
It's still not 2x over GCN, 50% over GCN & then the same over RDNA. Looks like 125% over GCN but that's not saying much because I'm pretty sure that's the best case scenario.
I think it is a moniker for the 2nd and 3rd gen aka x2, 3x and it has nothing to do with performance.
The other RDNA 50% more than GCN and then RDNA2 50% over RDNA then the RDNA2 is 2.25 times faster than GCN in my eyes.
Posted on Reply
#18
oxrufiioxo
R0H1T
It's still not 2x over GCN, 50% over GCN & then the same over RDNA. Looks like 125% over GCN but that's not saying much because I'm pretty sure that's the best case scenario.
Either way if they can get close to 50% per watt over a 5700XT while keeping prices sane it will be a nice card........... If its $800 it will be another fail. Well I guess that also depends on what Nvidia does they could raise prices again who knows.
Posted on Reply
#19
Chomiq
If they'll actually deliver the only thing that will prevent me from buying this would be the driver issues that we saw with almost every launch from team red's GPU camp.
Posted on Reply
#20
ratirt
Chomiq
If they'll actually deliver the only thing that will prevent me from buying this would be the driver issues that we saw with almost every launch from team red's GPU camp.
I think the driver will be ok. They had time with RDNA and 5000 series. Lets hope they can get these up to speed and stable.

oxrufiioxo
Either way if they can get close to 50% per watt over a 5700XT while keeping prices sane it will be a nice card........... If its $800 it will be another fail. Well I guess that also depends on what Nvidia does they could raise prices again who knows.
Yes but you have to keep in mind that the CU count would have to stay the same. They can still use more CUs and then maybe it would be even faster.
Posted on Reply
#21
delshay
I just want to see a Nano card this time around & HDMI 2.1.
Posted on Reply
#22
bogami
It is desirable to get better efficiency, but why do all AMD processors look like they were mushrooming on them?
Posted on Reply
#23
IceShroom
oxrufiioxo
Here's the slide I'm talking about


This slide says that RDNA based GPU has name that stats with - Navi1X, X=0,1,2....
RDNA2 based GPUs will have name like - Navi2X, where X=0,1,2.....
And RDNA3 based GPUs will have name like - Navi3X, where X=0,1,2.....
Posted on Reply
#24
W1zzard
oxrufiioxo
Here's the slide I'm talking about


Considering that currently shipping Navi is "Navi 10" and "Navi 14", which can be summarized as "Navi 1x", I would assume that the next GPUs are "Navi 20" and "Navi 30", so the x stands for "any number", not "x-times improvement"
Posted on Reply
#25
R0H1T
Oh in that case AMD should fire the wise guy who made that slide, I mean it makes little sense even now!
Posted on Reply
Add your own comment