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

#226
ARF
Valantar
It looks absolutely terrible, and changes the prefix from a series prefix (i.e. a name) to somehow being a count of something. Either it's counting the generation of the GPU (which the model number already does, i.e. it's redundant and only confusing - why is 11 (XI) the same as 6(xxx)?) or it's counting something else entirely, in which case the question becomes what?

RX is an equivalent of GTX for Nvidia, which is currently simply the prefix to what all their gaming-focused GPUs are named (there is also the entry-level, mobile-only MX series). RTX is then of course an extension of this - a gaming card with Ray tracing support. Or do you think Nvidia's X also stands for 10?
I think the X comes from eXtreme.

Previously, they had GT and GTS which were lower than GTX.

AMD also needs something to clarify that Ray-tracing support.

The "I" in RXI can come from Intersection.
Then AMD went from HD 8000 to R0 200.
Posted on Reply
#227
Valantar
ARF
Ok, now I see that R5 230 was a direct rebrand of HD 8450.

Then AMD went from HD 8000 to R0 200.
The HD 8000 series was pretty much OEM only. OEMs sadly tend to demand new product names each year regardless of whether there are new products available, which has led to a lot of silly midrange-to-low-end rebrands for both Nvidia and AMD across the years. Regardless, there have always been lower-tier R7 and R5 cards to the higher end, consumer-facing R9 cards.
ARF
I think the X comes from eXtreme.

Previously, they had GT and GTS which were lower than GTX.

AMD also needs something to clarify that Ray-tracing support.

The "I" in RXI can come from Intersection.
Intersection? Seriously? How is the average uninformed GPU buyer supposed to understand anything at all from that? At least R (for Nvidia) is the first letter in the actual feature it seeks to describe. "RXI" is a terrible idea. Period.

And you're right about X - in most product naming! - coming from extreme. That's likely why AMD fell back to it as well, as the letter X has become a sort of shorthand (much ridiculed, but still) for something cool/good/performant.
Posted on Reply
#228
ARF
Valantar
The HD 8000 series was pretty much OEM only. OEMs sadly tend to demand new product names each year regardless of whether there are new products available, which has led to a lot of silly midrange-to-low-end rebrands for both Nvidia and AMD across the years. Regardless, there have always been lower-tier R7 and R5 cards to the higher end, consumer-facing R9 cards.

Intersection? Seriously? How is the average uninformed GPU buyer supposed to understand anything at all from that? At least R is the first letter in the actual feature it seeks to describe. "RXI" is a terrible idea. Period.

And you're right about X - in most product naming! - coming from extreme. That's likely why AMD fell back to it as well, as the letter X has become a sort of shorthand (much ridiculed, but still) for something cool/good/performant.
The "I" can also invoke memories of Intel's "i" series CPUs and the customers may say "oh cool, this is I like iPhone and i7" ....." cool, man :D "
Posted on Reply
#229
Valantar
ARF
The "I" can also invoke memories of Intel's "i" series CPUs and the customers may say "oh cool, this is I like iPhone and i7" ....." cool, man :D "
Except that's a lowercase, single-letter prefix directly attached to a number. Does "RXI 6700" look like "i7-6700" to you? It sure doesn't to me. And besides, why would AMD try to sell products based on the naming of Intel CPUs when their own CPUs are kicking Intel's butt these days to such a degree that even people who don't care about PC hardware are picking up on it? Oh, and Apple got so much flack for their "it's pronounced iPhone ten, but we write it X" nonsense that anyone ought to understand that suddenly mixing in roman numerals into a series of non-roman numerals is a really bad and confusing idea. AFAIK most people still call it the iPhone "ex" (and even "ex eye" and especially "ex arr").


It's also worth mentioning that it's likely that part of why AMD abandoned the Rx naming was that they had long since seen that explicitly naming tiers like that has a detrimental effect on sales and marketing (you're very clearly telling your buyers that "this product is worse than something else", which isn't a good way of making people happy with their purchase), as Nvidia also saw and thus moved to all-over GTX branding. That's on top of it being redundant, of course. When your naming scheme consists of R[adeon][numbered tier] [space] [generation][numbered tier again?][0/5 if there's a new card/refresh] [X or no X depending if there's not enough room for a higher number] it doesn't take much brain power to tell that this scheme needs simplification. RX (named prefix, like "HD") [generation][three digits indicating performance level] [XT for higher end SKUs] is quite a lot simpler.
Posted on Reply
#230
ARF
Valantar
Except that's a lowercase, single-letter prefix directly attached to a number. Does "RXI 6700" look like "i7-6700" to you? It sure doesn't to me. And besides, why would AMD try to sell products based on the naming of Intel CPUs when their own CPUs are kicking Intel's butt these days to such a degree that even people who don't care about PC hardware are picking up on it? Oh, and Apple got so much flack for their "it's pronounced iPhone ten, but we write it X" nonsense that anyone ought to understand that suddenly mixing in roman numerals into a series of non-roman numerals is a really bad and confusing idea. AFAIK most people still call it the iPhone "ex" (and even "ex eye" and especially "ex arr").


It's also worth mentioning that it's likely that part of why AMD abandoned the Rx naming was that they had long since seen that explicitly naming tiers like that has a detrimental effect on sales and marketing (you're very clearly telling your buyers that "this product is worse than something else", which isn't a good way of making people happy with their purchase), as Nvidia also saw and thus moved to all-over GTX branding. That's on top of it being redundant, of course. When your naming scheme consists of R[adeon][numbered tier] [space] [generation][numbered tier again?][0/5 if there's a new card/refresh] [X or no X depending if there's not enough room for a higher number] it doesn't take much brain power to tell that this scheme needs simplification. RX (named prefix, like "HD") [generation][three digits indicating performance level] [XT for higher end SKUs] is quite a lot simpler.
I agree that it is simpler.
But AMD graphics division is not doing good and new image might become helpful.

For example, Navi 21 could be called Radeon iRT 900.
Navi 23 could be called Radeon iRT 700.
i just a cool letter, while RT from ray-tracing.
Posted on Reply
#231
Valantar
ARF
I agree that it is simpler.
But AMD graphics division is not doing good and new image might become helpful.

For example, Navi 21 could be called Radeon iRT 900.
Navi 23 could be called Radeon iRT 700.
i just a cool letter, while RT from ray-tracing.
But "i" still has zero relation to AMD's brand, and is already strongly related to both one of AMD's main competitors (Intel) and possibly the biggest brand on earth regardless of business (Apple). AMD adopting that would then just make them look like they're copying others to look cool, which will inevitably backfire. That's the type of marketing clueless rebranding shops do, not serious businesses, and the only reaction from the press if they did so would be to ask "...but why?". Of course RX is already very close to Nvidia's RTX, but at least in that case AMD had used the name for several years before Nvidia started using theirs, and Nvidia had a reasonable reason to switch.

As for replacing "RX" with "RT" ... why? That would suddenly make it AMD that's going after Nvidia's naming scheme rather than the other way around (which is a bad look, especially for an underdog), and they wouldn't gain much. While Nvidia's reason to switch from GTX to RTX was that they were adding RTRT and keeping both series alive, that is looking to be a single-generation thing - I don't think the GTX 16 series is getting a follow-up. For AMD to change their naming due to adding RTRT would then necessitate a wholesale name change for all RDNA 2-based cards, in which case "RT" or "iRT" would both be very poor choices (again, due to the resemblance to competitors' and other large brands' naming). Continuing with the established and relatively respected Radeon RX branding makes much more sense.
Posted on Reply
#232
ARF
Valantar
But "i" still has zero relation to AMD's brand, and is already strongly related to both one of AMD's main competitors (Intel) and possibly the biggest brand on earth regardless of business (Apple). AMD adopting that would then just make them look like they're copying others to look cool, which will inevitably backfire. That's the type of marketing clueless rebranding shops do, not serious businesses, and the only reaction from the press if they did so would be to ask "...but why?". Of course RX is already very close to Nvidia's RTX, but at least in that case AMD had used the name for several years before Nvidia started using theirs, and Nvidia had a reasonable reason to switch.

As for replacing "RX" with "RT" ... why? That would suddenly make it AMD that's going after Nvidia's naming scheme rather than the other way around (which is a bad look, especially for an underdog), and they wouldn't gain much. While Nvidia's reason to switch from GTX to RTX was that they were adding RTRT and keeping both series alive, that is looking to be a single-generation thing - I don't think the GTX 16 series is getting a follow-up. For AMD to change their naming due to adding RTRT would then necessitate a wholesale name change for all RDNA 2-based cards, in which case "RT" or "iRT" would both be very poor choices (again, due to the resemblance to competitors' and other large brands' naming). Continuing with the established and relatively respected Radeon RX branding makes much more sense.
Or maybe it's the best to allow the users to vote about how they want their cards to be christened?
Nvidia christened its graphics cards GeForce because the people said so.

As for AMD being backfired, they are backfired from the very start to begin with, have always been very bad in everything.
Look at one example - why is it reporting 6-bit colour when Radeon Settings is installed and 8-bit when it's uninstalled? :kookoo:

Posted on Reply
#233
Valantar
ARF
Or maybe it's the best to allow the users to vote about how they want their cards to be christened?
Nvidia christened its graphics cards GeForce because the people said so.

As for AMD being backfired, they are backfired from the very start to begin with, have always been very bad in everything.
Look at one example - why is it reporting 6-bit colour when Radeon Settings is installed and 8-bit when it's uninstalled? :kookoo:


Why hold a naming contest? Radeon is a well established and respected brand name. Nobody ever suggested "GT/GTS/GTX (and now RTX)" in a naming contest. Besides, that was in 1999.

Also, you would do well to look up the meaning word "backfire" and how it's used, as you can't say that someone/something "is backfired". That something backfires means that it has the opposite (or at least a very different) effect than what was intended. As for AMD's drivers being buggy, apparently YMMV there, as I have yet to have any serious issues across quite a few generations of AMD GPUs. I might be lucky, but serious bugs seem limited to an understandably vocal minority - but still a minority.
Posted on Reply
#236
EarthDog
ARF
AMD Big Navi Will Be 50% Faster Than RTX 2080 Ti According To Latest Leaks
If that is true... huuuuuuuuuuge. But not holding my breath.

Clicked play... heard it was Adored... stopped imediately.
Posted on Reply
#237
Master Tom
Dyatlov A
Will it support Windows 7?
What about Window 3.1?
Posted on Reply
#238
WeeRab
_larry
I'm just glad AMD is getting their $hit together GPU wise again finally. They have already done VERY well with their CPUs, now if they can get closer to what Nvidia delivers, it's gonna be another game changer. (Pun intended)

When the R9's came out I was stoked. I still have my R9 290 from 2013 and it still can handle most games at 1440p with some settings turned down. I was very disappointed with the Polaris architecture. All they did was make them more power efficient with the same performance as my 290. Hell, my 290 still beats the RX580 in some benchmarks... I am looking forward to getting a 5700XT when the new cards drop though :)
Oh I don't know. Their current mid-tier cards stack up well value wise against Nvidia. I recently bought an RX5700 for less than £300. Best bang-for-buck period.
Posted on Reply
#240
ECC_is_best
Using one of these can we assign a GPU per VM?
Posted on Reply
#241
Valantar
ECC_is_best
Using one of these can we assign a GPU per VM?
I guess we'll know when the cards arrive in... 6-8 months?
Posted on Reply
#242
ECC_is_best
Valantar
I guess we'll know when the cards arrive in... 6-8 months?
Well I figure their specs would say what it supports... But I'm not sure what the current AMD-branded technology is for this... DirectGMA?
Posted on Reply
Add your own comment