AMD's next-gen RDNA 2 'major leap forward' up to 225% faster than RDNA

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,898
"Big Navi / RDNA 2 launching in November key points
November 2020 launch: Firstly, AMD aiming for November means it puts the next-gen Radeon a couple of months behind NVIDIA is the rumors (and my sources) are true about an "August 2020 launch, September 2020 release" for the next-gen Ampere-based GeForce RTX 3000 series graphics cards.
How AMD benefits from a November launch: It will allow the company to have a couple of months to see where NVIDIA has priced its next-gen cards, and to gauge how performance is once it hits the hands of customers. AMD would then have a couple of months to be ready for a possible KO blow (at least for 2020) with Big Navi.
Thanksgiving launch: This is big for obvious reasons, but AMD being able to be inside of the next-gen Xbox Series X, the next-gen PlayStation 5, and have an NVIDIA Killer with Big Navi -- all for thanksgiving? What better gift can AMD bring to the table for Thanksgiving than that?!
Launching alongside PlayStation 5 and Xbox Series X: The stars have truly aligned for AMD to make this happen, and if performance of Big Navi is truly double that of RDNA1 and a 40-50% boost over the GeForce RTX 2080 Ti? Well, we could see AMD take out the end of 2020 with Intel and NVIDIA destroying products. Amazing.
Lisa Su did say Big Navi was coming in 'late 2020': You can read more about that here, but November is pretty damn spot on with "late 2020", right?!

Quick RDNA 2 / Big Navi specs
GPU cores: 72 SMs (two clusters of 36 CUs)
GPU game clock: 2.05GHz
GPU boost clock: 2.15GHz
Power usage: 300W or so
Performance: 40-50% faster than RTX 2080 Ti, 100% faster than RX 5700 XT"


Read more: https://www.tweaktown.com/news/7388...leap-forward-up-to-225-faster-than/index.html
 
That is exactly what people were expecting: you double the compute units of an RX 5700, then add some efficiency improvements, you're going to get double the performance of somewhere betwen the RX 5700 and the XT.

But we will see if they can actually scale to anywhere near that theoretical performance level (and if there's enough changes to the architecture over Navi 1.0 that they have to go through the same 12 months of driver mess?)
 
Last edited:
It'd be pretty exciting if AMD's high end is finally on par with NVidia. But also hard to believe.

Not a huge fan of the November delay. And not really buying the 'waiting to see what Nvidia prices their GPUs at'. I imagine they should have a ballpark idea where Nvidia is going to fall performance wise. As for pricing, they've shown in the past they have no problem adjusting their prices on the fly depending what Nvidia throws their way.
 
The article doesn’t say much... it throws around impressive numbers but with no relationship between them. So they are just big bold numbers that look and sound impressive. I’m not replacing anything until around April so both teams will have their stuff launched and should be available by then so I’ll get to wait and see.
 
Not going to hold my breath. Also, how's that RT performance going to be?

The latter is the real question. If Nvidia spends most of their new transistor budget on increasing ray tracing performance (and they need about 4x what the 20xx series has to match RT perf with raster perf) it wouldn't be too hard for AMD to pull ahead in raster performance even if they don't have any major architectural gains just by spending most of their transistor budget on that while only having a proof of concept level RT implementation (something similar to what the RTX 20xx cards offer).

It'd be the safer move for them, since we don't know how long it'll take for ray tracing to go from a forward looking feature to a must-have. If it's 2+ years out, shortchanging it now only hurts their resale value as they become obsolete faster. If mining seriously revives a shader heavy architecture would help their sales too; and historically that's been an AMD strength. As a gamer, it's one of the big reasons I want RT to take off; the more of the GPU is dedicated to things that can't mine the more likely Etherium and other coins that currently are GPU friendly will go off to custom ASICs. But from the perspective of an AMD bean counter, I can see the interest in staying the mining friendly product as long as possible.
 
Not going to hold my breath. Also, how's that RT performance going to be?

Edit: This is from MLiD. Not buying it for a second.

it's about as bad as Gamer Meld on YT too, i didn't like posting this. trust me
 
Oh yes more AMD hype that never delivers. How long did 5700 series take to get viable drivers, 6 months after release?? I wonder what Navi 2 drivers will be like...

I’m taking a wait and see approach this time. If Ampere gives me 40+% more raw FPS across the board I’ll likely grab it. If not I have no problem hanging on to my 2080 Ti for another year.
 
I hope the performance claims are true. However, if they knew it was going to be 50% faster than a 2080ti wouldnt they try to beat Nvidia to the punch(they are the enthusiast's darling right now) rather than engage in a price competition after the fact? That seems to say they dont know where this card will fall for sure in terms of performance.
 
I hope the performance claims are true. However, if they knew it was going to be 50% faster than a 2080ti wouldnt they try to beat Nvidia to the punch(they are the enthusiast's darling right now) rather than engage in a price competition after the fact? That seems to say they dont know where this card will fall for sure in terms of performance.
There could be a business reason, or maybe they just can't get the new cards ready any sooner than that. We'll likely never know for sure.
 
Oh it's about time to drink up those red Kool aid colored tears again. I remember hoping that they would be competitive and saying they aren't here to save us so stop believing those $250 1080ti performance dreams people are selling. Oh the horror when the product was out it was slower and more money to the point they had to drop their price to match them up to the proper Nvidia products. Someone is stiring the Kool aid again. The longer we go without AMD saying anything about performance the smaller they are in my mind share. I think they are capable of making products that are great value at the low end but have lost their way on high-end crown fighting technology. Ultimately I don't think the GPU division has enough R&D budget to fight with Nvidia yet.
 
Honestly based of how powerful the APUs are in the new console it could actually be pretty powerful in a discrete GPU. I am not holding my breath but damn it it is true.
 
  • Like
Reactions: Axman
like this
based of how powerful the apus are in the new console

Honestly looking back, I think Zen was good luck and the real hard work went into developing console APUs. The fact that they were able to spin that off into a stellar CPU architecture and capitalize on that in the near-term was happenstance.

Not that they didn't take Zen development seriously, just that I think it was a byproduct of their relationship with Microsoft and Sony.
 
Put up with drivers to get it on the cheap after a few months why not.
 
Honestly looking back, I think Zen was good luck and the real hard work went into developing console APUs. The fact that they were able to spin that off into a stellar CPU architecture and capitalize on that in the near-term was happenstance.

Not that they didn't take Zen development seriously, just that I think it was a byproduct of their relationship with Microsoft and Sony.
I don't think I've heard or even thought about this stance before but I find it interesting and funny if that's what happened.
 
Well, we will have to see what happens. One thing for sure, we are looking toward exciting times in the computing industry. :)
 
I would not shake your head to hard because, after all, Zen was far better than expected when it was released. Just have to wait and see what we get.

Ryzen was a total departure from Bulldozer. And it wasn't far better than expected on release. AMD made claims that it would be 40% faster than Bulldozer, which led to the conclusion that it would be around the same performance as Ivy Bridge. When it launched, it was a little better than that, but still well behind Intel. However, the increased core counts and excellent pricing strategy made the CPU's very popular. Don't forget that people were longing for an AMD product worth buying. Ryzen has then had iterative changes that have allowed AMD to leapfrog Intel in most respects. But the original Ryzen was only a little better than what AMD suggested. This was more of a case of AMD adopting the "under promise and over deliver" strategy rather than making claims about performance which ultimately turned out to be marketing BS. Prior to Lisa Su taking over, that's how AMD operated.

It seemed that AMD hoped that hype would drive enough initial sales for the CPU's to succeed despite being underwhelming upon release. In the case of Ryzen, AMD kept its claims realistic, and even understated. That way there was little chance for disappointment.

Now, the graphics division isn't the same beast as the CPU division is. So far, AMD has over promised and way under delivered. It's obvious that RDNA didn't scale well due to its inefficiencies. We can see that in how giving it more power improves its performance greatly, and that scaling it up (and increasing those requirements further) would be virtually impossible without massive changes and improvements in efficiency. I think people are right to be skeptical about what AMD can do with RDNA 2 given that its likely an iteration of RDNA and its biggest weakness is the very reason it wasn't competitive. It seems unlikely that AMD is going to be able to overcome the architecture's inefficiencies enough to take on Ampere.
 
I'm curious, why do you guys think he's so untrustworthy?

He regurgitates existing rumors and never adds anything new except random speculation. Claims to have sources but never has any real info. Basically zero track record or credibility but makes a lot of noise. Just YouTube clickbait at this point.
 
I hope the performance claims are true. However, if they knew it was going to be 50% faster than a 2080ti wouldnt they try to beat Nvidia to the punch(they are the enthusiast's darling right now) rather than engage in a price competition after the fact? That seems to say they dont know where this card will fall for sure in terms of performance.
I always hear this argument. You don't beat NVIDIA to the punch and hurry up a product. No thanks. They already haven't competed at high end for a while what is another month or 2 going to do after nvidia. At this point it makes better sense to have NVIDIA release their product first and then price your product accordingly to counter.
 
Disinformation campaign? Raja is gone so overpromising and underdelivering does not happen on its own.

Ohhh snap! Ol' Bennett is throwing out the bombs.

Maybe the November launch date will give their driver team a bit longer to cook the launch day drivers. The VII and 5700 XT launch drivers were a little suspect.
 
Maybe the November launch date will give their driver team a bit longer to cook the launch day drivers. The VII and 5700 XT launch drivers were a little suspect.
I've been a proponent of AMD drivers in the past, but I got bit by them this weekend on my 5700 XT.

Got a new monitor, swapped it out, and was stuck at 60Hz. Thought the monitor was broken at first (165Hz capable) but a AMD driver update fix it.

Then I was getting crashing after about 5 or 10 minutes in several games. Had to do a DDU safe mode uninstall and reinstall (of the same newest driver).

Now things are working fine, but it wasn't the best experience.
 
Overall I'm waiting for RDNA2 and Ampere to decide on my next GPU, but I'm definitely hoping that it will be AMD - I find the company and approach tends to be more open, ethical, and the like. It will all depend however on how well the card performs and its price, though. That said, if AMD wants to extricate themselves from these driver issues (admittedly I've not run into them myself ), I'd love to see them totally revamp the Windows driver the way they did so with Linux.

AMD drivers on Linux are an absolute dream, built from a single open source base with proprietary extensions something of an "add-on" if you wish it. Likewise, things like AMD's lack of OpenGL performance (those who are heavily involved in recent console emulation will no doubt run into this) on Windows are non-issues on the Linux side because of the changes. There was a time when AMD/ATI drivers on Linux were ghastly affairs at nearly every level but AMD put a lot of effort into dealing with the problem dumped into their lap and now they're better than ever while also more open and performant than ever. It would be nice to see the same thing happen to those Windows based and new GPUs are the perfect place to do so.
 
Back
Top