Question Speculation: RDNA3 + CDNA2 Architectures Thread

[uzzi38]

Man I have been dying to make this one for a while now.

First rumours for RDNA3 are here so new thread time!

Just going to start off with this one for now: kopite7kimi on Twitter: "@VideoCardz Ah, I mean a simple mcm design with 10240 cores is not enough. Because the lift from RDNA2 to RDNA3 is much bigger than from RDNA1 to RDNA2. We should expect many big improvements in GFX11. 🤔" / Twitter

 

[Timorous]

While AMD would never do it (they try to roughly maintain perf/$ with the competition), imagine what would have happened if they had released for $699 for the 7900XT and $799 for the XTX.

They would sell out and scalpers would be making a nice profit.

 

[Gideon]

If the performance of the broken chip is indeed between 4080 and 4090 in raster, and around 3090 in RT, it is a good value proposition (looking at the market) for 999$ while having a BOM way below that.

Agreed. I think it will even be "passable" if it has about the same raster performance of the RTX 4080 (say 6800XT vs RTX 3080, within 5% of each-other) just nothing spectacular. In that case RT performance will equal 3090 only in lighter RT games (such as Doom Eternal, etc).

The 3Dmark leaks seem to indicate this is the plausible worst-case scenario. I obviously hope it will do somewhat better, but this is where I set my expectation bar currently, not to be disappointed again.

 

[maddie]

They would sell out and scalpers would be making a nice profit.

One has to ask oneself, why has this scalping of GPUs exploded recently and was a fringe phenomenon before. The unique circumstances are gone, but like the mesmerized miners, hope for that glorious past returning, dies hard.

 

[eek2121]

There will be at least one more Navi 31 card. But we don't know name

https://twitter.com/x/status/1602294895674523648

Interesting, if true.

One has to ask oneself, why has this scalping of GPUs exploded recently and was a fringe phenomenon before. The unique circumstances are gone, but like the mesmerized miners, hope for that glorious past returning, dies hard.

Insufficient supply. Overwhelming demand.

 

[Kaluan]

I haven't seen anywhere, but does anyone know the exact time of date (including time and timezone) that the review embargo lifts?

Pretty much all big review embargos I've seen in the last months (Raphael/AM5, RTL, RTX 4090 & 4080) have been lifted at 2 PM, GMT timezone.
I don't expect 7900XT(X) reviews to be any different.

Sooo... in a few miniutes 😋

 

[jpiniero]

https://twitter.com/x/status/1602294895674523648

Interesting, if true

A 1% bin for $1099?

 

[amenx]

Some fresh leaks. Hope not true as it shows xtx closer to 4080 than earlier thought.

https://www.youtube.com/@Benchmarklab/videos

 

[Abwx]

Computerbase review, published right at 15h00 locally...

AMD Radeon RX 7900 XTX & RX 7900 XT im Test

ComputerBase hat mit den AMD-Grafikkarten Radeon RX 7900 XTX und Radeon RX 7900 XT die Gegenspieler zur GeForce RTX 4080 im Test.

www.computerbase.de

 

[linkgoron]

AMD Radeon RX 7900 XTX & RX 7900 XT im Test

ComputerBase hat mit den AMD-Grafikkarten Radeon RX 7900 XTX und Radeon RX 7900 XT die Gegenspieler zur GeForce RTX 4080 im Test.

www.computerbase.de

Basically same performance as 4080 with Raster performance (although their "new" games benchmark perform significantly better). 3090TI RTX performance.

Very disappointing all in all, very far from what I expected after the AMD "launch".

 

[jpiniero]

Computerbase review, published right at 15h00 locally...

AMD Radeon RX 7900 XTX & RX 7900 XT im Test

ComputerBase hat mit den AMD-Grafikkarten Radeon RX 7900 XTX und Radeon RX 7900 XT die Gegenspieler zur GeForce RTX 4080 im Test.

www.computerbase.de

Seems like it's right at 4080 performance in raster but decently worse in RT.

 

[Unreal123]

Told you all. This is the first time that nvidia is ahead of AMD in hardware.

 

[maddie]

Told you all. This is the first time that nvidia is ahead of AMD in hardware.

????????????

 

[Karnak]

+18% Perf/Watt according to ComputerBase (only review so far for me, will look at others after).

But wth? What went wrong there?

 

[Unreal123]

+18% Perf/Watt according to ComputerBase (only review so far for me, will look at others after).

But wth? What went wrong there?

RTX 4080 is more power efficient than RX 7900 XT leave alone XTX.

 

[Karnak]

It's not about the 4080 or Nvidia. But +18% is not +54% what AMD told us. That's not even close where you could argue with "first party benchmarks". I'm disappointed tbh.

 

[Unreal123]

????????????

RTX 3090 Ti is faster than RX 7900 XT in some games and mostly trading blows.

 

[PJVol]

+18% Perf/Watt according to ComputerBase (only review so far for me, will look at others after).
But wth? What went wrong there?

+40% according to ComputerBase TechPowerUp (only review so far for me, will look at others after).
Nothing wrong.

 

[Glo.]

Either AMD lost performance per CU, with going into Chiplets, or RDNA3 is very, very buggy indeed.

 

[exquisitechar]

Well, it’s quite a bit worse than AMD made it seem, which usually doesn’t happen. Very disappointed in them for that, I won’t trust their claims anymore in the least.

 

[Heartbreaker]

It looks like AMDs own earlier performance claims were a bit overblown. They seemed to have overpromised a bit.

 

[Cali3350]

Seems like a pretty decent card still though. 4k performance is generally a tid bit above 4080 for $200 less. RT is worse, but that is not particularly important to me.

Its not a blowout at all, but I would say its a little better than I expected after the "broken" comments that last few days.

 

[Tup3x]

It looks like AMDs own earlier performance claims were a bit overblown. They seemed to have overpromised a bit.

Performance is a bit all over the place and there's also stability issues... Also relative to previous gen product the results are rather weird - AMD pretty much lied in their presentation.

 

[Timorous]

Performance is a bit all over the place and there's also stability issues... Also relative to previous gen product the results are rather weird - AMD pretty much lied in their presentation.

I get the impression the dual issue shaders are going to need some driver work. I can see AMDs claims being true for games where they have done the ILP extraction for their shaders but in other games where they have not the gain is down to the increase in shader count and slightly higher sustained clocks.

I guess this is why it was designed to be high clocking to make up for some of the games where the drivers are not extracting the most ILP yet or where they are simply tapped out on that front.

 

[Glo.]

Is it only me, or others also feel as if this release is rushed, and AMD released an unfinished product?

 

[moinmoin]

Is it only me, or others also feel as if this release is rushed, and AMD released an unfinished product?

If it's a broken release it has to be rushed as well since it can't be planned as a broken release, nor was it delayed to polish it. So yeah, looks like a rushed release, 25 months after RDNA2 which points to the rumors of internal targets not being reached rather close to the launch to be at least partly right. And can't blame Raja for this one.