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

 

[GodisanAtheist]

Let's not hype too much..

-This guy

 

[gdansk]

Let's not hype too much..

Way too late for that. RDNA3 is bigger than Jesus/the Beatles.

 

[RnR_au]

Way too late for that. RDNA3 is bigger than Jesus/the Beatles.

Haven't see a single mention of a hype train for this release...



I do like this effort from 5 years ago. Lisa Su in a Ruby outfit makes it at least a little bit related to RDNA3

Already set my alarm for a 6am wakeup here on the Australian East coast. Only wish I had some AMD branded slippers...

 

[Leeea]

-This guy

This is the most under hyped launch I have seen a while.

Nobody is really expecting it to beat a 4090 enough to matter, to me the most interesting subject is around price.

 

[poke01]

Haven't see a single mention of a hype train for this release...

View attachment 70319

I do like this effort from 5 years ago. Lisa Su in a Ruby outfit makes it at least a little bit related to RDNA3

Already set my alarm for a 6am wakeup here on the Australian East coast. Only wish I had some AMD branded slippers...

Yeah same here. Aussie time(Melbourne). I am gonna watch on my 5300M powered laptop, the first RDNA arch.

 

[poke01]

This is the most under hyped launch I have seen a while.

Umm have you seen Twitter. Hype is off the charts, mainly the really hard-core AMD fanboys that the hype is beyond reality.

Like I say hype is great but in moderation. All that said hyped for AMD because their tech goes to consoles and handhelds too.
Plus AMD has been doing great last few years.

 

[SteveGrabowski]

Way too late for that. RDNA3 is bigger than Jesus/the Beatles.

Well Jesus was nailed to a cross at 33, while AMD still going strong at 53.

 

[Kepler_L2]

Umm have you seen Twitter. Hype is off the charts, mainly the really hard-core AMD fanboys that the hype is beyond reality.

Like I say hype is great but in moderation. All that said hyped for AMD because their tech goes to consoles and handhelds too.
Plus AMD has been doing great last few years.

Then you also have people like this...

 

[TESKATLIPOKA]

There will always be people who either underestimate or overestimate the performance. Nothing new there.

 

[Timorous]

Then you also have people like this...
View attachment 70324

That is probably a reasonable prediction for full N32.

 

[leoneazzurro]

Then you also have people like this...
View attachment 70324

Imagine believing that shader number and shader clock are two completely unrelated numbers. And imagine calling yourself a tech.

 

[coercitiv]

That is probably a reasonable prediction for full N32.

I don't think you get it, according to the way that youtuber is making performance estimates the full N32 would offer 90% of the 6900XT performance.

 

[leoneazzurro]

Navi 31 Die Shot - by SkyJuice - Angstronomics

"Performance, power and price will be revealed at the launch event. Unfortunately, our A0 performance profiling results were not indicative of production samples, so we have refrained from detailing our findings publicly. Silicon and board cost is definitely lower than NVIDIA’s RTX 4090, so let’s see where the two Navi 31 SKUs will be priced."

 

[poke01]

Laptops will be exciting with RDNA 3.

 

[Timorous]

I don't think you get it, according to the way that youtuber is making performance estimates the full N32 would offer 90% of the 6900XT performance.

I got it fully, just pointing out that 30/40% faster than 6900XT is probably where 7800XT will land give or take a bit of course.

 

[Timorous]

Navi 31 Die Shot - by SkyJuice - Angstronomics

"Performance, power and price will be revealed at the launch event. Unfortunately, our A0 performance profiling results were not indicative of production samples, so we have refrained from detailing our findings publicly. Silicon and board cost is definitely lower than NVIDIA’s RTX 4090, so let’s see where the two Navi 31 SKUs will be priced."

So that die looks exactly the same as what Igor posted in his CG mockup. I presume then his 3 8-pin are going to be AIB OC cards with the 2 8-pin being reference.

 

[moinmoin]

I got it fully, just pointing out that 30/40% faster than 6900XT is probably where 7800XT will land give or take a bit of course.

You got that the math is wrong, and are still pointing to the result of said math?

 

[Timorous]

You got that the math is wrong, and are still pointing to the result of said math?

The gain he calculated for the 7900XT/XTX vs the 6900XT is very likely wrong but the end number of 30/40% is probably ballpark for 7800XT. Is this really hard to get? He fluffed the numbers for the comparison he did but by accident those numbers probably apply to a different comparison.

 

[moinmoin]

Is this really hard to get? He fluffed the numbers for the comparison he did but by accident those numbers probably apply to a different comparison.

No, its just looks very silly.

 

[Karnak]

Anything else got that feeling that the MCD are somewhat... bigger than "excpected"?

 

[leoneazzurro]

Anything else got that feeling that the MCD are somewhat... bigger than "excpected"?

Not much, you have to consider that the expected MCD size was around 38 mm^2, so 6*38=228mm^2 and you can see that anyway these 6 MCD still don't cover the area of the GCD (supposedly 308 mm^2)

 

[Timorous]

No, its just looks very silly.

What is silly?

I am ignoring their method entirely and just saying that 30/40% increase over 6900XT looks about right for 7800XT / Full N32. I am basing that off of the laptop N32 ~= 6950XT at around 150W/160W info that has gone around. So 30/40% more at 220/230W roughly adds up. If that is silly then what is the point in a speculation thread because everything is silly.

 

[zinfamous]

You got that the math is wrong, and are still pointing to the result of said math?

No. he's saying that despite the stupid math, the tuber managed to stumble on a likely correct performance difference, but between 2 different cards that the tuber wasn't applying the stupid math to.

 

[RnR_au]

Hehehe

https://twitter.com/x/status/1588106223790002176

 

[Timorous]

Hehehe

https://twitter.com/x/status/1588106223790002176

No "We love gamers"?