AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

queamin · 9 Oct 2020 at 00:21

LePhuronn said:
um no, that's what the 28th is for. Surely the concept of 2 separate events isn't that big a challenge to comprehend. Plus everybody has been crying about zero Big Navi info, demanding AMD show everything immediately, yet when we get official ******* confirmation that AMD has some form of Big Navi that draws parity with the 3080 it's still not enough? There's just no pleasing some people.

The reviewers were moaning last time they launched them together, not enough time for the reviews and the Cpu revel over shadowed Gpu revel.

LePhuronn · 9 Oct 2020 at 00:37

g67575 said:
Why on earth would they show the performance of their lower end GPU first, thats just stupid.

I dunno, for the same reasons they showed their lower end CPU first? It's called build-up and expectation and then leaving with a little something bigger and special at the end. if AMD teased their biggest Navi there would be zero reason to tune in on the 28th. It's ridiculous to me that people genuinely don't understand how this works.

FoxEye · 9 Oct 2020 at 00:37

OMGsideboob said:
Very interesting reveal at the end there which answers some questions and confirms a few leaks. The lower tier card they showed off is roughly on par with the 3080 while consuming much less power - we knew one of their lower tier cards would compete with the 3070 but to have it competing at the high end is absolutely huge. They're further ahead than most of us imagined they would be.

This then puts their top tier card decisively above the 3080 in performance while using much less power. The only question I'm left with now is how close the top end RDNA 2 card will be to the 3090. Considering how closely they are guarding RDNA 2 information and their approach to marketing this time around, signs firmly suggest they'll claim both CPU and GPU performance crowns very soon.

Eh? Where are you getting all this from?

LePhuronn · 9 Oct 2020 at 00:41

Grim5 said:
You guys forgot about Moore's Law is Dead, that troll once again has been proven wrong

shut up Grim, you're hardly in any position to call other trolls.

And I'd like to know exactly what information has been revealed that's proven anything. I'd also like to know exactly what MLID has gotten wrong so far.

Zeeflyboy · 9 Oct 2020 at 00:47

LePhuronn said:
shut up Grim, you're hardly in any position to call other trolls.

And I'd like to know exactly what information has been revealed that's proven anything. I'd also like to know exactly what MLID has gotten wrong so far.

Oh you know. This and that... oh and the time he said that thing. Oh boy did he get that wrong!

Lokken86 · 9 Oct 2020 at 00:48

Didn't he just say it would be close to a 3080? And it looks like it is so....

EastCoastHandle · 9 Oct 2020 at 00:57

Proof that with new games Radeon has improved

5700xt: 40 CUs, 64 ROPs, 160TMUs, 2560 Shading Units, 10.3B Transistors @ 225watts with a 251 mm² die size
2070s : 40 SMs, 64 ROPs, 160TMUs, 2560 Shading Units, 13.6B Transistors @ 215watts with a 545 mm² die size

Newest PC title is Star Wars Squadrons. As you can see the 5700xt yet again beats the 2070s in these new titles. Lets remember 5700xt was barely keeping up with a 2070 at one point.
So if a 40 CU 5700xt can compete with a newly released 2070s (a post release sku to compete against Navi) then an 80CU sku should be an exceptional part. We will see.

DDH · 9 Oct 2020 at 00:58

g67575 said:
Im happy with the info that was shown today. The only people who aren't are the ones claiming (for no apparent reason), that there is still more to come this year for RDNA 2 graphics cards.

No apparent reason? What have we gotten this year from rdna2? Some benches from AMDs zen presentation, and some info about XBSX soc.

Hmmm what else could they possibly tell us about rdna2 this year? What about..... Names, SKUs, performance, clocks, memory capacity, availability, price... Basically everything.

Are you intentionally this daft or is it accidental?

And the reason we believe this isn't the top die on the benches is simple. AMD have been so cagey about rdna2 even their aibs don't know what the top card is. And all of a sudden they release benchmarks 3 weeks prior to its reveal? Not likely. Lisa Su said it was big Navi, she didn't say it was the biggest Navi and herkleman wants that point to be known.

Rroff · 9 Oct 2020 at 01:26

KentMan said:
Careful @Rroff will be along shortly to accuse you of turning over stones to mudfling at Nvidia

You are right to point it out his post is ridiculous.

EastCoastHandle · 9 Oct 2020 at 02:03

Herkelman stressed that AMD didn’t state which Radeon RX 6000 graphics card ran these benchmarks. We don’t know whether these results come from the biggest Big Navi GPU, or a more modest offering. (Herkelman also said there’s still fine-tuning left to do before launch.)

https://www.pcworld.com/article/3585090/amd-radeon-rx-6000-big-navi-performance-tease-rtx-3080.html

g67575 · 9 Oct 2020 at 02:04

DDH, if we can just agree that there's no way you will admit the GPU that AMD have shown us is the highest end GPU on offer, that would save a lot of time.

Your attitude is nothing new.

If it's true there's still some fine tuning to do with the top RDNA 2 GPU, it suggests (unsurprisingly) that it's been a bit of a rushed launch.

Rroff · 9 Oct 2020 at 02:22

g67575 said:
If it's true there's still some fine tuning to do with the top RDNA 2 GPU, it suggests (unsurprisingly) that it's been a bit of a rushed launch.

With a teaser ahead of the final touches to launch there is always going to be a caveat of last minute fine tuning, etc. I don't think that necessarily means a rushed launch it happens with pretty much any product launch.

I think people have lost sight of the 5700XT with some of the comments above - it is already a 2nd gen 7nm part which they've basically doubled the performance of and taking in mind that they will need a little extra silicon at the same performance level to accommodate new features like the RT changes to shaders that will mean a die about as big as it makes commercial sense on TSMC 7nm for a gaming GPU was showed off here aka big Navi (this is taking into account process gains and refinements to switching power, etc.) - if they have a higher product it is around the order of 10-15% more performance tops but I see no reason from the tone AMD brought that they'd have held that back if they had it.

g67575 · 9 Oct 2020 at 02:24

Fair enough, Rroff.

Anyone know what RX-532 means (written at the bottom of the slide), if anything?

https://hexus.net/media/uploaded/2020/10/e7631a23-2e9d-43ed-b6b6-1cdb390bbc51.PNG

Maybe AMD just wants more speculation before the launch

EDIT - Found this on a spanish forum, which is part of the endnotes:

https://cdn.videocardz.com/1/2020/10/AMD-Ryzen-5900X-Radeon-6000.jpg

DDH · 9 Oct 2020 at 02:35

g67575 said:
DDH, if we can just agree that there's no way you will admit the GPU that AMD have shown us is the highest end GPU on offer, that would save a lot of time.

Your attitude is nothing new.

If it's true there's still some fine tuning to do with the top RDNA 2 GPU, it suggests (unsurprisingly) that it's been a bit of a rushed launch.

Sure we can agree on that, if we also will agree that you wont admit you have no idea what gpu was shown and it could have been anything

Ive no doubt my attitude it not new to you, i bet youve encountered people pushing back on your dubious opinion present as facts many times before

g67575 · 9 Oct 2020 at 02:43

Sure, it's possible. I'd just rather go by the information we already have.

Blackjack Davy · 9 Oct 2020 at 02:45

Lokken86 said:
Didn't he just say it would be close to a 3080? And it looks like it is so....

Not heard any of his his earlier stuff but his recent predictions for Big Navi were "within striking distance of the 3080" and far exceeding the 3070. Also the 3080FE cards are essentially loss leaders and will always be in very limited supply as they exist simply to keep the official MSRP low, oh and the prices for AIB's would rise come october. All of which has been spot on so far.

SpudMaster · 9 Oct 2020 at 02:49

Even if AMD's top card does turn out to be 10% slower than a 3080 it will no doubt be cheaper. If it does infact have 16gb vram as all the rumours have suggested then I know which card I would rather have for 4k gaming when we start breaking the 10gb barrier

CAT-THE-FIFTH · 9 Oct 2020 at 02:52

SpudMaster said:
Even if AMD's top card does turn out to be 10% slower than a 3080 it will no doubt be cheaper. If it does infact have 16gb vram as all the rumours have suggested then I know which card I would rather have for 4k gaming when we start breaking the 10gb barrier

Going by the Zen3 pricing I don't expect AMD to really undercut Nvidia that much. They might just price it the same as the RTX3080,but have more VRAM.

g67575 · 9 Oct 2020 at 02:57

I think it will be priced about the same as the RTX 3070. It's no coincendance Nvidia is launching it the day after the RX 6000 series launch date.

The 5700 XT launched at 400 dollars, so hopefully it won't be a massive increase in cost. Maybe they could sell an identical version of the GPU, but with less VRAM.

Emlyn_Dewar · 9 Oct 2020 at 03:00

Is there any other enthusiast area in the world, that can talk so much about so little?
I'm hopeful from those 3 numbers AMD posted, and they do seem to be on a roll. Still, we've all been there for the last several AMD card releases. It's not set the world on fire.
It feels like some are offended by the idea AMD could release a decent card though. They should give their head a shake. Some competition would be great!
I don't care who makes my GPU, I just want a strong mid-range card.