Question Speculation: RDNA2 + CDNA Architectures thread

[uzzi38]

All die sizes are within 5mm^2. The poster here has been right on some things in the past afaik, and to his credit was the first to saying 505mm^2 for Navi21, which other people have backed up. Even still though, take the following with a pich of salt.

Navi21 - 505mm^2

Navi22 - 340mm^2

Navi23 - 240mm^2

Source is the following post: https://www.ptt.cc/bbs/PC_Shopping/M.1588075782.A.C1E.html

 

[Kenmitch]

Its going to be crap, from top to bottom.

Your overdoing it. Find somewhere in the middle ground. The true bi-polar folks will get angry at you.

Red dragon slay blue dragon already, tired red dragon may slay the green dragon. Green dragon playing lame duck, red dragon resting getting ready to pounce! Red dragon has mighty battle ahead, if wins becomes mighty red dragon!

 

[amrnuke]

While this new Glo is entertaining, i prefer the old. I have a spare ticket for the Choo Choo, it stops for nothing and no one

Good: Navi21 will beat 3090
Bad: Navi21 will cost $100 more than 3090
Ugly: Navi21 will draw 450W at peak gaming, and contrary to pre-release photos, will only come as a reference three-slot blower, no AIBs

 

[pandemonium]

A lot of nay-sayers recently seem to embrace AMD failing in the past and accept it as the only course of action to happen from here on out. While I understand that logic, I fail to reason with it given a lot of facts:

1. Two major consoles have their hardware based on AMD's architecture and we're pretty well acquainted with how they can perform
2. The road map for GCN to RDNA to RDNA2
3. The compute task handling theoretical research paper

Further, I'm wondering why would one want AMD to fail and not give Nvidia a run for their money? Good competition = better prices and better results for us, the end users.

Think about that the next time you want to bring up AMD's previous history of failure. Also, read the thread, look at the architecture, and understand the origins of the speculations here before commenting.

I have nothing to contribute to this thread aside from this summary. Also, I hope AMD wipes the floor with Nvidia because I'm tired of their market controlling shenanigans.

 

[Veradun]

Barely matching RTX 3070.

That's what the 3060Ti is for, right?

 

[DisEnchantment]

I just wished DF did a deeper look into that, because it doesn't match their earlier narrative that Series X = 2080.

https://twitter.com/x/status/1317513390756122625

DF seems to think PS5 is 2060 level.
Games shown by Sony indicate otherwise.
Anyway we will get the parts for PC and other reviewers can make the compare.
Also, there is too many mention of NV in so many console related things which have absolutely nothing to do NV which makes me wonder if something is up.

 

[Veradun]

Anyway, there is too many mention of NV in so many console related things which are absolutely got nothing to do NV which makes me wonder if something is up.

It's called damage control

 

[dzoni2k2]

https://twitter.com/x/status/1317513390756122625

DF seems to think PS5 is 2060 level.
Games shown by Sony indicate otherwise.
Anyway we will get the parts for PC and other reviewers can make the compare.
Also, there is too many mention of NV in so many console related things which have absolutely nothing to do NV which makes me wonder if something is up.

Same incompetent DF that benchmarked Nvidia and AMD GPUs in I think it was Mirrors Edge Catalyst? And forgot to mention that game dynamically adjusts settings on lower VRAM GPUs? Which meant Nvidia GPU ran at lower settings.

 

[Vope45]

A lot of nay-sayers recently seem to embrace AMD failing in the past and accept it as the only course of action to happen from here on out. While I understand that logic, I fail to reason with it given a lot of facts:

1. Two major consoles have their hardware based on AMD's architecture and we're pretty well acquainted with how they can perform
2. The road map for GCN to RDNA to RDNA2
3. The compute task handling theoretical research paper

Further, I'm wondering why would one want AMD to fail and not give Nvidia a run for their money? Good competition = better prices and better results for us, the end users.

Think about that the next time you want to bring up AMD's previous history of failure. Also, read the thread, look at the architecture, and understand the origins of the speculations here before commenting.

I have nothing to contribute to this thread aside from this summary. Also, I hope AMD wipes the floor with Nvidia because I'm tired of their market controlling shenanigans.

what the heck are you talking about ? 2/3 of posters are on the choo choo train while the rest, including me, saying better keep expectation in check. How can you infer several pages of non content to your conclusion ?

 

[Veradun]

what the heck are you talking about ? 2/3 of posters are on the choo choo train while the rest, including me, saying better keep expectation in check. How can you infer several pages of non content to your conclusion ?

It's so boring to be on an hype train, it's still so quiet and comfortable here at destination :>

 

[Vope45]

Same incompetent DF that compared Nvidia and AMD GPUs in I think it was Mirrors Edge Catalyst? And forgot to mention that game dynamically adjusts settings on lower VRAM GPUs? Which meant Nvidia GPU ran at lower settings.

Techtubers and tech journalists make me want to puke. Especially Jayz2cents. I still remember him begging nvidia for free cards and throwing a tantrum over AMD products.

 

[Vope45]

It's so boring to be on an hype train, it's still so quiet and comfortable here at destination :>

I'll hop back on the hype train once AMD has something special to counter nVidia and i'm not talking about 5% to 10% slower at lower price and lower power consumption. 255W TGP is basically 300W card. For 20W more you can get a 3080 with DLSS, better ray tracing performance and a load of nvidia gimmicks but only if an average user is willing to wait.

 

[Kuiva maa]

https://twitter.com/x/status/1317513390756122625

DF seems to think PS5 is 2060 level.
Games shown by Sony indicate otherwise.
Anyway we will get the parts for PC and other reviewers can make the compare.
Also, there is too many mention of NV in so many console related things which have absolutely nothing to do NV which makes me wonder if something is up.

I wouldn't bother too much with DF estimations at this point. With PS4/Xbone they initially claimed that a 750Ti would roughly mirror the settings and performance of consoles on the PC. It worked like that for a couple of years but in late 2015, when the (arguably) first true next gen title arrived (SW:BF) with its photogrammetry and compute shaders etc, the 750Ti couldn't keep up with PS4. By mid 2016, whem DOOM came around, the 750Ti was getting trounced so they stopped using it as a PC point of reference. If they lowball again with this 2060S estimate for PS5, I expect them to backtrack even sooner this time.

 

[Timorous]

- This has likely been covered somewhere in the 130+ pages of this thread, but I'm curious as well of the logic of having a 80CU chip, then 40/32/24 CU chips.

I understand the huge gap between the 80 and 40, plenty of die harvesting possible to fill that gap, but I would have figured we'd get a 40 CU Chip and a 20CU chip mirroring RDNA 1 (naturally as chips get smaller, you have to artificially segment rather than naturally harvesting dies).

Its been a helluva thread, can anyone refresh the argument for this?

My problem with 40CU N22 is that it is rumoured to be over 300mm² and with the Series X we know AMD could design a 56CU part with a 320bit bus and 64 ROPs in a similar die area.

In fact this is my problem with all the rumours. The die size vs performance vs CU count make little sense. For a 300mm² N22 to make sense it needs to be at 2080Ti/3070 performance, any less and they may as well have made desktop versions of the Series X GPU because 56CUs @ 2.2 Ghz would be 2080Ti level.

If N22 is closer to 350mm² then it needs to perform even better. This would be between 3070 and 3080 performance so if AiBs only have N22 it does line up with the leaked performance expectations but that implies the teased big navi part was the bottom part.

So yea, the rumours make little sense full stop.

 

[majord]

I'm confused. 40CU RDNA1 @ 1.8Ghz already = a smidge shy of 2070 super

They think 40CU RDNA2 @ up to 2.2Ghz is slower? I mean if it turns out "up to" is, like "up to" in the menu, 1.5ghz in gaming or something, sure. That would be pretty weak.

 

[uzzi38]

I'm confused. 40CU RDNA1 @ 1.8Ghz already = a smidge shy of 2070 super

They think 40CU RDNA2 @ up to 2.2Ghz is slower? I mean if it turns out "up to" is, like "up to" in the menu, 1.5ghz in gaming or something, sure. That would be pretty weak.

Because they're just looking at how the 5700 performs because it and the PS5 both are 36CUs.

 

[Timorous]

I just wished DF did a deeper look into that, because it doesn't match their earlier narrative that Series X = 2080.

That was in a quick and dirty port just to show off the new features. I bet if they did optimise it they stopped as soon as performance was acceptable to show the features they wanted. That was always a lower bound for performance.

 

[majord]

Because they're just looking at how the 5700 performs because it and the PS5 both are 36CUs.

argh, sorry I keep thinking it (PS5) is 40CU not 36.

Makes more sense.. but is still a very pessimistic

 

[dzoni2k2]

My problem with 40CU N22 is that it is rumoured to be over 300mm² and with the Series X we know AMD could design a 56CU part with a 320bit bus and 64 ROPs in a similar die area.

In fact this is my problem with all the rumours. The die size vs performance vs CU count make little sense. For a 300mm² N22 to make sense it needs to be at 2080Ti/3070 performance, any less and they may as well have made desktop versions of the Series X GPU because 56CUs @ 2.2 Ghz would be 2080Ti level.

If N22 is closer to 350mm² then it needs to perform even better. This would be between 3070 and 3080 performance so if AiBs only have N22 it does line up with the leaked performance expectations but that implies the teased big navi part was the bottom part.

So yea, the rumours make little sense full stop.

Die sizes are probably the least reliable leaks we have. They could all be BS. Well with exception of N23 which is suppose to be 236mm2.

 

[Timorous]

Because they're just looking at how the 5700 performs because it and the PS5 both are 36CUs.

That is dumb but DF like to hedge, they just overdo it sometimes.

PS5 should be at or slightly ahead of 5700XT (assuming it holds 2.23Ghz as much as Cerny claimed) which is 2070S level.

 

[Vope45]

I wouldn't bother too much with DF estimations at this point. With PS4/Xbone they initially claimed that a 750Ti would roughly mirror the settings and performance of consoles on the PC. It worked like that for a couple of years but in late 2015, when the (arguably) first true next gen title arrived (SW:BF) with its photogrammetry and compute shaders etc, the 750Ti couldn't keep up with PS4. By mid 2016, whem DOOM came around, the 750Ti was getting trounced so they stopped using it as a PC point of reference. If they lowball again with this 2060S estimate for PS5, I expect them to backtrack even sooner this time.

But that is how the game is played isn't it ?

 

[Timorous]

Die sizes are probably the least reliable leaks we have. They could all be BS. Well with exception of N23 which is suppose to be 236mm2.

If N23 is 236mm² then it needs to be at or slightly above 5700XT.

I really don't get the leaks. Take the series X GPU at around 300mm². Build it using the same library as renoir and it shrinks down to 205mm². That would explain the smaller bus because you would not fit a 320 bit bus on a die that small and 256 might be a challenge so you make it 192bit and those transistors you save on the bus you use for Infinity Cache, and maybe then some so you grow it back to 230mm² ish like the N23 leak but now you have a 2080Ti/3070 tier part in a 230mm² die area. Give it 12GB ram and you are good to go.

An 80CU, 56CU, 40CU, 24CU stack makes sense and it would make Big Navi not that big (350mm²-400mm²) and explain why they have a smaller bus.

I guess I will have to wait till the 28th but I keep hoping something will leak that makes sense of this nonsense.

 

[moinmoin]

https://twitter.com/x/status/1317513390756122625

DF seems to think PS5 is 2060 level.
Games shown by Sony indicate otherwise.
Anyway we will get the parts for PC and other reviewers can make the compare.
Also, there is too many mention of NV in so many console related things which have absolutely nothing to do NV which makes me wonder if something is up.

DF consists of multiple persons, Rich, Dark1x, Dictator...
Dictator on messages boards like Neogaf, ResetEra etc. is a very(!) staunch Nvidia fan, so I personally wouldn't take any classification of RTG products from him without any less biased second opinions. In general Nvidia's following on PC is irrationally strong, that's the uphill battle AMD has to fight in the DIY GPU market.

 

[PhoBoChai]

I'll hop back on the hype train once AMD has something special to counter nVidia and i'm not talking about 5% to 10% slower at lower price and lower power consumption. 255W TGP is basically 300W card. For 20W more you can get a 3080 with DLSS, better ray tracing performance and a load of nvidia gimmicks but only if an average user is willing to wait.

You sound like you've already made up your mind, especially with your repeated claims of 3080 having better ray tracing performance.

I would just like to refresh your knowledge on the current situation.

We have the 3080, in Metro Exodus, 4K native with ONE ray tracing effect, global illumination and FPS tanks badly. We also have the same GPU in Fortnite, and using multiple RT effects, FPS drops to 23. Yes, 23.

Then we have the PS5, in Spider Man, with multiple RT effects, 4K30. DMC5, multiple RT effects, 4K30.

Don't be too quick to dismiss RDNA ray tracing potential.

 

[Head1985]

I think this is -redacted-.No way big navi eat more than 3x more power than consoles.I think igor lost it.Well 8 more days.

AMD Radeon RX 6000 based on Navi 21 XT to feature 320W TBP, 16 Gbps GDDR6 memory - VideoCardz.com

A fresh report from Igor’sLAB puts a new claim on Navi 21 power consumption. AMD Navi 21 XT to feature 320W TGP/TBP Igor Wallossek published his first ‘big leak’ on Big Navi. According to his data, AMD Navi 21 XT will feature TGP at around 320W, not 255W as reported earlier. The difference comes...

videocardz.com

Btw TGP is for entire card.Dont know what the redacted igor doing?

Graphics Cards: TDP and TGP (and don't forget TBP, GCP and MPC...) | Geeks3D

Graphics Cards: TDP and TGP (and don't forget TBP, GCP and MPC...)

www.geeks3d.com

As you already know, Profanity is not allowed in the tech forums.

Daveybrat
AT Moderator

 

[PhoBoChai]

I think this is bull. No way big navi eat more than 3x more power than consoles.I think igor lost it.Well 8 more days.

AMD doesn't use TGP. As we discussed this already. That's an NVIDIA terminology.

For AMD, its package power (die) and TBP = Total BOARD Power. Its been this way for ages, then a few years ago, we started getting all these random different terms and different power allocations.

As for 320W total power, I don't see why its not possible. They can crank it well beyond the efficiency curve to chase more performance at the expense of perf/w just like NVIDIA did. It seems the market is A-OK with high power GPUs now with Fermi 2.0.