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

 

[AtenRa]

wrong number, nm

 

[Tup3x]

Yeah, pretty much AMD is going all out, with balls to the wall, unlocking all of what those GPUs are capable of.

IMO it's stupid to release a product that doesn't do that. Of course, there's a fine line where thing get out of hand but with Zen AMD has been pretty spot on with it. No point in overclocking. Intel's Sandy Bridge and Haswell for example though... They didn't even bother and there's a massive marginal that they didn't use.

Also NVIDIA did set the W bar rather high so I guess that could encourage others to do so...

(I'd prefer something that's not anywhere near 300W though.)

 

[CastleBravo]

I think 505mm^2 N21 is nonsense.

That would give you 83 good dies per wafer. Even if they were all 6900XT and the 6900XT is 3080+5% it means AMD can look to charge around $700 per card. That is revenue of $58,100.

OTOH a wafer of Zen3 dies is 782 good CCDs. Even if all are used in 5600X's that would be a revenue of $230k.

A 500+ mm^2 GPU is far far far too costly for AMD when they can generate 4x the revenue by making Zen3 parts instead.

Edit To Add: The other option is that availability is like the 3080/3090, non existent.

The third option is that the 505mm^2 die beats the 3090 by ~15% and AMD sells it for ~$3k. Bonus points if they put "Rage" in the GPU's name.

 

[Timorous]

The third option is that the 505mm^2 die beats the 3090 by ~15% and AMD sells it for ~$3k. Bonus points if they put "Rage" in the GPU's name.

AMD can put a $3k price tag on it but I doubt it would sell at that price point.

 

[Kenmitch]

The third option is that the 505mm^2 die beats the 3090 by ~15% and AMD sells it for ~$3k. Bonus points if they put "Rage" in the GPU's name.

The leaks misinterpreted rage mode. It's what JHH will go into if Big Navi takes the crown.

 

[Krteq]

Summing up all leaked info...

 

[Glo.]

AMD can put a $3k price tag on it but I doubt it would sell at that price point.

Looking at the people who bought RTX 3090, in volume, I presume at this point of stupidity, AMD would sell ALL of its RX 6900 XT Rage edition inventory, if that would be the case(15% faster than RTX 3090 at 3k USD).

 

[Mopetar]

Yeah pretty much it. Plus his performances on events were always full cringe.

Raja had a special way of sucking out any excitement you might have had for a product going in to an event he was headlining. Physicists should study him to see if he emits some kind of anti-hype particle because if anyone could manage to stop the current Navi hype train dead in its tracks, it would be Raja.

I honestly think that he should leave Intel and go to work for some abstinence only education group, because Raja is the only person I can think of that could manage to make a concept like "horny college coeds in your area" completely unappealing. Then again maybe not. He might make sex sound so boring that humanity goes extinct.

 

[linkgoron]

Raja had a special way of sucking out any excitement you might have had for a product going in to an event he was headlining. Physicists should study him to see if he emits some kind of anti-hype particle because if anyone could manage to stop the current Navi hype train dead in its tracks, it would be Raja.

I honestly think that he should leave Intel and go to work for some abstinence only education group, because Raja is the only person I can think of that could manage to make a concept like "horny college coeds in your area" completely unappealing. Then again maybe not. He might make sex sound so boring that humanity goes extinct.

Obligatory

https://imgur.com/vI8DM0u

 

[Veradun]

To me, the strange feeling in the RDNA2 lineup is the uncommon rumored gap between the top of the line (80CU) and the next smaller die (40 CU) and the very small gap between Navi 22, 23 and 23 to 24. Granted that we will see cut down Navi21 dies, but seeing that a certain performance target could have been covered with a slightly larger die (60-64CU) compared to the rumored 40CU of Navi22, while 40 is also quite near the 32 CU figure for NAVI23 (and both will be also cut down). One possibility is that AMD is trying to target these GPUs not only for the desktop marked but especially for the mobile market, where a smaller bus for the memory and possibly better clock/power/voltage curve compared to the competition could be a winning factor for OEMs.

The gap between N21 and N22 can be filled both with cuts (72 and 64, that also supposedly have lower frequency) and with the higher rumoured frequency of N22. All in all Polaris was 36CUs and Vega 64CUs (with only ONE cut).

The rumoured 32CU chip is the outlier here, it might be a mobile/apple/microsoft only SKU (low freq, hbm -> very low power)

 

[Veradun]

I think 505mm^2 N21 is nonsense.

That would give you 83 good dies per wafer. Even if they were all 6900XT and the 6900XT is 3080+5% it means AMD can look to charge around $700 per card. That is revenue of $58,100.

OTOH a wafer of Zen3 dies is 782 good CCDs. Even if all are used in 5600X's that would be a revenue of $230k.

A 500+ mm^2 GPU is far far far too costly for AMD when they can generate 4x the revenue by making Zen3 parts instead.

Edit To Add: The other option is that availability is like the 3080/3090, non existent.

Change that 700$ to 999$ and it will work.

 

[linkgoron]

The gap between N21 and N22 can be filled both with cuts (72 and 64, that also supposedly have lower frequency) and with the higher rumoured frequency of N22. All in all Polaris was 36CUs and Vega 64CUs (with only ONE cut).

I believe that a smaller Vega 11 was canned as it didn't make financial sense vs Polaris

 

[Glo.]

The gap between N21 and N22 can be filled both with cuts (72 and 64, that also supposedly have lower frequency) and with the higher rumoured frequency of N22. All in all Polaris was 36CUs and Vega 64CUs (with only ONE cut).

They could also add more CUs to Navi 22, that were taken out from Navi 21.



Im sorry...

 

[Veradun]

They could also add more CUs to Navi 22, that were taken out from Navi 21.



Im sorry...

 

[Glo.]

Im genuinely surprised, that we are not discussing one of the features, that appear to be confirmed, at this point, to be featured, and working for RDNA2 architecture, and for RDNA1.

I wonder if anyone knows what Im talking about, which landed in recent times in upcoming driver release.

 

[Kenmitch]

They could also add more CUs to Navi 22, that were taken out from Navi 21.



Im sorry...

So that's what AMD's HyperTransport is used for. It's about time they put it to good use!

 

[Panino Manino]

These crazy clocks.
I hope AMD isn't pushing things too far out of necessity leaving no reserve for latter.

Or just maybe it has something to do with them moving from massively cash starved to actually having the resources to do a decent job? That's much more a priori probable than a single management person changing.

I sure it does, but what I believe we are really discussing here on the subject of Raja is, "did he made things worse (for RTG and himself)"?.

I say that we aren't being unfair to Raja.
He wasn't just one more working there at RTG, he was supposed to be someone important, someone that could make a difference (for better or worse). It's the same ruler we are applying for the other big players at AMD.
Lisa, Jim, David, Michael, did they made things better with their presence or worse? Did they inspired trust and doubt?
If the answer to this question is different for Raja he has no one to blame than himself.

 

[Kenmitch]

You guys trying to jinx AMD again with all the Raja talk?

Poor Raja!

 

[Glo.]

You guys trying to jinx AMD again with all the Raja talk?

Poor Raja!

Poor AMD!

 

[blckgrffn]

Looking at the people who bought RTX 3090, in volume, I presume at this point of stupidity, AMD would sell ALL of its RX 6900 XT Rage edition inventory, if that would be the case(15% faster than RTX 3090 at 3k USD).

I bought AMD stock. I encourage this behavior.

(I really think they can have a lot of longer term upside in 1-5 years as they steadily pick up server CPU market share, the GPU business outside of the APUs for consoles seems like a bit of a side hobby )

 

[Glo.]

I bought AMD stock. I encourage this behavior.

I only invest in crypto.

Ha! Buy more GPUs and mine on them so I can earn more!

 

[blckgrffn]

I only invest in crypto.

Ha! Buy more GPUs and mine on them so I can earn more!

Fiat for crypto is so much easier. I'll let you do the hassle, I had my turn too briefly too late to be awesome but early enough for actual returns.

 

[HurleyBird]

Or just maybe it has something to do with them moving from massively cash starved to actually having the resources to do a decent job? That's much more a priori probable than a single management person changing.

It definitely has something to do with RTG borrowing some excellent Zen engineers (clockspeeds), but cash starved? Throwing more cash at the same engineers doesn't make them better engineers. Granted, they may have hired some additional talent, but there's a reason why simply throwing cash at a technical problem doesn't go very far. Just look at Intel right now.

 

[kurosaki]

500mm ZEN 4 FTW!

But seriously, why aren't CPUs a little bigger and far more performant if the *individual* core size was doubled, so that IPC was doubled (in theory) - Is the know-how just not there to scale IPC with individual core size?

I think this has been run trough an earlier x86 review by Anand him self many years ago, or something like that. But I'll give it a try:

The X86-64 CPU core is like hardware for a specific instruction set. (ISA). That ISA works in a specific manor and demands a specific set of work to be done, the core does that job. You can work in parallell, eg, more executions per clock or, serial, more MHz.
To drive the parallell workload per core you need more transistors, but they also eat more energy. A wider front end in a x86 sense has to have a lot more to it to make any use of the parallellism. more transistors again.
As the nodes shrink, the main x86 core has gone so small that you produce an enormous wattage at a small surface are = W/cm2 has gone up = bad.
to work on improvements on such a small core for more performance without burning up would itself be a feature. there's where several-core comes into the picture. If you have reached the watt per mm^2 ceiling, duplicate for more performance and so on.
If they could have come away with just doubling everything and get away with it, they would have long time ago. It's all connected to w/mm^2 in the end and a game of trade-offs.

I think ARM64 is better suited for the further development of the single core, where they in some instances are far beyond both AMD and Intel in IPC. The future isn't now, but soon.

In five years, it would be sick if I ran a gaming rig consisting of a Win for ARM, a ARM CPU and an INTEL GPU. The odds on that, would not bet any serious money... But it is a plausible scenario, everything can happen, and it's happening faster and faster. If a platform is on its peak, another will follow soon..

 

[caswow]

It's all forum psychology, he is a scapegoat. AMD didn't deliver a card that could beat Nvidia while he was there and he has now left. Putting all the blame on him means that people can hope RDNA2 can now beat Nvidia because he's not in charge. Where as in-fact he was just a cog in a very big wheel and him being there, or not being there will have in the end made a very small difference to how RDNA2 turns out.

people need to decide either people who get paid big bucks need to bring good things if not take all the blame. thats how it should be. on one side people get told "hurr durr only few people can do xy job thats why they get big bucks" on the other hand we have people telling us dont put the blame on person xy