It seems like AMD may have some prospective supercomputer bids in progress.
AMD may be the only game in town when it comes to scientifically oriented supercomputers, after Intel withdrew itself following the Aurora disaster, and NVidia is too busy counting the money from AI. (And its cards have nerved FP64 even more than before)
adroc_thurston
Diamond Member
- Jul 2, 2023
- 6,260
- 8,815
- 106
Works the opposite way too, RDNA SIMD and cachemem models are far, far superior.
Hmm. Let's see here. Who is currently ahead in performance? Yea, AMD .... and from a node behind no less.
Who just wet the bed with PTL yields? Yep.
Lets look at WHY companies design new processors. It isn't because they screwed up the engineering on the current generation. It's because new process technology has come along that enables higher transistor density and thus larger transistor budget in the same die area. This transistor budget is used to improve the performance.
Since process technology is slowing down rapidly, so will performance increases. It's just math.
I will say that AMD has been working very smart with TSMC. Chiplets, 3D stacked cache, etc. These are all great ways to do more with less.
To directly answer your question though, I don't think AMD is going to feel any pressure to release Zen 6 ..... rather on Zen 7.
The way things are going over at Intel, I wouldn't be surprised if AMD sold most of its desktop and laptop models on N3P and didn't even bother with N2 except for DC. With no threat of 18A and 14A (high NA) to push them, I think AMD will resort to profit taking .... like any company would.
Speaking of Transistor Densities, AMD has been delivering CPU / GPU chip dies with insane transistor densities, higher than its peers.
IMO, this could be a data point that AMD is employing (well) some AI tools.
I don't think AMD is going to take the foot off the gas pedal. There are the Arm hordes, Arm itself, hyperscalers using Arm, NVidia using Arm.
Volume is what accelerates the profits. Being a small minor player in number of markets - most importantly client laptop, and then, of course datacenter GPU, the easiest way to greater profit is to gain market share, to move up from being a bit player.
In fact, AMD is doing the opposite of what you suggest - to reach the state where I think more profits are:
In datacenter GPU, AMD has moved to N3P with Mi355 ahead of NVidia which is using N4P, in order to gain more share in the market. And later, in 2026, NVidia moves only to N3P and AMD moves to N2P on Mi400.
I think that any company with a comfortable lead will take the foot off the gas. It's called profit taking.
If AMD can maintain a lead without releasing a new processor on a new node, they will do it.
In the case of nVidia, AMD is NOT in the lead. Still your point is valid. in DC where AMD is in the lead, they are paying out the nose for the leading edge process .... but then again, in DC it is worth every penny!
So are they going to be diverging the ALU design between MI450X and MI430X?
Because as of CDNA2 the ALUs are 64b native ALUs not 32b native ALUs... so they would need a different ALU design in order to do what you are suggesting without it still physically being there...
So... if it's the 1/32 like you said then it should be 8 FMA per cycle per CU for GFX1250...
I am also going to question that they are getting rid of the extra ALU lanes that GFX1200 has along with cutting the matrix performance significantly... By the math you have there, assuming it's FMA ops per cycle per CU and not FLOPS per cycle per CU, then they are cutting the FP16/BF16 Matrix performance in half compared to CDNA4... If you are talking about the FMAs per clock then it's a halving of the FP8, FP6, and FP4 Matrix FLOPs per cycle and a quartering of the FP16/BF16 Matrix performance compared to CDNA4...
And I am doubtful that they are going to spend the RnD time for the different GFX125X ALUs... That isn't AMD's style...
I personally think you are extrapolating far too much from Github repos...
yeah they are using densest libraries available with more metal layers to get a denser but performant as well
Given they have managed to get in at least 2 of the last faster supercomputers that isn't a stretch.
fastandfurious6
Senior member
- Jun 1, 2024
- 657
- 835
- 96
100% even if amd decides to do most volume of desktop/laptop/client one node behind N3P etc it still needs a cutting edge N2 top chip to keep the real substantial lead
btw how much are latest PTL yields?
AMD has tremendous pressure to release Zen 6 as fast as possible. It's part of the MI400 ecosystem. Intel is also using leading-edge on laptop and desktop, so AMD can't sit behind them either. If AMD wants to offer more affordable SKUs, they will likely rename Zen 5.
I guess it makes sense to expect the last generation on an AMD platform to reuse the IOD from the previous generation. We'll see, I guess.
A bit of a tangent, but the gaslighting over Kepler's reply to his claim that that the next 4 generations would use lga 1954 is hilarious. He did this in the last video before this one too.
The hard to believe part isn't that NVL isn't going to use DDR6, it's the part where NVL+3 will still only be on DDR5.
SteinFG
Senior member
- Dec 29, 2021
- 719
- 855
- 106
RPL refresh is still compatible with ddr4, so not that crazy
This is a poor long term strategy.
Intel did this and look at them now. They stagnated performance and ASPs during a period of minimal competition and then they missed with their fabs and a couple of architectures when that competition increased putting them behind.
NV did not do this and look at them now. They kept pushing performance and increased ASPs during a period of minimal competition and moved into new markets that have enabled them to really leverage their advantage through closed eco systems like CUDA. Even with misses like Fermi or Blackwell their very reliable execution and their performance and feature leadership means even during periods where they have increased competition they don't lose much if any ground.
I think given what has happened to Intel and what NV have done AMD are far more likely to take the NV path of just going for it. On top of that compute demands are increasing so there will be customers for more expensive and more performant parts which will allow AMD to just charge more for that extra performance. It is a better strategy and NV have proven that to be the case. If and AMD shareholders want AMD to follow the Intel path then they are morons.
Honestly Zen 7 even with 32 Cores would probably be fine with 8000+ MT/s which will be super common by the time it comes out. Like quality 8800+ MT/s with tight timings will be easily enough especially if CAMM2 DDR5 comes to AM5 along with LGA 1954 (wouldn't be surprised if CAMM2 becomes the standard on that socket, near 100% of people probably buy new memory with a new motherboard when upgrading anyways) it'll be enough for Zen 7. And for those with a 16C single CCD; why would even DDR6 show a benefit in the vast majority of cases? Especially with gaming on X3D?
That said I do wonder if they'll do 16 Core Zen 7 X3D that can do all core boost to 7Ghz at minimum, would be insane for gaming especially for 10th gen gaming, would exceed 2x PS6/Xbox Next in CPU performance (which likely probably have cut down cache just like this gen). Very capable for 240Hz gaming.
That said I do wonder if they'll do 16 Core Zen 7 X3D that can do all core boost to 7Ghz at minimum, would be insane for gaming especially for 10th gen gaming, would exceed 2x PS6/Xbox Next in CPU performance (which likely probably have cut down cache just like this gen). Very capable for 240Hz gaming.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter