Golden Cove client and server appear completely different in die shots, but Raptor Cove and Golden Cove server are indistinguishable. Don't think you can find any Raptor Lake where the AVX-512 isn't fused off.
Well, it's new silicon. Took the server variant Golden Cove cores from Sapphire Rapids, called the Raptor Cove, and improved the e-core implementation a bit.
I generally agree, but I also think eventually you reach a point where a sufficient deficit in ST just makes any advantage in MT irrelevant.
If a product gets 0.9x ST/1.2x MT, that's viable.
But if a product gets 0.66x ST/1.66 MT, that's a really touch sell tbh.
I'd take 8 Zen1 cores over 4...
It doesn't exist in large part because there aren't any good components for it.
The main draw for STH in a Mac Studio type form factor is you can include gobs of unified memory for AI workloads.
The analysis *really* depends on how AMD decides to construct the thing. Strix Point suggests the existence of an 8c Zen5c CCX (assuming its 4+8 config isn't all implemented on one big, shared CCX). It could be that for Strix Halo, there's, for example, an 8 core Zen5c CCX sharing the same die...
It should have around 3x the bandwidth.
That's an assumption, but probably a safe one. Strix Point is 4+8, so 8+8 makes sense for Halo. 16+0 would be a waste of power and die space. But, bear in mind AMD traditionally sacrifices a lot of stuff on the alter of getting parts ready sooner.
I...
Lots of variables. Halo obviously wins all bandwidth bound tasks. Monolithic improves latency but otoh the whole memory subsystem is likely geared towards bandwidth at the expense of latency.
I'm assuming Halo is 8 Zen5 + 8 Zen5c, so in that case 9950X obviously wins all embarrassingly parallel...
Good response.
I think there's a phenomenon that when someone is not all that deep, but is sufficiently wide, and especially if they are wider than you are, you err towards giving yellow flags the benefit of the doubt. And, when the alternative is someone like RGT who throws a dozen red flags a...
It wouldn't be entertaining if the fakes were actually good, or they put in a solid effort to appear authoritative. The whole point of the trap is that it's easy to avoid.
I've defended him before. He has had some legit and important leaks, but this is bad. Very sloppy journalism.
Hard to say whether his hit-rate will go up or down after this. He might become more diligent, yet at the same time legit sources may be less likely to contact him now that his image is...
The issue with the Terrascale gen was less about the volume play, and more about using a mid-range die as their high end. Keep everything the same, but add a 4970 as a halo card, which thanks to AMD's monumental architectural advantage would have blown GT200 out of the water, and that gen would...
On one hand, AMD prefers to be a bit conservative. On the other hand, there's always someone making outlandish claims on any future product for attention. I assume the truth is somewhere in-between.
Actually, let me expand on this. The odds of any hype train being realized in an environment where rather unexciting internal projections are dismissed as sandbagging and hyper-conservatism, and where whispers of development problems and missed targets are starting to materialize, is effectively...
I hope so too, of course. But if I had to put money on it either way, it would have to go on the side of the hype train derailing, as it almost always does.
More MLID. Claims that Z5 is still strong, but not as strong as originally intended due to bugs. Reasserts he considers the 30%+ people dead wrong.
Well, he was first out the gate with Z5 delay, and generally speaking hype trains are ultimately derailed far more often than they are realized.
I believe that frequency and voltage being equal, the dense cores do consume less power, but I think that's all based on transistor count, distances, and maybe process, rather than anything architectural.
I think they're dropping the ball a bit on marketing (It is AMD, after all).
Even if AMD doesn't think it will be important, they should hammer on the fact that they are beating Nvidia with a much more general approach. They hint at this a bit by comparing to both tensor and CUDA flops, but...
It would be right in line with AMD's history of self destructive GPU business strategies. G80 wasn't the beginning of the end for AMD's aspirations of graphics leadership, RV770 was. AMD needed a R600/G80 moment, and could have had it just one generation after G80, but despite a far superior...
So, here's an issue:
This sounds exactly like your "I have insider info" voice, and it just makes your entire persona seem unlikely.
I'm willing to entertain the idea, at least until disproven, that you might have confidential information regarding AMD server tech, AMD in general, or server...
Or maybe optimizing for power first, IPC second-ish, and frequency last-ish just tends to get you to a better place than when you try to do it the other way around.
Anyway, I suppose it's possible Adroc only knows of Turin-D and there's some other bottleneck there that makes the SMT yield look bad, but if you look at the full context of the whole SMT discussion back then he's clearly claiming that the yield is way down (and Kepler makes similar claims).
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.