Not really, if it's on N2 (or even N3P). 6 GHz was the target frequency for Zen 4 in fact.
Not really, if it's on N2 (or even N3P). 6 GHz was the target frequency for Zen 4 in fact.I think the 1.1 is a very solid prediction. 5.7GHz>6GHz seems like a stretch though.
6 core CCD 💀6Ghz Zen 6 with 600MB of L3 and DDR6 on 6/6 of 2026...
They cant pass that up.
I think the new IOD is going to help as will the larger shared cache (due to 12 core CCX). I think a few tweaks to the core itself will also contribute, and of course, support for faster memory.So what are current thoughts on overall IPC (throughput/MHz) increase for ST for Zen 6?
Are there still some who are going to make outlandish claims like they "invented the question mark" and that Zen 6 will be +30% ST over Zen 5?
I'm a little late answering this question, but my philosophy is to mostly ignore rumors and just base my expectations off what has been achieved in prior generations. AMD has averaged low 20s percent performance-per-core every couple years since Zen1.So what are current thoughts on overall IPC (throughput/MHz) increase for ST for Zen 6?
Are there still some who are going to make outlandish claims like they "invented the question mark" and that Zen 6 will be +30% ST over Zen 5?
You'll get stuff and a lot of frequency.So what are current thoughts on overall IPC (throughput/MHz) increase for ST for Zen 6?
Gotta make use of nanoflexYou'll get stuff and a lot of frequency.
Client is out of N-1 ghetto for once!
10GHz was the target frequency for P4.Not really, if it's on N2 (or even N3P). 6 GHz was the target frequency for Zen 4 in fact.
10GHz was the target frequency for P4.
Yes, only because we have been in the ~ 5Ghz window for quite some time. Increasing frequency beyond a certain point starts getting you exponentially higher thermals.... and failures.If Zen 5 is 5.7 GHz on N4P, is it that inconceivable that on the originally planned Zen 5 on N3 could reach the original target of 6.0 GHz?
I think that this is much more likely than the 3nm Zen 6 reaching 6Ghz! In fact, this is exactly the kind of thinking and strategy that would make a server based Zen 6 chip an enormous success.I hope they can achieve 5.1GHz under 15 watts that would good. That would instantly make them leaders in x86 efficiency.
That also means they can cram in more cores for mobile. Its doable on N2 i think
a) it's N2I think that this is much more likely than the 3nm Zen 6 reaching 6Ghz
I have heard the potential that the desktop and laptop might be N3P while the server is N2.a) it's N2
b) speed is funny (also matters)
old old old old stuff.I have heard the potential that the desktop and laptop might be N3P while the server is N2.
not anymore!This is essentially how AMD is running now (N4P desktop/laptop and N3E server dense).
oh hell no.Intel seems oblivious lately to costs
I agree with this. Currently with the 9950X 5.7GHz is only reached with 1 or 2 core or very light loads without extreme cooling. We will see what happens but I doubt boost frequency will increase and even it it did it wouldn't matter in day-to-day use situations.Yes, only because we have been in the ~ 5Ghz window for quite some time. Increasing frequency beyond a certain point starts getting you exponentially higher thermals.... and failures.
More than anything else, it just hasn't been economically sound. Intel pushed the boundary and paid the price. I am relatively certain AMD isn't going to rush out and try to duplicate their mistake.
Of course, I could always be wrong .
Is that the sound Raptor Lake makes when it fails?bazinga
Yes and also it has the double meaning of signifying 1.5+ volts of Vcore.Is that the sound Raptor Lake makes when it fails?
Arrow Lake is considerably more expensive to produce than Zen 5, but is under performance.
I'd be satisfied to just have 16 cores at 3.5-4.0 Ghz under 65W.I hope they can achieve 5.1GHz under 15 watts that would good. That would instantly make them leaders in x86 efficiency.
My point is that Intel needs to focus on design for profit..... and I agree, they don't have a choice OTHER than to start designing for profit or they wont exist.Intel doesn't have a choice. For the long run it's also a war of attrition and AMD remains unchallenged
That's what Novalake is for, brother.My point is that Intel needs to focus on design for profit.