Question Zen 6 Speculation Thread

Page 277 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Thibsie

Golden Member
Apr 25, 2017
1,136
1,338
136
Yet Arrow Lake was launched only 2 months later than Zen 5...on N3. That makes no sense.

On N3B. Intel got f*** or rather they did it to themselves.
AMD judged N3B wasn't worth the cost and moved back to N4p. Which seems to have been the smart choice.
 

reaperrr3

Member
May 31, 2024
136
394
96
Intel is competitive in desktop and laptop, less so in DC.
Yeah, which is why using an equal or better process than Intel for premium laptop and most desktop SKUs is even more important in premium client, and Intel will use N2(P) for NVL, so AMD will too and only use N3P for cheap mainstream laptop APUs (which also serve as IOD for premium laptop SKUs with additional N2P CCD).

I have been speculating this for some time.
You mean you've dug in and stubbornly refuse to climb out of that hole...
(and Tigerick is totally the right person to listen to...🫠)
Intel just bought/reserved huge N3(B) capacity early, to get a discount (which Pat killed with his stupid public remarks regarding TSMC) and compensate for their inferior big cores in terms of PPA/PPW, while AMD was more careful and ported Zen5 back to N4P/X early for everything but the high-margin DC chiplets, where higher wafer cost and worse yields of N3E were easier to offset.

And by the time Intel realized N3(B) was meh for its cost, it was too late for them to take the same route as AMD.
So of course it makes sense, you're just selectively ignoring facts that don't fit your narrative.

I really can't wait for your reaction when the process for the 12C CCDs is finally revealed.
But I can already guess how you'll react:

"It makes not sense, it would have been smarter of them to use N3P, why are they doing this, I'm making business decisions like this for a living so I know better, they're making a business mistake by doing this, yadda yadda..."

Big kudos to you if you actually just admit you were wrong when that time comes.

It sucks as much as AMD's core. both LNC and Zen 5 have similar perf/w and IPC.
Zen5 isn't super great (for client), but
- it's on a slightly less efficient process
- it reaches that IPC with a less wide core and less total cache per core
- in MT workloads, ARL is only competitive thanks to Skymont; basically, without E-cores, the lack of SMT would've been catastrophic, while with SMT, LNC would've been even bigger and even less power efficient
 

Fjodor2001

Diamond Member
Feb 6, 2010
4,235
594
126
idk maybe their big core sucks?
idk maybe their E core rules?

You have to look at the complete P+E+LPE+iGPU+etc offering.

What matters is the total solution (ST and MT perf, perf/watt, iGPU, NPU, etc), and the price of that. And what parts of the total solution are important depend on market segment.

So the statement that "But Intel's just not gonna be competitive till UC." does not hold, if you by that mean that only P core performance matters regardless of market segment.
 

jpiniero

Lifer
Oct 1, 2010
16,849
7,292
136
idk maybe their E core rules?

You have to look at the complete P+E+LPE+iGPU+etc offering.

What matters is the total solution (ST and MT perf, perf/watt, iGPU, NPU, etc), and the price of that. And what parts of the total solution are important depend on market segment.

So the statement that "But Intel's just not gonna be competitive till UC." does not hold, if you by that mean that only P core performance matters regardless of market segment.

What matters is gaming performance, not Cinememe.
 

Fjodor2001

Diamond Member
Feb 6, 2010
4,235
594
126
And in those markets there is much more glue to play with.
For client segment, the glue is in the mobile market, not DT.

And for most mobile/laptop users, what matters is battery life, low weight, and price. Their P cores will be idling 95% of the time. They want good E+LPE+iGPU+NPU, where the last one will be forced upon them by MS Windows requirements whether they like it or not.
 

gdansk

Diamond Member
Feb 8, 2011
4,598
7,731
136
For client segment, the glue is in the mobile market, not DT.

And for most mobile/laptop users, what matters is battery life, low weight, and price. Their P cores will be idling 95% of the time. They want good E+LPE+iGPU+NPU, where the last one will be forced upon them by MS Windows requirements whether they like it or not.
No, as usual Apple shows the right way if cost isn't an issue. Monolith works best for laptop. Neither Intel or AMD is even close in efficiency. But if customers want more they can buy 2 glued together.

The disaggregated monstrosities AMD and Intel are conjuring for consumers are suboptimal. AMD so they can reuse disaggregated server designs. Intel because ???.
 
  • Like
Reactions: CouncilorIrissa

adroc_thurston

Diamond Member
Jul 2, 2023
7,346
10,085
106

Fjodor2001

Diamond Member
Feb 6, 2010
4,235
594
126
No, as usual Apple shows the right way if cost isn't an issue.
If you need a certain fruit stamped on your computer then Apple has monopoly on that.

But for the Windows world, it's AMD vs Intel (and Qualcomm & friends if you venture into Windows on ARM). And then it's a price competition.
 

gdansk

Diamond Member
Feb 8, 2011
4,598
7,731
136
If you need a certain fruit stamped on your computer then Apple has monopoly on that.

But for the Windows world, it's AMD vs Intel (and Qualcomm & friends if you venture into Windows on ARM). And then it's a price competition.
No, it is purely technical. Logos do not matter. What AMD and Intel are doing is suboptimal, cost-saving methods for reuse.

So far when Apple does that for their biggest chips only because it would simply be too big to fit in reticle and unyieldable.
 

Fjodor2001

Diamond Member
Feb 6, 2010
4,235
594
126
24с Olympic Ridge.
And that'll outperform NVL-S 52C in MT?
Yeah? Better or cheaper.
So what specs for CPU + iGPU + NPU are you expecting for Zen6 on mobile/laptop/DT? And why are you expecting that to outperform the offer from Intel NVL, including in perf/watt?

Just as an example: What iGPU are you expecting on Zen6 DT that'll be better than what NVL-S will have?