• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Intel current and future Lakes & Rapids thread

Page 697 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
nicalandia said:
Yes, I agree. But I blame that on Buggy Bios, Buggy MB and plain old Buggy ES Sample SPR.

Still 90,000 for a 2S System(96C/192T) it's pretty average, there has to be more room for performance, I want 120,000+ on release date. 😛
Click to expand...
So you're expecting SPR to match Genoa? Alright man.
 
uzzi38 said:
I don't know about you but I'd personally want to make assumptions on this once someone does some more in-depth testing on power at varying clock speeds. I'd do it myself like I did for RDNA2 but I haven't seen an A380 on sale in the UK yet...
Click to expand...


I was mainly talking about the clock speed wall, the upper clock speed limit of Xe LP on Intel 7 is low compared to AMD iGPUs. With Xe HPG and TSMC 6nm they upped the clock speed wall by about 50%. The Asrock A380 challenger has a GPU power limit of 55W which is enough for a GPU clock of 2450 Mhz. Mobile A370M requires 30-32W at 2200 Mhz if it's fully utilized.
 
mikk said:
I was mainly talking about the clock speed wall, the upper clock speed limit of Xe LP on Intel 7 is low compared to AMD iGPUs. With Xe HPG and TSMC 6nm they upped the clock speed wall by about 50%. The Asrock A380 challenger has a GPU power limit of 55W which is enough for a GPU clock of 2450 Mhz. Mobile A370M requires 30-32W at 2200 Mhz if it's fully utilized.
Click to expand...
Different games/workloads can require different levels of power for the same clock, even if the GPU reports 100% usage. It's the reason why I chose TimeSpy GT2 very specifically when I did the RDNA2 testing a while back - even more so than GT1 I found that it was the heaviest test I could run.

That's why I'm saying I'd wait until you can get a proper power/frequency curve in a specific workload you can compare to other architectures.
 
uzzi38 said:
Different games/workloads can require different levels of power for the same clock, even if the GPU reports 100% usage. It's the reason why I chose TimeSpy GT2 very specifically when I did the RDNA2 testing a while back - even more so than GT1 I found that it was the heaviest test I could run.

That's why I'm saying I'd wait until you can get a proper power/frequency curve in a specific workload you can compare to other architectures.
Click to expand...


Tested on different games and workloads of course.
 
Markfw said:
"Intel's recently delayed its Sapphire Rapids launch again, with the chips now slated to arrive in 2023. Gelsinger remarked that while the new chips are 'better than the AMD alternatives" in power and performance and will win in some benchmarks, the advantages aren't dramatic enough to slow AMD's advance.
Click to expand...

There's no Genoa yet and early shipment versions are shipping so it may be a very stretched definition of "better". He said in the same interview they'll be "competitive" in 2024 meaning they aren't going to be.

@witeken

Didn't you also claim Meteorlake uses an active interposer like Lakefield which is totally wrong?

I doubt it. Gelsinger said 10-12% gain over what it would have been originally so that's Redwood Cove. It's not bad but not a major architectural change.

@Exist50 When asked about LNC not being Royal Core the twitter guy few posts above says Royal Core Project is a PROJECT not a core and it encompasses more than one generation.
 
Last edited:
Markfw said:
This quote below, I just don't get, tell me what I am missing. AMD has it all over SR, Milan does and Genoa is even better yet.
Click to expand...
He might still be comparing to Milan. That, or something that makes heavy use of either SPR's accelerators or AMX. Doesn't really matter though; everyone knows SPR won't be competitive in the general market with Genoa.
msj10 said:
this twitter account got some details right about the Zen 4 architecture before the announcement but still take it with a grain of salt
Click to expand...
If it's right, that's not good in all sorts of ways. Shipping RWC in 2024 when LNC was supposed to be ready for '23? That's a bad look in its own right, and then there's the reality check on what RWC is...
IntelUser2000 said:
Interesting. He's saying Royal Core project starts from Lion Cove. That's pretty exciting. Arrowlake is the first.
Click to expand...
If that's what his comment means, then he is absolutely, 100% BSing. "I'm THE SOURCE" my butt! Royal has nothing to do with either Lion Cove or Panther Cove. That I can say with complete certainty.
 
Yea that tells us how bad the original Granite Rapids would have been if they said porting to Intel 3 ended up being a much better decision despite coming later than expected.

Exist50 said:
If that's what his comment means, then he is absolutely, 100% BSing. "I'm THE SOURCE" my butt! Royal has nothing to do with either Lion Cove or Panther Cove. That I can say with complete certainty.
Click to expand...

With two camps saying they are absolutely right(even though I lean on you being correct) they are doing a good job hiding secrets.
 
Exist50 said:
If it's right, that's not good in all sorts of ways. Shipping RWC in 2024 when LNC was supposed to be ready for '23? That's a bad look in its own right, and then there's the reality check on what RWC is...
Click to expand...
he also says LNC design was finalized early this year. how long does it usually take from the core design being finalized to release?
 
IntelUser2000 said:
Yea that tells us how bad the original Granite Rapids would have been if they said porting to Intel 3 ended up being a much better decision despite coming later than expected.
Click to expand...
If this is all they could manage, then I'm not convinced it was the better decision.
IntelUser2000 said:
With two camps saying they are absolutely right(even though I lean on you being correct) they are doing a good job hiding secrets.
Click to expand...
Tbh, I think it's just some "leakers" parroting each other in the hope one of them actually knows what they're talking about. But that falls apart completely when they do not.

Honestly, I find it absurd now some of these purported "leakers" have such a fundamental misunderstanding. I don't think Intel's even trying to hide it. It's just people unwilling to do any due diligence.

But do feel free to hold me to my word on this. There's not a single doubt in my mind. Royal and LNC/PNC are independent projects, and there will be absolutely zero room for doubt when Royal does arrive. That I promise you.
msj10 said:
he also says LNC design was finalized early this year. how long does it usually take from the core design being finalized to release?
Click to expand...
Depends what you mean by finalized. RTL complete? Val complete? PD/SD complete? Taped out? Regardless, I think the product roadmap would be much more insightful, and in that regard, it seems that Lion Cove will arrive in Arrow Lake and Lunar Lake sometime in 2024.
 
Last edited:
IntelUser2000 said:
@Exist50 When asked about LNC not being Royal Core the twitter guy few posts above says Royal Core Project is a PROJECT not a core and it encompasses more than one generation.
Click to expand...
Oh, somehow I missed this part. Link to the tweet in question.
https://twitter.com/x/status/1562718580046954498

I know I'm being redundant, but this is 100% wrong. Royal will be a new core, and the start of a new series, but one that has nothing to do with LNC, PNC, etc.
 
igor_kavinski said:
Are you Jim Keller? Coz you are way too sure about this.
Click to expand...
Hah, I wish. But nah, I'm no insider. Remember what I said about most of my knowledge being people-centric? Well it's pretty darn hard to hide the fact that they've created a whole new(?) CPU team. Nor do some of these "leakers" set a high bar to beat.
 
igor_kavinski said:
Anyone special in that team? Like ex-Apple, ex-AMD?
Click to expand...
In terms of names people would recognize from other companies, none that I'm aware of. The foundational architecture team seems to mostly be Intel old guard that were wandering around corporate no-man's-land for a while. Some from the Willamette days. Some more from Intel Labs. One or two hired from academia. A couple emerging from retirement. Etc. Though that said, the team's apparently been growing a lot (as they'd really have to to ship a core), and I don't have a clear picture on where all the new hires are coming from, so maybe there's someone there.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top