• 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 712 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Yes, but one 96 core chip at 360 vs 2 360 watt chips that perform the same ? Intel needs to get with it.
SPR seems to top out at 350W. And no one was debating that there's a huge efficiency gap, but you claimed some very particular numbers that clearly don't match either platform.
 
SPR seems to top out at 350W. And no one was debating that there's a huge efficiency gap, but you claimed some very particular numbers that clearly don't match either platform.
I was estimating, and not far off.... Jeesh, give me a break. Even you agree its like 2 to 1 or 100% more efficient. Why haggle on numbers that are close.
 
I was estimating, and not far off.... Jeesh, give me a break. Even you agree its like 2 to 1 or 100% more efficient. Why haggle on numbers that are close.
Off by several hundred watts all told. And I don't think anyone was making into a big deal, but at the same time, if you're interested in how they compare in power, that information has been both readily available and posted to this forum in several places. It would be surprising for such a prolific member to not have seen either data point yet.
 
Off by several hundred watts all told. And I don't think anyone was making into a big deal, but at the same time, if you're interested in how they compare in power, that information has been both readily available and posted to this forum in several places. It would be surprising for such a prolific member to not have seen either data point yet.
So you are saying I read every single post in every single thread ? You just like to troll me every place I post.
 
So you are saying I read every single post in every single thread ? You just like to troll me every place I post.
You didn't claim to be guessing. You said, and I quote, "you can bet that 2P system is running over 1000 watts". So why was it ok for you to make that claim, but not me to correct it?

And why is it that your "guesses" always seem to radically favor AMD beyond all reality?

You constantly claim to be personally concerned with power numbers, but can't be bothered to make accurate claims?
 
I have watched the whole about "Chalk Talk" event...

It looks like, because of various accelerators, SPRs achieves x3~6 perfomance (cf. ICL-SP) in major DC workloads.
This is better than I had thought.
Are SPRs and Eagle stream great jump similar to the Nehalem-EP?
 
Are SPRs and Eagle stream great jump similar to the Nehalem-EP?
Well to be honest there was nothing impressive about Core2 Xeons back then(Or Server Grade CPUs back on those days), But yeah in Niche Products(mind you that GPUs and ASIC are many times more efficient and more powerful at doing that than anything SPR-SP can mustard), this is a pretty big jump in performance.
 
Last edited:
Core for Core this is the fastest(Multi-Threaded) Intel Sapphire Rapids-SP processor and the most competitive with Genoa(Core for Core)

The Intel Xeon Platinum 8468 48C/96T
1667500558831.png


2S System(Two Sockets for 96C/192T)

Geekbench5: 74,853 Points
1667500616341.png

Cinebench R23

MT: 90,411 Points
1667500654070.png

V-Ray
MT: 85,766 vsamples
1667500636106.png


Sisoftware Sandra Native Arithmetic:

1,900.91GOPS

1667502703561.png
 
Last edited:
This appears to be the original source:

Not much info to go by. They don't even explicitly state that it's Granite Rapids. But if it is GNR, then it would be a promising sign if they have DDR5 working better on first/early GNR silicon than on mature SPR/EMR silicon. But obviously way too early to really tell.
 
sp-.jpg

You must include personal commentary when posting links/pics. You've already been warned once.

Daveybrat
AT Moderator
 
Last edited:
Xeon Max (HBM) 2023

l_si7101-Max-01.jpg




All three of your first posts violate the no posting of images or links only. Read the rules before you post again


esquared
Anandtech Forum Director
 
Last edited by a moderator:

I guess that Genoa-X has twice Milan-X performance by 96cores, 12ch-DDR5 and AVX512.
Is Xeon MAX without "AMX effect" good product?

Maybe Xeon MAX boasts perfomance/cost and perfomance/watt.
We know that DIMM slot(12ch-DDR5) is expensive equipment and DRAM access is a waste of energy.

However, basically, I think that EMRs with HBM3 should compete Genoa-X.
 
You know Intel is in a Bad Shape when they can't even fact check their own product line and are looking for a headline.

1668014671336.png


Xeon Knights Landing with HBM and same memory configurations.
1668014831111.png
 
Last edited:
Hey someone tell Intel that their Intel Xeon Phi Knights landing was a x86 CPU with HBM with the same HBM Memory Options.

View attachment 70652

View attachment 70653

It was HMC (Hybrid Memory Cube). There were multiple competing stacked memory solutions in development at that time.


 
Back
Top