Meteor Lake seems on track.

L

Lakados

[H]F Junkie
Joined
Feb 3, 2014
Messages
9,719
https://wccftech.com/roundup/intel-meteor-lake-core-ultra-cpus/

Uses a combo of Intel 4 (EUV 7), TSMC5, TSMC6, and Samsung 14
First commercial consumer use for Foveros, new architectures for CPU and GPU cores, new layout, new interconnects, and 8 lanes of PCIe5 dedicated for 3'rd party GPU all crammed into one beautiful mobile SoC, with options for external SoDIMM memory, or on-chip DDR5 (up to 96GB) or LPDDR5x (up to 64GB).
Will also be coming with Arc Alchemist with that 128Mb cache option.


Everything about it looks fun, no clue on how well it performs in the wild, but it looks like Intel is playing with a full deck this time around.
 
Last edited:
So assuming this is for desktop, what is the lineage going to be? Raptor Lake refresh (2023), Arrow Lake (2024), Arrow Lake refresh (2025), Lunar Lake (2026)?
 
wandplus said:
So assuming this is for desktop, what is the lineage going to be? Raptor Lake refresh (2023), Arrow Lake (2024), Arrow Lake refresh (2025), Lunar Lake (2026)?
Click to expand...
This is for the upcoming Mobile refresh, Desktops get Raptor Lake which is supposed to be Oct 17.
Oct 17 is the date for the K and KF series, general consumer and OEM models probably at CES in 2024.
 
LukeTbk said:
It would be so strange if Intel 4 is in between TSMC 3 to 5, they could have a more than doubling of density in one go ?

https://www.anandtech.com/show/1744...il-2x-density-scaling-20-improved-performance
Click to expand...
Not too surprising really, TSMC, Intel, and Samsung are all using the same ASML hardware for these nodes at this size, Intel is just doing something different with the light frequencies than what TSMC is doing for their 3nm process, and doing something else different than Samsung for the transistor design and masking process than their 3nm process.
But at this stage when they need help figuring out why something not working the 3 of them have the same 1800 number, and the same Nvidia-powered hardware running the software that they design the chips on.
Yeah, Nvidia built the AI that ASML uses for chip and process optimization and troubleshooting, Nvidia cuLitho, it's a beautiful piece of work.
https://developer.nvidia.com/culitho
 
Last edited:
I had more in mind for the punditry around the industry, having Intel not a power hog all of the sudden that can do 13900k MT and more performance with 90w will be a bit strange.
 
LukeTbk said:
I had more in mind for the punditry around the industry, having Intel not a power hog all of the sudden that can do 13900k MT and more performance with 90w will be a bit strange.
Click to expand...
Unless I'm misunderstanding you, that's not happening (yet). Meteor Lake is mobile, and maxes out (IIRC) at 6E+8P.
 
1_rick said:
Unless I'm misunderstanding you, that's not happening (yet). Meteor Lake is mobile, and maxes out (IIRC) at 6E+8P.
Click to expand...
I don't know what to expect from the Desktop parts, the leaks have been all over the map so what ever they announce in Oct will be interesting for sure.
 
Lakados said:
Exploits for everybody at this stage, if it's multi-threaded and x86, it seems to have a memory hole in there somewhere.
Click to expand...
I know, on the plus side, Intel usually gets 2-3 years of decent performance uplift before plugging the holes takes it away again.
 
You must log in or register to reply here.
Back
Top