Intel Demoes Core "Lunar Lake" Processor from Two Generations Ahead

erek

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,674
Better than Meteor Lake

“Intel at the 2023 InnovatiON event surprised audiences with a live demo of a reference notebook powered by a Core "Lunar Lake" processor. What's surprising about this is that "Lunar Lake" won't come out until 2025 (at least), and succeeds not just the upcoming "Meteor Lake" architecture, but also its succeeding "Arrow Lake," which debuts in 2024. Intel is expected to debut "Meteor Lake" some time later this year. What's also surprising is that Intel has proven that the Intel 18A foundry node works. The Compute tile of "Lunar Lake" is expected to be based on Intel 18A, which is four generations ahead of the current Intel 7, which will be succeeded by Intel 4, Intel 3, and Intel 20A along the way.

The demo focused on the generative AI capabilities of Intel's third generation NPU, the hardware backend of AI Boost. Using a local session of a tool similar to Stable Diffusion, the processor was made to generate the image of a giraffe wearing a hat; and a GPT program was made to pen the lyrics of a song in the genre of Taylor Swift from scratch. Both tasks were completed on stage using the chip's NPU, and in timeframes you'd normally expect from discrete AI accelerators or cloud-based services.”

1695209159816.png

Source: https://www.techpowerup.com/313823/intel-demoes-core-lunar-lake-processor-from-two-generations-ahead
 
I didn't know Taylor Swift was a genre...

Anyway, creating an image that small could be done pretty quickly on ages old cards, so it's not terribly impressive. Pretty sure they just picked a seed that looked good and just generated it. Definition of NPU says it's specifically a machine learning accelerator. As in, it's 100% focused on that. So... where's the CPU? What exactly is this article saying? That they're deciding to try to get in on the AI bandwagon...? Okay.
 
StoleMyOwnCar said:
I didn't know Taylor Swift was a genre...

Anyway, creating an image that small could be done pretty quickly on ages old cards, so it's not terribly impressive. Pretty sure they just picked a seed that looked good and just generated it. Definition of NPU says it's specifically a machine learning accelerator. As in, it's 100% focused on that. So... where's the CPU? What exactly is this article saying? That they're deciding to try to get in on the AI bandwagon...? Okay.
Click to expand...
Lunar Lake is going to be another low-power laptop design, the successor to Meteor Lake: https://www.techspot.com/news/100216-intel-confirms-arrow-lake-lunar-lake-panther-lake.html Arrow Lake is next year's desktop platform. Panther Lake will be back to a unified desktop and laptop line.
 
StoleMyOwnCar said:
I didn't know Taylor Swift was a genre...

Anyway, creating an image that small could be done pretty quickly on ages old cards, so it's not terribly impressive. Pretty sure they just picked a seed that looked good and just generated it. Definition of NPU says it's specifically a machine learning accelerator. As in, it's 100% focused on that. So... where's the CPU? What exactly is this article saying? That they're deciding to try to get in on the AI bandwagon...? Okay.
Click to expand...
Yeah but that was a 15-65w laptop APU.
Sure you can do all that now with a good 600w plus on a desktop easily enough, but that thing was an ultraportable.
 
Haven't they heard.... all the lakes are drying up, global warmin & all that...:shame:
 
dogDAbone said:
Haven't they heard.... all the lakes are drying up, global warmin & all that...:shame:
Click to expand...
Well if they are naming their GPUs after fantasy jobs, then why not name their CPUs after mythical landmarks?

The first stacked cache consumer parts, Daltigoth, followed by Losarcum, and Istar.

Why not.
 
1_rick said:
Lunar Lake is going to be another low-power laptop design, the successor to Meteor Lake: https://www.techspot.com/news/100216-intel-confirms-arrow-lake-lunar-lake-panther-lake.html Arrow Lake is next year's desktop platform. Panther Lake will be back to a unified desktop and laptop line.
Click to expand...
Yeah, but what I'm confused about is "The demo focused on the generative AI capabilities of Intel's third generation NPU". An NPU is specifically a machine learning accelerator, afaik. It's not the CPU portion. I'm not aware of how recent gen intel chips are, but do they include the NPU bundled in with the CPU or somehow intertwined? Is the NPU supposed to provide some sort of functionality to the CPU to improve its performance? I guess this article is mostly about Intel's advances in the AI space specifically, moreso than any talk about its next processor as a CPU unit...

Mobile (ie your phone) processor units are already using NPU's for doing things like facial recognition.
https://semiconductor.samsung.com/s...t-npu-a-brainy-next-generation-semiconductor/

Lakados said:
Yeah but that was a 15-65w laptop APU.
Sure you can do all that now with a good 600w plus on a desktop easily enough, but that thing was an ultraportable.
Click to expand...

I'm not talking about a 600w desktop. I mean my 2080 (with its tdp capped at 180W) with an 11400F, probably both consuming much less than 230W, could easily generate an image that small pretty quickly. I kind of know because it's doing that every day and even as I type this, in another room in my house. The 2080 is pretty old at this point. I wish this article would tell us how fast it generated it and how large it actually is.
 
StoleMyOwnCar said:
Yeah, but what I'm confused about is "The demo focused on the generative AI capabilities of Intel's third generation NPU". An NPU is specifically a machine learning accelerator, afaik. It's not the CPU portion. I'm not aware of how recent gen intel chips are, but do they include the NPU bundled in with the CPU or somehow intertwined? Is the NPU supposed to provide some sort of functionality to the CPU to improve its performance? I guess this article is mostly about Intel's advances in the AI space specifically, moreso than any talk about its next processor as a CPU unit...

Mobile (ie your phone) processor units are already using NPU's for doing things like facial recognition.
https://semiconductor.samsung.com/s...t-npu-a-brainy-next-generation-semiconductor/



I'm not talking about a 600w desktop. I mean my 2080 (with its tdp capped at 180W) with an 11400F, probably both consuming much less than 230W, could easily generate an image that small pretty quickly. I kind of know because it's doing that every day and even as I type this, in another room in my house. The 2080 is pretty old at this point. I wish this article would tell us how fast it generated it and how large it actually is.
Click to expand...
They did it live on stage, so it couldn't have been that long.
Trying to find a video because these things are usually broadcast somewhere.
 
So they have a working 1.8nm foundry? I'm not up to date on Intels' foundries and their capabilities.
 
GoodBoy said:
So they have a working 1.8nm foundry? I'm not up to date on Intels' foundries and their capabilities.
Click to expand...
They have it working in their test foundry, which is a single machine in a building smaller than most Costco’s. They use that place for some 5 or 6 different nodes at once.

The 18A and 20A foundries are still being built and ASML is still assembling the equipment so it can be shipped for delivery.
 
Lakados said:
They have it working in their test foundry, which is a single machine in a building smaller than most Costco’s. They use that place for some 5 or 6 different nodes at once.

The 18A and 20A foundries are still being built and ASML is still assembling the equipment so it can be shipped for delivery.
Click to expand...
What are the different tiles in Lunar Lake & which tile is from which foundry ?
 
Marees said:
What are the different tiles in Lunar Lake & which tile is from which foundry ?
Click to expand...
With Lunar Lake, Intel Foundry Services (IFS) is expected to debut the company Intel 18A foundry node, which offers transistor densities and power/thermal characteristics comparable to 2 nm-class nodes by TSMC. From the looks of it, the entire CPU tile will be built on the TSMC N3B (3 nm) foundry node

https://www.techpowerup.com/315941/...c-with-on-package-lpddr5x-memory-detailed?amp
 
Lakados said:
With Lunar Lake, Intel Foundry Services (IFS) is expected to debut the company Intel 18A foundry node, which offers transistor densities and power/thermal characteristics comparable to 2 nm-class nodes by TSMC. From the looks of it, the entire CPU tile will be built on the TSMC N3B (3 nm) foundry node

https://www.techpowerup.com/315941/...c-with-on-package-lpddr5x-memory-detailed?amp
Click to expand...
Will the p-cores & e-cores be on same tile or different tiles

Will the gpu be on a separate tile or the same tile as p-cores ??
 
Marees said:
Will the p-cores & e-cores be on same tile or different tiles

Will the gpu be on a separate tile or the same tile as p-cores ??
Click to expand...
There are 2 Tiles, one doing all the compute, the other doing all the IO.

The CPU tile contains the compute complex along the high bandwidth North Fabric. The key components here are the Performance Compute cores, the Low-Power core clusters; the next-generation NPU which accelerates AI, the iGPU based on the Xe2 "Battlemage" graphics architecture, and the LPDDR5X memory controllers. There are other minor bandwidth-hungry components, such as the IPU (image processing), and media engine (video accelerators). The SoC die is now back to being a glorified PCH, with the various platform interfaces, PCIe, USB, and Thunderbolt I/O.
 
You must log in or register to reply here.
Back
Top