Intel powers on its first 14th-gen Meteor Lake CPU prototype

[mongeese]

Something to look forward to: On an earnings call last week, Intel CEO Pat Gelsinger announced that Intel had recently powered on their first prototype of a Meteor Lake processor, a member of the 14th-gen Core series and their first consumer-oriented product to use the Intel 4 node.

"I am happy to share that Intel 7, Intel 4, Intel 3, Intel 20A, and Intel 18A are all on or ahead of the timelines we set out in July. For example, on Intel 4, we said we had taped out our compute tile for Meteor Lake," Gelsinger told investors.

"In this quarter, it came out of the fab and powered up, and within 30 minutes, with outstanding performance, right where we expected it to be. All told, this is one of the best lead product startups we have seen in recent memory, which speaks to the health of the process."

Meteor Lake will be a major advancement in the consumer space. When Intel announced it in July, they said that it will be their first series to separate the compute, I/O, and GPU components into separate chiplets, or as they call them, "tiles."

Intel 7 is a rebrand of the 10nm Enhanced SuperFin node and Intel 4 is a rebrand of the 7nm node.

AMD and Intel are already leveraging chiplet technology inside their CPUs, but in a way that reduces their performance when compared to non-chiplet designs with the same number of cores. Intel wants to reduce that compromise for Meteor Lake by using Foveros, a packaging technique that leverages an active interposer to improve inter-chiplet communication.

Meteor Lake specifically targets the 5-125 W range and will be used for both laptop and desktop processors. It can be integrated with Xe GPUs that have between 96 and 192 EUs, which will be comparable to current entry-level and midrange discrete GPUs.

Before Meteor Lake arrives, however, both Alder Lake and Raptor Lake need to debut. Intel has been forthcoming with details about Alder Lake and has indicated that the series will launch soon, while Raptor Lake is a mystery slated for next year. Meteor Lake will follow in 2023.

Permalink to story.

https://www.techspot.com/news/91906-intel-powers-first-14th-gen-meteor-lake-cpu.html

 

[passwordistaco]

Hot as a meteor, needs a lake to cool it?

 

[captaincranky]

Hot as a meteor, needs a lake to cool it?

Dude, that's rude.

That said, I'm wondering how they'll escalate the code name from here. "Asteroid Crater Lake, maybe? .

 

[envirovore]

Intel's marketing team needs a good punch in the **** over this pants on head stupid naming scheme.

Not that AMD needed the 5, 7, & 9 branding at all before the specific CPU model either.

 

[theruck]

Intel 7 - 10nm
Intel 4 - 7nm node.
of course! so intelish

 

[Sausagemeat]

So basically now that Intel is back on schedule they will dominate for another decade. Potentially until we all ditch X86 and go ARM.

 

[Lionvibez]

So basically now that Intel is back on schedule they will dominate for another decade. Potentially until we all ditch X86 and go ARM.

lol

 

[Lounds]

If they're so ahead, then why is arc going to TSMC? Honestly I don't believe a word they're saying. It's all marketing tosh.

 

[kiwigraeme]

Most of us are not chip designers - well not this type .
Remember articles from a long time ago - about making thin wires , then etching thin connecting conduits

Probably designing a tiny powerhouse chiplet is one of the easiest parts - just a guess.
But doing the communication , optimising placement/chains- keeping thermal constraints and other limits. Ensuring it can be manufactured - is probably the harder part.

There must be a good article about that a willing layman could read on the evolution of chip communication.
At these sizes chips can only have 3 states I think from memory 0 , 1, and something else .
Again output is only those 3 - could be wrong .
I assume a channel can only send 1 stream at a time - unlike say fibre optics, or quantum
I know there's work on tiny variable outputs .

Anyway such an article would be fun to read of history to future

 

[Reehahs]

The only thing ahead of target here seems to be Intel's marketing department.

 

[seeprime]

If they're so ahead, then why is arc going to TSMC? Honestly I don't believe a word they're saying. It's all marketing tosh.

Smoke and mirrors from Intel, until their new Arizona fabs go online. Then we'll know if they've caught up, let alone be ahead. Until then, any wafer that's destined for smaller than 10nm trace chips will be outsourced to the more capable TSMC. Intel margins are so high that the outsource cost won't hurt them.

 

[QuantumPhysics]

My Youtube streaming computer is a 5960x with a 3090 FTW3.

I'm still able to run virtually any game on the market just fine - including Far Cry 6 which I'm playing now.

The only thing that will make me upgrade, possibly to this 14th gen is DCS WORLD, but I definitely need DDR5 on that motherboard.

 

[hahahanoobs]

Smoke and mirrors from Intel, until their new Arizona fabs go online. Then we'll know if they've caught up, let alone be ahead. Until then, any wafer that's destined for smaller than 10nm trace chips will be outsourced to the more capable TSMC. Intel margins are so high that the outsource cost won't hurt them.

56% is a margin far from being "so high". Especially when it's projected to drop to 50% before it goes back up.
By their own admission, they don't expect to be on par with AMD until 2024. That's when they say they'll be back on top.

TSMC isn't chosen because they are "more capable". They simply have the capacity.

Side note (rumour):
Alder Lake is already faster than M1 Max.

 

[hahahanoobs]

Intel 7 - 10nm
Intel 4 - 7nm node.
of course! so intelish

More like a valid adjustment that also counters the noobs that think a smaller process is better without taking into account the actual chip designs and the process technology itself. It's like the Apple vs HTC and the "bigger GB's" from years ago, or higher clock speeds means better CPU since the Pentium 4 days. Times change. Your average consumer doesn't. That's life.

 

[hahahanoobs]

"For the first time in years, they're running on schedule"

But I thought the delay was Intel sandbagging, because nothing was coming out of AMD?
Make up your minds!

 

[Geralt]

So basically now that Intel is back on schedule they will dominate for another decade. Potentially until we all ditch X86 and go ARM.

Eeeer... NO.

 

[Hardware Geek]

So basically now that Intel is back on schedule they will dominate for another decade. Potentially until we all ditch X86 and go ARM.

I'm actually looking forward to the new exynos processors when they come to chromebooks. With the rdna2 gpu, it could potentially be a decent little laptop. I don't know that arm will take the place of x86, but I do think they are finally at a point where they will make a dent in the market. Especially now that Apple has gone arm, the rest of the industry is going to be pushed to offer laptop worthy performance from other arm manufacturers.

 

[Nintenboy01]

I'm cautiously optimistic. I think AMD's resurgence and losing Apple's business did finally light a fire under their butts.

 

[BadThad]

Intel will be behind until they tackle the thermal issues or start marketing their CPU's as processors AND room heaters.

My son walked into my computer room this weekend and ask "why is it so hot in here". I told him 18 Intel CPU cores running 100% load 24/7. LOL

 

[Yenega]

Intel 7 - 10nm
Intel 4 - 7nm node.
of course! so intelish

Intel 10nm Superfin is just as dense as TSMC 7nm, so..

You can't compare nanometer across different fabs.

TSMC 6nm is optimized 7nm, TSMC's answer to + but has nothing to do with 6nm

It has been known for years that nanometer is a dead metric, unless used internally

 

[Nintenboy01]

Intel 10nm Superfin is just as dense as TSMC 7nm, so..

You can't compare nanometer across different fabs.

TSMC 6nm is optimized 7nm, TSMC's answer to + but has nothing to do with 6nm

It has been known for years that nanometer is a dead metric, unless used internally

and the 8nm Samsung uses for Ampere isn't too different from 10nm LPP I think