• 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 765 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I wouldn't read too much into clock speeds for an engineering sample. I recall a few years back when people were worried about low clock speeds on a Zen engineering sample. Clock speeds were fine with shipping silicon.
 
If Intel really wanted to add some "pop" to Raptor Lake with the refresh they could remove AVX512 from the P cores. They might be able to add another pair of P's or another E cluster maybe?
 
If Intel really wanted to add some "pop" to Raptor Lake with the refresh they could remove AVX512 from the P cores. They might be able to add another pair of P's or another E cluster maybe?

Physically? Not without a layout change.

BS Buster is a second tier leaker at the best in line with RGT, if he had any legitimate leaks before.
 
if ES2 isn't available yet, I expect MTL to be pretty delayed then, at the very least not launching in 2023.
But ye tech twitter has a problem with taking stuff out of context I swear.

There's a story on semianalysis (grains of salt perhaps required) saying almost that: that Meteor Lake is now on target for a Week 52 2023 launch.
 
It's not looking good for Intel...

1. Rialto Bridge is discontinued
2. The next Max Series XPU/GPU will be Falcon Shores in 2025 (Initially GPU only)

Probably not a lot of customer interest in Ponte Vechio of they're cancelling its successor. Hopefully 2026 sees Intel regain their footing as they are falling woefully behind in a very lucrative market. One which they used to completely dominate.
 
It's not looking good for Intel...

1. Rialto Bridge is discontinued
2. The next Max Series XPU/GPU will be Falcon Shores in 2025 (Initially GPU only)
One hand, Intel cancelling/delaying products is pretty shitty.
On the other hand, Intel cancelling products years before they launch to face reality rather than indefinitely delaying them or every 3 months right before it's supposed to launch is a good thing.
As a consumer, IDRC about this. And I don't nearly know as much about XPUs and the server space to know how popular this product could have been. But I don't doubt this will hurt Intel since I'm assuming server products are always pretty high margin and high profit segments.
 
Called it.
I think there's basically no chance we'll see Falcon Shores before 2026
But to answer your actual question, I consider this neutral to slightly negative for Falcon Shores, and remain firm in my belief that we won't see it till 2026-27.
Intel's GPU division is a mess. They've already delayed Falcon Shores at least two years, and it still has another 3+ to go. They haven't even taped out yet! They're not going to be competitive anytime soon, so let's see how long the board is willing to dump money into graphics with Intel's claims of cost sayings.

Also means we aren't going to see a Granite Rapids HBM.
 
It's not looking good for Intel...

1. Rialto Bridge is discontinued
2. The next Max Series XPU/GPU will be Falcon Shores in 2025 (Initially GPU only)
Considering the counterpart by Nvidia and AMD will include a CPU, makes Falcon Shores initially being GPU only an even bigger deal and worth a separate bullet point. Hard to imagine that Falcon Shores manages similarly efficient CPU - GPU communication and shared memory with an approach like this.
 
Considering the counterpart by Nvidia and AMD will include a CPU, makes Falcon Shores initially being GPU only an even bigger deal and worth a separate bullet point. Hard to imagine that Falcon Shores manages similarly efficient CPU - GPU communication and shared memory with an approach like this.
Yeah, without the CPU component, it's little more than an MCM GPU (maybe also with DDR support?). That's not going to be anything special in 2025.
 
Wait, so this lego thing is not happening? Could it be because of the complexity of interconnecting all these tiles?

Rialto-Bridge_car_678x452.png



BTW not having everything under one hood beside that interconnection losses brings more flexibility how the system can be configured...
 
Last edited:
Looks like it indeed. This all better be because they want to "do it right", not due to even more delays.
It's surely mostly the latter, plus a heavy dose of cost cutting. They've been laying off quite a few people from the GPU/HPC teams. Probably will be more to follow with these roadmap changes.

I don't think Intel's GPU efforts are dead, but all this talk of aligning with customers etc is clearly PR nonsense. They've been missing their targets by a mile, and are bleeding money at a time when they've now promised investors to save ~$10billion. They're going to cut anything unprofitable in the short term to the bone.
 
They're going to cut anything unprofitable in the short term to the bone.
Great, so we are essentially back at that point in the cycle that Intel is optimizing its business for short term effects again, ignoring the goal of having sustainably competitive products in the future. Only difference between now and ~5 years ago is that back then they were optimizing for highest possible short term profit while buying back stocks, now they are optimizing for lowest possible short term losses while not completely ditching dividends. I guess some things never really change.
 
Back
Top