• 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 639 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
That is an insane amount of money for a company that is behind in so many areas, server most notably, the most lucrative part of a CPU company.

Well, Intel had to lure him away from VMware, not surprised he was promised with a handsome compensation package.

However, investors were also reported to be unhappy about Gelsinger's plans to revitalize the business with a significant expansion in manufacturing facilities that will impact the Intel's bottom line for several years to come.

Good. They shouldn’t have listened those short term profit chasers since 2013.
 
Has Intel published any first Party Benches on Sapphire Rapids non-HBM? All I see are numbers for the Sapphire Rapids HBM with apps that can take advantage of it. But general FP/INT benchmarks are lacking
 
Compared to a typical 12900K similar or little higher 1C-8C performance , MT over all cores Raptor Lake-S is more like 25% faster (3900 vs 3100 points)

edit:
Actually a typical 12900K is even slower
Thanks for the edit. The 25% faster MT was comparing this Raptor Lake chip to an overclocked 12900K. Stock conditions were closer to 31% faster in MT which is roughly what you'd expect given 33% more threads (note of course that not all threads are equal in computational performance so you can't just go by thread count).

The turbo clocks of 4.6 GHz average seem low. I assume this was an engineering sample.
 
4.6 Ghz is the average clock speed, we don't know the max boost from this sample.
UserBenchmark for that Raptor Lake chip says "CPU throttled at 87% by Windows. Ensure maximum processor state is set to 100% via Settings > System > Power & sleep > Additional power settings > Change plan settings > Change advanced power settings > Processor power management > Maximum processor state."

If we take that as accurate (a big if) then that puts this chip at a 4.6 GHz / 0.87 = 5.3 GHz turbo. Could be that this chip just runs hot and reaches thermal limits in the test.
 
Has Intel published any first Party Benches on Sapphire Rapids non-HBM? All I see are numbers for the Sapphire Rapids HBM with apps that can take advantage of it. But general FP/INT benchmarks are lacking

Interesting that you mention that.

If we take that as accurate (a big if) then that puts this chip at a 4.6 GHz / 0.87 = 5.3 GHz turbo.

I think we'll see higher tvb values out of Raptor Lake-S in the highest-end SKU. This has got to be an early chip.
 
So the reason Meteorlake SoC tile is so large is because not only it has the PCH, but the uncore portions of the current CPU as well. Some are speculating that they are moving the media engine, which is why the GPU is so small at 23mm2..

The PCH is over 50mm2 and the Uncore portion of the CPU is also over 50mm2. Some will move to the I/O tile, but that's only 10mm2 so majority of 50mm2 uncore is going to the SoC. Now add in the media engine, enhanced functionalities like the Atom low power block and new capabilities for the chipset and it's easy to see why it's close to 100mm2.

As for the GPU, it's likely it's a 96EU variant. We'll see 192EU variant someday but that's not the die that's being shown.
 
Last edited:
6% single core performance at 4.6ghz the is still 25% clock heardroom.. so my prediction is 15% single core performance uplift 😁
I was right 😁 raptor lake 15% faster in single core performance.. Raptor lake will be a gaming beast ✔️
 
Wow. I'm sure that Raptor Lake is great work.
Raptor Lake must be optimized superior for Intel 7 process.
It is insanity mission to implement higher clock & larger cache in same process.

Nowadays, we realize "Coming Back" INTC with their IDM advantage.
I can't wait Intel 4/3_MTL/ARL family.
 
Wow. I'm sure that Raptor Lake is great work.
Raptor Lake must be optimized superior for Intel 7 process.
It is insanity mission to implement higher clock & larger cache in same process.
what you mean like Zen3.........

Nowadays, we realize "Coming Back" INTC with their IDM advantage.
I can't wait Intel 4/3_MTL/ARL family.
Where do all these clowns keep crawling out from.
Intel release 1 performant but a very larger area/power cost Core and all these insane takes start crawling out from everywhere.

If anything Zen4 and Raptor Lake are looking quite equal in a high end mainsteam desktop setting with every other market looking very much in AMD favour. Intel is super lucky that 6X00U is Zen3 on 6nm because otherwise that would be just beyond embarrassment.

/enough amd shilling in the Intel thread, but man do posts like quoted pull my chain.......
 
Intel is super lucky that 6X00U is Zen3 on 6nm because otherwise that would be just beyond embarrassment.


Rembrandt availability would have been even worse, not sure if this is a good thing. AMD is very lucky they can use a much superior 5nm for Zen 4, otherwise things would look rather dull for them. They seem to rely on a full node advantage over Intel with Zen 4 against Raptor Lake.
 
So the reason Meteorlake SoC tile is so large is because not only it has the PCH, but the uncore portions of the current CPU as well. Some are speculating that they are moving the media engine, which is why the GPU is so small at 23mm2..

The PCH is over 50mm2 and the Uncore portion of the CPU is also over 50mm2. Some will move to the I/O tile, but that's only 10mm2 so majority of 50mm2 uncore is going to the SoC. Now add in the media engine, enhanced functionalities like the Atom low power block and new capabilities for the chipset and it's easy to see why it's close to 100mm2.

As for the GPU, it's likely it's a 96EU variant. We'll see 192EU variant someday but that's not the die that's being shown.

Meteor Lake should be interesting to say the least.
 
Rembrandt availability would have been even worse, not sure if this is a good thing. AMD is very lucky they can use a much superior 5nm for Zen 4, otherwise things would look rather dull for them. They seem to rely on a full node advantage over Intel with Zen 4 against Raptor Lake.
Scale wise intel is a way bigger company than amd.. intel is showing that massive RnD budget in alder lake and meteor lake comong fast
 
Wow. I'm sure that Raptor Lake is great work.
Raptor Lake must be optimized superior for Intel 7 process.
It is insanity mission to implement higher clock & larger cache in same process.

Nowadays, we realize "Coming Back" INTC with their IDM advantage.
I can't wait Intel 4/3_MTL/ARL family.
Meteor will be a revolution in cpu+ gpu design.. intelv4 is a huge node advantage over tsmc n4 or n5😁
 
It is HIGHLY Likely that Raptor Lake will use BSM To Improve thermal conductivity(like Zen4), it's being used on Sapphire Rapids and Ponte Vecchio already.

People were surprised to see The Gold/Silver Plating on that Zen4 Prototype, but Intel is also putting some of that Gold on their CPU/GPU that will be releasing at the end of the year(Sapphire Rapids and Ponte Vecchio)
 
Last edited:
It is HIGHLY Likely that Raptor Lake will use BSM To Improve thermal conductivity(like Zen4), it's being used on Sapphire Rapids and Ponte Vecchio already.

People were surprised to see The Gold/Silver Plating on that Zen4 Prototype, but Intel is also putting some of that Gold on their CPU/GPU that will be releasing at the end of the year(Sapphire Rapids and Ponte Vecchio)
Funny how we've come a long way from toothpaste as TIM.
 
Back
Top