Discussion Intel current and future Lakes & Rapids thread

[Harry_Wild]

Intel’s 12th Gen “Alder Lake” desktop processors will be announced October 27th and released November 4th! Stay tune!

 

[jpiniero]

GPU Benchmark - Ashes of the Singularity

Video Card GPU Benchmarks, High quality testing of your video card and gaming system.

www.ashesofthesingularity.com

AOTS benchmark of the 12900H. That one is 6+8.

 

[eek2121]

Intel Core i9-12900K overclocked to 5.2 GHz on all Performance cores reportedly consumes 330W of power - VideoCardz.com

Supposedly the user achieved an all core 5.2 GHz overclock. I wonder if he disabled Gracemont for the test. I don't see the little cores getting to 5.2 GHz unless Intel was sandbagging.

 

[dullard]

Intel Core i9-12900K overclocked to 5.2 GHz on all Performance cores reportedly consumes 330W of power - VideoCardz.com

Supposedly the user achieved an all core 5.2 GHz overclock. I wonder if he disabled Gracemont for the test. I don't see the little cores getting to 5.2 GHz unless Intel was sandbagging.

It was addressed in your link: "he kept Gracemont/Efficient/Atom/Small cores at stock 3.7 GHz."

 

[Hulk]

Intel Core i9-12900K overclocked to 5.2 GHz on all Performance cores reportedly consumes 330W of power - VideoCardz.com

Supposedly the user achieved an all core 5.2 GHz overclock. I wonder if he disabled Gracemont for the test. I don't see the little cores getting to 5.2 GHz unless Intel was sandbagging.

Report reads the Gracemont cores were kept at stock 3.7GHz. I don't think there is anyway the 12900K could achieve that MT score without the Gracemont cores. It would mean Golden Cove would have better than 75% IPC advantage over Zen 3 (accounting for difference in clocks and core count) in CPUz.

 

[IntelUser2000]

@Hulk I believe it. If you actually go to the Videocardz article, the 12900K by Yukki got 11400 points, while this one is getting 11900. So Either Golden Cove is 70% faster per clock or Gracemont clusters are helping a lot. Gracemont seems to add 35-40% to the multi-threaded cores. Pretty damn decent considering the reputation of Atom and Gracemont's known clock speed and lack of Hyperthreading.

The 5.2GHz overclock itself is only helping it by 5%.

 

[eek2121]

It was addressed in your link: "he kept Gracemont/Efficient/Atom/Small cores at stock 3.7 GHz."

Report reads the Gracemont cores were kept at stock 3.7GHz. I don't think there is anyway the 12900K could achieve that MT score without the Gracemont cores. It would mean Golden Cove would have better than 75% IPC advantage over Zen 3 (accounting for difference in clocks and core count) in CPUz.

Didn’t see that when I initially read it. Maybe it was edited? Regardless, I am curious if Gracemont has any headroom for higher clocks.

 

[Hulk]

@Hulk I believe it. If you actually go to the Videocardz article, the 12900K by Yukki got 11400 points, while this one is getting 11900. So Either Golden Cove is 70% faster per clock or Gracemont clusters are helping a lot. Gracemont seems to add 35-40% to the multi-threaded cores. Pretty damn decent considering the reputation of Atom and Gracemont's known clock speed and lack of Hyperthreading.

The 5.2GHz overclock itself is only helping it by 5%.

You believe the score or you believe it was achieved with Golden Cove only?

 

[IntelUser2000]

You believe the score or you believe it was achieved with Golden Cove only?

I believe it was achieved with the help with the Gracemont cluster.

 

[inf64]

You believe the score or you believe it was achieved with Golden Cove only?

There is 0% chance that it was done only with GC.

 

[Det0x]

Poster says this was done at "XMP".

Seems like this benchmark war for next gen will be fought around 50ms for both Intel and AMD
May the biggest L3 cache win

 

[JoeRambo]

View attachment 51694
Poster says this was done at "XMP".

Seems like this benchmark war for next gen will be fought around 50ms for both Intel and AMD
May the biggest L3 cache win

Thats not bad actually. DDR4 3200CL20 + some 7-10ns. The previuos mad latencies seem to be fixed at gear2 level. Stock tests @ Anandtech JEDEC timings will be horribad, but that was expected anyway.
One has to wonder what are the GEAR2->GEAR4 switching points?

 

[dullard]

View attachment 51694
Poster says this was done at "XMP".

Seems like this benchmark war for next gen will be fought around 50ms for both Intel and AMD
May the biggest L3 cache win

There already is DDR5 memory with both a higher frequency and noticeably lower latency than what that poster used:

https://cdn.videocardz.com/1/2021/10/03-gskill-ddr5-6600-cl36-36-36-16gbx2-screenshot.png

6600 MT/s with CL36 is a 10.9 ns latency for the memory, who knows if the specific implementation would add more to that.

 

[mikk]

https://twitter.com/i/web/status/1451109571544555530

From 4.9-->5.2 Ghz= +100W
From 5.2-->5.3 Ghz=+70W

Voltage too high, Golden Cove scaling is off. At 4.0 Ghz (i5-12400) Golden Cove is probably a very efficient core but not at around 5 Ghz.

 

[dullard]

https://twitter.com/i/web/status/1451109571544555530

From 4.9-->5.2 Ghz= +100W
From 5.2-->5.3 Ghz=+70W

Voltage too high, Golden Cove scaling is off. At 4.0 Ghz (i5-12400) Golden Cove is probably a very efficient core but not at around 5 Ghz.

It isn't supposed to be an efficient core. It is supposed to clock as high as possible for single-threaded and lightly-threaded tasks. That is why we are getting the efficient Goldmont cores. 8 now, 16 next year, who knows how many with Meteor lake in a year and a half, 32 with Arrow Lake. That is also why mobile will have just 6 Golden Cove cores and Ultra mobile will just have 2.

 

[Asterox]

Same story as with Rocket Lake, CPU-s are sold before official launch=still NDA in active.

Intel Core i9-12900K Alder Lake CPUs Being Sold & Shipped To Customers Weeks Before Launch For $610 US

Intel unveils its Alder Lake CPUs including the Core i9-12900K next week but some lucky customers have already got the chip in their hands.

wccftech.com

Interestingly strange, Intel obviously doesn't care about Desktop.The rules are there, hm so that they can be broken at will.MindFactory again obviously not caring about the rules.

 

[John Carmack]

It looks like architecture and/or 10nm++ process is still behind the competitors.

 

[mikk]

It isn't supposed to be an efficient core. It is supposed to clock as high as possible for single-threaded and lightly-threaded tasks. That is why we are getting the efficient Goldmont cores. 8 now, 16 next year, who knows how many with Meteor lake in a year and a half, 32 with Arrow Lake. That is also why mobile will have just 6 Golden Cove cores and Ultra mobile will just have 2.

It's about Golden Cove multithreading clock speed.

 

[Abwx]

https://twitter.com/i/web/status/1451109571544555530

From 4.9-->5.2 Ghz= +100W
From 5.2-->5.3 Ghz=+70W

Voltage too high, Golden Cove scaling is off. At 4.0 Ghz (i5-12400) Golden Cove is probably a very efficient core but not at around 5 Ghz.

Smaller cores are actually quite inefficient at full CPU throughput because they are fed the full voltage required by the big cores running at 5GHz.

A separate voltage rail for the small cores is necessary for full throughput being efficient, currently small cores are overvolted by something like 20% over what would be enough to get good stability at 3.7, that s roughly 40% more power than necessary.

 

[mikk]

It looks like architecture and/or 10nm++ process is still behind the competitors.

Based on what? Based on 5Ghz Golden Cove vs 4.5 Ghz Zen 3? It's flawed.

 

[JoeRambo]

Smaller cores are actually quite inefficient at full CPU throughput because they are fed the full voltage required by the big cores running at 5GHz.

Yeah, i was disappointed to read they share same VCC domain as big cores. Those things really need their own VID tables and separate voltage rail to really shine in OC scenarios. Mobile origins of this CPU are showing up once more.

 

[Det0x]

https://twitter.com/i/web/status/1451109571544555530

From 4.9-->5.2 Ghz= +100W
From 5.2-->5.3 Ghz=+70W

Voltage too high, Golden Cove scaling is off. At 4.0 Ghz (i5-12400) Golden Cove is probably a very efficient core but not at around 5 Ghz.

Other known twitter guys are also posting the same

https://twitter.com/i/web/status/1451105188643672069

 

[mikk]

Getting worse..

https://twitter.com/i/web/status/1451105188643672069

Same as Raichu.

 

[Abwx]

Yeah, i was disappointed to read they share same VCC domain as big cores. Those things really need their own VID tables and separate voltage rail to really shine in OC scenarios. Mobile origins of this CPU are showing up once more.

On a DT SKU this would necessitate a 100W VR just for the small cores, that s too much added cost and complexity for the benefit, and in mobile frequency of the P cores will be low enough as to render dual rails useless, seems to me that AMD s such solution for CPU and iGPU wasnt a success and that the common supply ended being generalised in OEMs designs.

 

[inf64]

For about 10% IPC increase versus Zen3, Golden Cove looks to be very inefficient power wise. I don't know if it's the issue with the process node or uarch. itself, but perf./watt is noticeably worse. Especially if you take into consideration die size and how much intel invested in beefing up the core.