N3B Lion Cove in LNL vs 18A Cougar Cove(LNC+)

[siliconbruh999]

One of the Intel PTL slides comparing the nearly same uARch on N3B and 18A as close as we can get

 

[Xebec]

Interesting the "est" at the bottom - these are estimates for both performance and power, and not actually measured.

The benchmark itself *appears* to be single threaded, which I actually take as a good sign here - it's easy to show higher performance with more cores. But per-core performance is what matters most to users at this point.

 

[MKWVentures]

View attachment 3724
One of the Intel PTL slides comparing the nearly same uARch on N3B and 18A as close as we can get

so 18a is 10% higher performance than N3B at given power. Is that correct?

 

[Xebec]

so 18a is 10% higher performance than N3B at given power. Is that correct?

Estimated performance * estimated power * estimated architecture differences = ??? .

 

[blueone]

Estimated performance * estimated power * estimated architecture differences = ??? .

I agree. That chart looks really suspicious. I think that would have made even Gelsinger blush.

 

[tooLongInEDA]

I agree. That chart looks really suspicious. I think that would have made even Gelsinger blush.

Never trust a slide with unmarked axes. Instant lab fail when I was at school and university. Also, why do they use the wording "at similar performance/power" ? The normal wording would be "higher performance at the same power".

 

[siliconbruh999]

so 18a is 10% higher performance than N3B at given power. Is that correct?

Yes or 40% lower power at the same performance

 

[sonchoy]

N3B is FinFET. 18A is GAA. Better gate control and steeper subthreshold swing are expected in GAA devices over FinFETs. This means either "more performance at iso-power" or "lower power at iso-performance." The Intel chart is consistent with this expectation.

 

[eding42]

Some tech commenters on Twitter noted that the Cougar Cover P cores looked smaller than Lion Cove on N3B. Eagerly awaiting die shots! Let's see the REAL density of 18a.

 

[MKWVentures]

Never trust a slide with unmarked axes. Instant lab fail when I was at school and university. Also, why do they use the wording "at similar performance/power" ? The normal wording would be "higher performance at the same power".

Lawyers..... companies have lawyers check these items now LOL