Intel Core i9-12900HK performance and power efficiency comparison: Significant gains with potential to put even upcoming AMD offerings at a disadvantage

[_MT_]

I find comparison between 12900HK and 5900X quite interesting. It's to be expected that 8-core Zen3 processors can't keep up with a 6+8 Alder Lake in loads that scale well. A 12-core 5900X is a better match and it's not outrageously off even in terms of TDP. 5900X was tested here in a Eurocom chassis. Not only can the 12900HK outperform 5900X, it manages better efficiency as well. Not by that much (+10 %), but it's still interesting and quite positive. I think 12900HK would make for a very interesting desktop processor. On the other hand, 12900K is not that much more expensive than 5900X where I live (about €60) and you can restrict its boost if you want a more efficient/ cooler/ quieter machine. And if you want to go all out, you can squeeze roughly 5950X performance out of it at the expense of efficiency.

[Vaidyanathan]

think 12900HK would make for a very interesting desktop processor.

That's a very good observation. Desktop variants of these could be really interesting for those not wanting a 12900K.

[ArsLoginName]

Misleading because:

1 - Title : "Significant gains with potential to put even upcoming AMD offerings at a disadvantage" As shown by the *current* thermally constrained 5800U vs thermally unconstrained 12900hk (16 threads vs 20 threads at 30W) , AMD's offerings only lag by <6% from your and Anadtech's scores. So unless there is no gain whatsoever between Rembrandt and Cezanne, Intel's newest and best is just ahead of AMD's last gen. So to say "put upcoming offerings at a disadvantage" is rather disingenuous and misleading.

2 - As others point out, efficiency isn't about 'power' only. It is about Joules which is the integration of power multiplied by time. But regardless, your "CB score/power" efficiency benchmark even shows this 12900hk lags behind Cezanne. But let's examine this further by using Geekbench 5.3 scores as an example. 12900hk scores 1918 pts in single core at let's say 25 W. An iPhone 13 scores 1755 pts in single core at let's say 10 W. Does that mean we all should just put A15 Bionics in our computers and call it a day since they are roughly 2.5x as efficient?

3 - the old 11980hk ge76 raider scored 5189 pts in CB R20 with a PL1 of 91.75 W while the newer 12900hk scores 6849 points with a PL1 of 110 W with 16 vs 20 threads. That's a 32% performance gap but Intel claims the new 'Intel 7' node is 10-15% more efficient than the "10 nm ESF" of the 11th gen processors and the 12900hk is rated to consume 20% more power (110 W PL1 /91.75 W PL1 rating). So those two factors alone lead to a 20%+ performance advantage for the newer generation.

Hope this helps clarify things. Again this site has a reputation for letting scores and values of laptops speak for themselves. The only subjective scores are keyboard and trackpad.

[Imglidinhere]

think 12900HK would make for a very interesting desktop processor.

That's a very good observation. Desktop variants of these could be really interesting for those not wanting a 12900K.

I still don't understand the point of trying to make an "efficient" core paired with a "performance" core. All this does is add complexity to an already complex system. Zero reason to even bother with "efficiency cores" given Zen 3 is already equally efficient.

Not to mention once Zen 3+ comes to the market en masse, Intel's lead in performance will largely be made irrelevant, given power draw is too high for day to day operations. Unless you can somehow dynamically shut off or make the performance cores go to sleep, I just don't see how it's capable of matching AMD's offerings. Battery life matters a lot for mobile devices, and even gaming laptops these days can casually manage 7-8 hours of time and not really lose out on too much perceived performance.