Intel Core i7-12800H Alder Lake-P Mobile Processors Spotted in Geekbench

[AleksandarK]

Intel's upcoming lineup of mobile processors with the novel hybrid core technology are codenamed Alder Lake-P. Contrary to the desktop Alder Lake-S, the P variant was envisioned with a lower power budget in mind to fit various form factors. Today, we get to see some of the first benchmarks of the Alder Lake-P processors and get to compare them to AMD's competing products. In the Geekbench 5 listing discovered by BechLeaks, Intel's Core i7-12800H processor with six performance and eight efficiency cores appear. The CPU ran at a base frequency of 2.8 GHz, while Geekbench didn't show boosting clocks in the submission.

The CPU managed to score 1654 points in single-core results and 9618 points in multi-core runs. If we compare this to AMD Ryzen 7 5800H, a direct competitor, the CPU is faster by 25% and 35% in single-core and multi-core results, respectively. If the previous Tiger Lake-H generation is a reference, the Alder Lake-P chip manages 12% and 20% higher single-core and multi-core scores. This specific processor is part of the GIGABYTE AORUS 15 YE4 laptop used for the Geekbench 5 benchmark test run.


View at TechPowerUp Main Site

 

[Richards]

This will compete well against Apple's m1 max cpu

 

[qcmadness]

Power consumption is vital in determining the performance of a mobile processor.

 

[ToTTenTranz]

Shouldn't we be a little cautious with these pre-release results?

IIRC the desktop Alder Lake models also got some headlines showing ridiculously high scores on geekbench and cinebench leading up to the release date when the actual results were still very good but a lot more grounded.

This could be e.g. a 12800H mounted on a desktop environment with a 800W PSU and a cooler for 250W dissipation.

 

[ICWiener]

Shouldn't we be a little cautious with these pre-release results?

IIRC the desktop Alder Lake models also got some headlines showing ridiculously high scores on geekbench and cinebench leading up to the release date when the actual results were still very good but a lot more grounded.

This could be e.g. a 12800H mounted on a desktop environment with a 800W PSU and a cooler for 250W dissipation.

The 12900k gets around 2000 in single core and 18,000 in multi. 1600/10,000 for the 12800h suggests a very reasonable TDP and honestly seems a bit low.

 

[Flanker]

Why do we always see Geekbench scores first?

 

[ToTTenTranz]

The 12900k gets around 2000 in single core and 18,000 in multi. 1600/10,000 for the 12800h suggests a very reasonable TDP and honestly seems a bit low.

Intel's H-series have a TDP ranging between 28 and 45W. Alder Lake's desktop 12900K pushes almost 300W during a Cinebench run. If it's getting 55% the performance out of 15% the power consumption then that's a massive win for Intel, and not low at all IMO.

This particular 12800h is probably based on the same 8p + 4e cores as the 12700K, which get ~15000 on Geekbench while pushing ~230W. Again, getting ~66% the performance out of 20% or less of the power consumption would be spectacular.

 

[ICWiener]

Intel's H-series have a TDP ranging between 28 and 45W. Alder Lake's desktop 12900K pushes almost 300W during a Cinebench run. If it's getting 55% the performance out of 15% the power consumption then that's a massive win for Intel, and not low at all IMO.

This particular 12800h is probably based on the same 8p + 4e cores as the 12700K, which get ~15000 on Geekbench while pushing ~230W. Again, getting ~66% the performance out of 20% or less of the power consumption would be spectacular.

Not really, that's just how power scales. Also in reality H series processors can use much more than 45W.

12800H will almost certainly use the 6+8 mobile die. Not sure if it will be 6+8 or 6+4 though. The 11800H already gets 1500/8000, if this is 6+8 then i really don't see how 10,000 is a massive win.

 

[RandallFlagg]

This should not be surprising for those following some of the testing with the 12900K. Short story is, we've already seen a 12900K power limited to 35W and undervolted matches an M1 Pro at performance/watt using Cinebench as a metric.

Since this is almost certainly the 6+8 part vs the 12900K 8+8, my guess is this is at 28W. A 35W part should perform about 15-20% better on multi-core, similar to how the power limited 12900K did.

This really is what the big.LITTLE architecture of Alder Lake is for. The laptop space was always going to be Alder Lakes strong point. The fact it also performs very well on desktop is the bonus.

 

[Richards]

This should not be surprising for those following some of the testing with the 12900K. Short story is, we've already seen a 12900K power limited to 35W and undervolted matches an M1 Pro at performance/watt using Cinebench as a metric.

Since this is almost certainly the 6+8 part vs the 12900K 8+8, my guess is this is at 28W. A 35W part should perform about 15-20% better on multi-core, similar to how the power limited 12900K did.

This really is what the big.LITTLE architecture of Alder Lake is for. The laptop space was always going to be Alder Lakes strong point. The fact it also performs very well on desktop is the bonus.

Raptor lake is gonna scale it up even bigger

 

[ratirt]

Not really, that's just how power scales. Also in reality H series processors can use much more than 45W.

We all know the 12th gen processors can use more power than 45W. Actually they will suck in as much power as you throw at it. The problem is not to exceed the given power limit yet still keep the performance. It is a mobile CPU and power restriction are crucial. You would need a tremendous cooling solution and with a laptop that is very limited. You can go too high with power due to heat but yet you want to keep the performance at a certain level. Balance here is very important not how much power can the CPU use because you wont be able to cool it.