Intel i9-11980HK claims 20%+ FPS advantage over AMD, uses different TGP GPU

[schwellmo92]

Summary

Intel just wrapped up their COMPUTEX opening keynote, in which they showed a single benchmark which was Crysis Remastered comparing the new 10nm Intel Core i9-11980HK versus the AMD Ryzen 9 5900HX. In this benchmark Intel are claiming a win of 117.62 FPS versus 95.86 FPS for AMD which represents a 22% higher average FPS. The laptop specifications are located at https://intel.com/computex21 (link only works on mobile for me)

 

Intel used the MSI GP76 Leopard which is a 17.3" 2.9kg laptop with a 4.5 hour battery life to represent the 11980HK and a Asus ROG Strix G533QS which is a 15.6" 2.3kg laptop with 8 hour battery life to represent the 5900HX. Other than that the MSI laptop also has a 155w TGP whereas the Asus laptop has a 130w TGP. Intel also made no mention of performance modes on either laptop, which do directly affect power limits in some cases (for example the Asus laptop is only 130w in 'Turbo' mode).

  

Quotes

Quote

 

 

My thoughts

There is no doubt in my mind that the new Intel chips are likely to best their AMD competition but 22%~ seems unlikely. Given Intel only displayed a single benchmark this must be the benchmark that shows the greatest difference in performance between the two competitors offerings. Obviously some of that performance delta is also down to the different GPU power limits. Intel has also not documented the power profile used on the laptops, the Asus only runs at 130w in 'Turbo' mode so could be running as low as 115w, I was not able to find if the MSI had similar functionality or if it ran always at 155w.

 

Sources

 

[The_Mafia69]

Seems a bit off is it another principled technology like  scandal

[PeachGr]

I don't find laptop processors interesting to be fair. They depend on power, cooling, cleaning and many more. Big hopes for desktop processor innovation at the same pace

[LAwLz]

Sigh... Can we please stop with the witch hunting? 

"marketing material seems to show best case scenario for company who put it out" is kind of a boring thread to have every week. 

 

As for the performance claims, 22% seems to be on the high side but certainly not unreasonable. The 11980HK is very fast. I am not sure about the power characteristics of that AMD chip but my guess is that it typically uses a lot less power than the Intel chip. That might be completely wrong but I doubt it. 

Power consumption usually don't matter in desktop replacement laptops though so... 

Who knows. Maybe Intel actually has a lead in the "desktop replacement laptop" segment of the market. Too bad it's a segment barely anyone cares about. 

[RejZoR]

It was never quite a question of Intel delivering. It was more of "at what cost". And the "what cost" is still rather ridiculous thermal and power consumption metrics. And those seem to still apply. It's not problem make something stupid fast and just clock to stupid heights. Ryzen CPU's are still the balance sweet spot. They may not be "the fastest gaming CPU's anymore", but they win on all fronts. Intel's just on performance front. Maybe that's less of an issue on desktops where you can just use larger AiO or stick more powerful PSU in the case, but you don't have such luxury with laptops.

[schwellmo92]

I had a read of the Intel specifications, the laptop representing Intel did have a higher TGP so have updated the topic.

[StDragon]

44 minutes ago, RejZoR said:

It was never quite a question of Intel delivering. It was more of "at what cost". And the "what cost" is still rather ridiculous thermal and power consumption metrics.

That a custom or AIO water loop just became the Intel standard requirement?

[samcool55]

The run are 17-30 seconds.

Doesn't that mean it can do the whole benchmark while being in the turbo timeslot where it won't stay during actual gaming?

 

I'm gonna wait until guys like GN review these things because something seems off...

[leadeater]

2 hours ago, LAwLz said:

As for the performance claims, 22% seems to be on the high side but certainly not unreasonable.

Sure it's not unreasonable difference when the GPUs are of different power calibers.

 

Intel: *Proves more powerful GPU renders game at higher frame rate*

Everyone: *slow clap* "Duh, are you stupid or do you think we are"

 

Note: To be fair to your comment, it was made before topic was updated with the extra GPU power information

[Blademaster91]

I think that "22% faster" claim is pointless if its only in 1 benchmark and Intel has a higher TGP. Nice try Intel lol.

[crystal6tak]

Reminds me of the intel partner connect slides that occurred nearly a year ago where they pitted a 2060 mobile against a 2060 MaxQ. That one was even more painfully obvious.

 

Why is intel even pulling this BS again? Isn't tiger lake genuinely better compared to ryzen in per core performance? Come on intel...

[TheReal1980]

Why doesn't popular PC games have a built in "FPS Tester" where it runs the game in different scenarios and give you the results in the end so we have some sort of standard to go by?

Having people manually test the FPS in different places in a game is almost useless.

[schwellmo92]

4 hours ago, TheReal1980 said:

Why doesn't popular PC games have a built in "FPS Tester" where it runs the game in different scenarios and give you the results in the end so we have some sort of standard to go by?

Having people manually test the FPS in different places in a game is almost useless.

They do exist in many games and that was what was used in this instance.

[leadeater]

6 hours ago, crystal6tak said:

Why is intel even pulling this BS again? Isn't tiger lake genuinely better compared to ryzen in per core performance? Come on intel...

It is, real independent benchmarks exists for the Tiger Lake H series CPUs and they are actually in terms of raw performance for gaming faster. Other lightly threaded applications can be faster too. The only real down side is the higher power draw and thus heat (and less battery run time) so H series in thin and light laptops will be slower than Ryzen Mobile, so buyer has to be very careful to evaluate the cooling performance before they choose to go with Intel H series.

[Radium_Angel]

6 hours ago, crystal6tak said:

Come on intel...

I say the following as a fan of Intel CPUs:

 

Intel seems desperate and is trying anything to buy enough time for their dev group to pull out a win no matter what the cost.

[CTR640]

I can't take Intel anymore seriously, they have become nothing but a bunch of clowns trying to make balloon animals that keeps bursting.

[CephDigital]

That 20% difference in performance can also be due to optimus. Asus are optimus only while I bet that MSI was running the GPU in discrete mode. Videos showing the difference in performance are below and they can be MASSIVE.

 

[IkeaGnome]

How were their tests set up? 22% difference makes sense..

When the 11980hk is using 20% more memory...

Is there something I'm missing on the memory usage?

[tishous]

It's very hard to do an apples to apples comparison on laptop CPUs because it heavily depends on the power draw of both the CPU and GPU, cooling, and boosting differences of the CPU and GPU. This is definitely a cherry picked example of heavily cherry picked examples, and I highly doubt it will be anywhere close to 20% faster than Ryzen, but more than likely at least a few percent faster.

 

That being said, Intel regaining the gaming performance crown in laptop CPUs gives me hope for their next generation of CPUs which are hopefully on 10nm as well. With DDR5 likely on their next generation too, I can imagine the performance uplift will be quite significant for Intel and for AMD if they want to stay competitive.

[TheReal1980]

5 hours ago, schwellmo92 said:

They do exist in many games and that was what was used in this instance.

Oh ok, thanks.

[TetraSky]

Intel be like : Our newest product is better than AMD's by this much!! It's also consuming way more power and we got half the battery life compared to them. As such, we're still better at gaming! Hell yeah! Go us!

 

As always, wait for independent reviews... Intel lying/misrepresenting their numbers to look better, is nothing new.

[LAwLz]

3 hours ago, TetraSky said:

Intel lying/misrepresenting their numbers to look better, is nothing new.

I'd like to add that ALL companies do this, including AMD and Nvidia.

Never trust first party benchmarks. 

 

Remember when Nvidia showed the 30 series and were really pushing the 4K and 8K results because they saw a much bigger increase than lower resolutions?

Remember when AMD were comparing Zen 1 to Intel and in their demo they were pretty much just looking up into the sky on the AMD demo, and the screen was full of explosions on the Intel system?

 

First party benchmarks are almost always the best case scenario for whatever picture the company wants to paint. Always wait for third party ones.

We already got benchmarks for the 11980HK and it's pretty good, for a very niche use case (desktop replacement laptops where performance is a much higher priority than power consumption).

[Sauron]

18 hours ago, schwellmo92 said:

laptop

ok, so all of this is completely meaningless because we don't know power management parameters

[lostcattears]

Ok so... Fluff until reviewers do the review