• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Intel's 12th Gen 12900K i9 beat AMDs Ryzen 9 5950X. At 2.3x the power.

Intel announced its Alder Lake-S desktop processors the other day, with benchmarks claiming a 30% increase in gaming performance compared to the fastest AMD CPU. The average delta between the Core i9-12900K and the Ryzen 9 5950X varies between 11-16%, however. And on top of that, there’s a caveat. Something Intel didn’t mention in its webcast.

The Ryzen 9 5950X was tested with the TDP (PL1) set to 105W, while the Core i9-12900K had its PL1 set to 241W. That’s 2.3x more for the latter, making it a very unfair comparison. This means that with the TDPs equalized the performance deltas between the 5950X and the 12900K will be in single-digit figures. Launching a “revolutionary” architecture that’s only 5-10% faster than competitor products which have been around for over a year doesn’t sound very hot, does it?

Keep in mind that this is the PL1 (stock) TDP and not the boost TDP we’re talking about here. It’s unclear what the boost PL2 power for the 12900K is like with the PL1 set to 241W, but the 5950X usually settles around 150-170W.

Fortunately, you can take the content creation benchmarks seriously as both the Alder Lake and Rocket Lake parts had their PL1 values set to PL2. Ironically, Intel didn’t compare its new processors against the Ryzen 5000 parts in the content creation segment which itself says a lot. This means that while Alder Lake will be faster (~10%) in gaming workloads, it’s likely going to be on par with the Ryzen 5000 parts in content creation. The former should be offset with the launch of Zen 3D in H1 2022.

Finally, the last tidbit that we need to understand is that these benchmarks were taken before the Windows 11 fix for the Ryzen processors landed. The L3 cache fix improves the lows in most games, while the firmware patch advances the single-threaded performance by letting the OS choose the fastest available core (correctly).

All this means that the third-party benchmarks from reviews may be a fair bit different from the 1st party benchmarks shared by Intel yesterday. Either way, we recommend against pre-ordering any Alder Lake part before the third-party reviews are released.


The source is Intel's own material.

 

manfestival

Member
Wasn't this posted the other day? Also Intel admitted that they initially did their testing during the Cache bug that hurt AMDs performance in w11.
 

Silver Wattle

Gold Member
I don't see the appeal of these until the next node reduction to get the power and thermals to more manageable levels.

Zen3D is soon and it's going to make anyone that bought into this series regret there impulse buy.
 
Title of the thread should have included:

Intel-sanctioned benchmarks that don’t take into account the Win 11 patch that fixed Ryzen cache latency.
I’m not the intel fanboy, but surely this graph takes into account the windows 11 fix?

Anyways, ryzen 3 with v cache is going to be at least on par, probably better than alder lake at way less watts.
 

01011001

Member
Intel announced its Alder Lake-S desktop processors the other day, with benchmarks claiming a 30% increase in gaming performance compared to the fastest AMD CPU. The average delta between the Core i9-12900K and the Ryzen 9 5950X varies between 11-16%, however. And on top of that, there’s a caveat. Something Intel didn’t mention in its webcast.


look at the RAM setup of their test rigs... not representative IMO... running a Ryzen CPU paired with 3200mhz DDR4 vs a new chip that runs on DDR5 is a bit weird to say the least.

they didn't even try to get high performance DDR4 ram for these tests, and DDR4 can run above 4000mhz
 

Chiggs

Member
I’m not the intel fanboy, but surely this graph takes into account the windows 11 fix?

Anyways, ryzen 3 with v cache is going to be at least on par, probably better than alder lake at way less watts.

No, actually, it doesn't. And that''s called out in the article, too.

Finally, the last tidbit that we need to understand is that these benchmarks were taken before the Windows 11 fix for the Ryzen processors landed. The L3 cache fix improves the lows in most games, while the firmware patch advances the single-threaded performance by letting the OS choose the fastest available core (correctly).
 
Last edited:

Kenpachii

Gold Member
240 watts! How hot is that thing getting at compared to AMD? Also good luck trying to cool that thing down in a tiny case for a smaller pc build.

A lot more hot. I am curious to see how this thing will fare on normal air coolers.
 

Bboy AJ

My dog was murdered by a 3.5mm audio port and I will not rest until the standard is dead
My Ncase M1 is happy to house my 5900X.
 

dave_d

Member
At the very least hopefully Alderlake isn't a paper launch like last years Ryzen was. (Yes, that's why I have a 3800xt and not a 5600x or 5800x which probably would have been a better cpus for me.)
 

Dream-Knife

Member
At the very least hopefully Alderlake isn't a paper launch like last years Ryzen was. (Yes, that's why I have a 3800xt and not a 5600x or 5800x which probably would have been a better cpus for me.)
Zen 3 has been easy to find since February.
 

Black_Stride

do not tempt fate do not contrain Wonder Woman's thighs do not do not
Just a few hours to go.
Please have supply......ohh pretty please.
And motherboards better not be that X570 luck of the draw bullshit at launch.
 

Azurro

Member
Ah, so Intel either misled, exaggerated or downright lied about the performance of their new CPUs? Glad to see some things in life don't change. ;)
 

MikeM

Member
I'm unimpressed. Wayyy too much power for such little gains over AMD. If building today with high end CPUs in mind, i'd still buy the AMD chip.
 

Draugoth

Gold Member
Very unimpressive performance. Releasing CPUs every year shows that most of these are paper launches
 

//DEVIL//

Member
24h to go right?

I can't wait to see what these chips are capable of and how they compare to the current AMD offerings.
Same here . I have a 5600x and it’s really an amazing cpu . I am really tempted to get the i7 12th gen but ... looking at the bundles in Canada computers .. just for the motherboard cpu and ram are like 1000$ US.

honestly that is allot of money and I quite frankly do not think it’s worth this much
 
it’s almost like you have no clue how to look at a graph showing performance at different power/performance ratios?
Because that graph does not show intel winning the way you think.
?

Figure It Out What GIF by CBC


Everything checks out.

5900X 24 Threads v 12900K 24 Threads

8545 <<<<<<<<<<<<<<<<<< 10180

Intel 12900K up by 19% over same thread count 5900X.

What are you looking at? Power draw? :messenger_grinning_sweat:
 
Last edited:
Top Bottom