fevgatosTo answer your question about gaming, yes alderlake is insanely more efficient than zen 3, and not just in gaming but in every lightly threaded or mixed workload. You can check igorslab review or phoronix that run some efficiency benchmarks, alderlake wipes the floor with zen 3. The 5950x is only efficient in all core workloads, and thats mainly because of the insane pl2 limit intel decided to put into their cpu. But their actual architecture is very efficient

For example my 12900k @35w does 12600 on cbr23, beating even the m1 pro max in efficiency. The 5950x scores somewhere around HALF.

ratirtYou wont keep your 12900k at 35w where the efficiency is would you?

CrackongObviously you did not read what I said and posted, After a whole month.

What I said was, Error cores are the shortest plank of a barrel
They DO make the barrel larger so holds more water ( more MT performance )
But they also hinder the performance of the taller planks ( The P cores )
Intel could build a barrel holding the same amount of water using taller planks only but they didn't.

Now the barrel have to use up more space (because it is FAT)
and the barrel is unstable (because there are holes between un-matching planks)

If you really want to argue,
Please watch Buildzoid 's analysis about Error-cores limiting ring clock and IMC performance test, also the Performance / Watt test and projections test.
Then we could start over.

CrackongI wasn't talking to you.

From you own words ,
You enjoy pissing into the wind ?

fevgatosIf i actually cared about efficiency, i would. If i pretended to care about efficiency only to defend one cpu company over another like some people, i wouldnt.

If you care about efficiency first and foremost, alderlake is your man. No matter the task, be it gaming rendering encoding or any other professional or daily task, it does more work per watt. Period.


Thats nonsense. Yes you can get more gaming performance with e cores off cause you can clock the cache higher, but you know in what scenario you can notice it? I kid you not, 720p dlss ultra perfromance with a 3090. Thats what i needed to run on cyberpunk to see if the higher cache makes a difference.

Im running a high overclocked cpu (5.3 all core) with 6000c32 manual tuned kits and a 3090 with 550w custom bios. Cant tell the difference with e cores on or off in gaming unless i drop to 32p resolution. For example, in cp the difference was 234 with e cores off and 217 with e cores on. You know what gpu youd need to get 220 fps on full rt ultra cyberpunk for the difference to even matter? I dunno, propably a 6090 or something

fevgatosThats nonsense. Yes you can get more gaming performance with e cores off cause you can clock the cache higher, but you know in what scenario you can notice it? I kid you not, 720p dlss ultra perfromance with a 3090. Thats what i needed to run on cyberpunk to see if the higher cache makes a difference.

Im running a high overclocked cpu (5.3 all core) with 6000c32 manual tuned kits and a 3090 with 550w custom bios. Cant tell the difference with e cores on or off in gaming unless i drop to 32p resolution. For example, in cp the difference was 234 with e cores off and 217 with e cores on. You know what gpu youd need to get 220 fps on full rt ultra cyberpunk for the difference to even matter? I dunno, propably a 6090 or something

TiggerSeems it's you that cannot read or remember, i said "so imo you are pissing into the wind"

CrackongThen you just agreed to what I said, about the e-cores being the shortest plank of the barrel.

And
If you are so gaming focus, then you don't need e-cores in the first place.
The e-cores are there for bumping up the MT benchmark which, gaming doesn't need and, the presents of e-cores are hindering your cache performance, what you've mentioned.
The 12900k would be a much better gaming CPU if the e-cores are replaced with 2 big fat P-cores.

For MT workload,
Who wants a Hybrid CPU for MT workloads?
Myself doing rendering and heavy VM applications, I sure don't want a Hybrid CPU.
Who wants their money making workload getting slow down randomly just because the CPU itself decides to thrown the thread into subpar e-cores ?

Please consider the fact that Intel themselves don't put hybrid CPU into their mission critical MT workload Sapphire Rapids lineup, only pure P-core design there.



That's normal
Who pays attention to your troll reply.

CrackongThen you just agreed to what I said, about the e-cores being the shortest plank of the barrel.

And
If you are so gaming focus, then you don't need e-cores in the first place.
The e-cores are there for bumping up the MT benchmark which, gaming doesn't need and, the presents of e-cores are hindering your cache performance, what you've mentioned.
The 12900k would be a much better gaming CPU if the e-cores are replaced with 2 big fat P-cores.

For MT workload,
Who wants a Hybrid CPU for MT workloads?
Myself doing rendering and heavy VM applications, I sure don't want a Hybrid CPU.
Who wants their money making workload getting slow down randomly just because the CPU itself decides to thrown the thread into subpar e-cores ?

Please consider the fact that Intel themselves don't put hybrid CPU into their mission critical MT workload Sapphire Rapids lineup, only pure P-core design there.



That's normal
Who pays attention to your troll reply.

fevgatosIf i actually cared about efficiency, i would. If i pretended to care about efficiency only to defend one cpu company over another like some people, i wouldnt.

If you care about efficiency first and foremost, alderlake is your man. No matter the task, be it gaming rendering encoding or any other professional or daily task, it does more work per watt. Period.

ratirtI'm not defending any company because like the companies, defending them wont bring me anything from it. Hope you do. I disagree with INSANELY efficient because counterparts does not fare off by much in that regard. So if there is a case in which it is efficient like 35w, it is not for me so thanks but that's not my man.

fevgatosGreat, so you dont care about efficiency. Neither do I, I only care about perfromance thats why i bought the 12900k

ratirtYes I kinda dont. I bought two Threadripper for performance and yet you may disagree they are efficient in the workload i throw at them and pretty fast. To be clear, I dont play on them. I don't understand why gaming is the new efficiency test around here either. Because that is what you where talking about right? Lightly threaded and gaming. Try loading it up to the brim.

fevgatosNope, was talking about everything. There is not a single task on planet earth that a 12900k with appropriate power limits isnt the most efficient cpu on planet earth (talking about mainstream cpus, the TRs might be better, have no idea honestly).

A 12900k @35w for example scores 12600 on cbr23. Thats 360 point per watt, its untouchable.

ratirtWhat do you mean appropriate? I understand that as stock. it does show some efficiency here and there. OC it and it goes out.
35w again. It is a desktop processor. This metric is useless for a desktop computer you know. It is like arguing which one is more efficient at idle.

fevgatosWhy is 35w useless if you care about efficiency? Lets say it scores 25k cbr23 at 35w and 27k at 240w, would that be useless? I argue it wouldnt i would be running it at 35w all day long.

Again, if you care about efficiency then the most efficient cpu is the 12900k.

ratirtI'm sorry. If i wanted an efficient chip i would have gone for a laptop based computer. I went with a desktop for a reason. So I'd rather use stock setting and measure efficiency/performance not necessarily focusing on gaming and lightly threaded apps as a metric with a 35w cap on it.
You are absolutely missing the point here. You cant say F1 is efficient in fuel consumption when you drive 20Mph would you?
Just because it is efficient in 35w it does not mean it still is when the power goes to 125w which is the ballpark of this processor. And so that Insanely efficient is incorrect and that is what bothers me with your statement because that's stretching the truth about the CPU and what it's supposed to be for.

Sabishii HitoI want to see some BIOS screen shots from the people that got their 12900KS chips early from BLT.

ThrashZoneHi,
Doubt bios are any different from 12900k or even 11900k.
Online manuals are readily available.

Sabishii HitoI mean really from an Asus ROG BIOS that shows SP rating.

ThrashZoneHi,
Okay yep asus is the only manufacture that does prediction evaluation that I know of.

fevgatosAgain, i think you are the one missing my point. If your first concern IS efficiency then yes, youd care about maximizing score per watt. And the 12900k is the best cpu at that on planet earth. If you don't care at all or dont care that much about efficiency then why are we even having this discussion

ratirtYeah at 35w power sure but in the MT apparently it is not that efficient as you say unfortunately. The Ecores help with efficiency in MT or like you say light threads and gaming but if you have some heavy MT load for a noticeable time, the efficiency goes away. Efficiency is not just power draw you know.
We have this discussion since you said INSANELY efficient remember? I disagree with that statement and you have not convinced me with your 35w story.

fevgatosWhat so you mean in the mt it's not that efficient? It scores 12600 on cinebench r23 at 35 watts. That's multithreaded, and its the most efficient cpu on planet earth

ratirtOk. 35w for a 12600 and 12900k it is. Perhaps Intel should have locked those to 35w why even bother with 125 and up.

ratirtOk. 35w for a 12600 and 12900k it is. Perhaps Intel should have locked those to 35w why even bother with 125 and up.