mechtechI'd like a updated 5700G with double the L3 cache and with at least 1024 rdna2 shaders at 95W or less.

That would probably be a 5600XT level GPU!!!!

Dr_b_What is this based on? There are numerous youtubers/articles, one which i already linked, showing how E Cores don't matter in games, did you look at them at all, or do you want to just deny they exist and keep asking for evidence?

I only care how my PC performs, not how someone else's does who may have a completely different setup. You think i am going to disable my E cores because of something i see on the internet? lol Or because of the opinion of some random guy on a forum? well no i am not. My PC runs fine as i said with the E cores enabled. If you think otherwise, as i said twice, disable them on your PC if you even have ADL.

TiggerIt might have been worth getting a 12700 non k and just disabling the E cores. At least you would have an extra 2 P cores. The E cores have no detrimental effect on my PC so imo you are pissing into the wind.

I have no interest to pay for useless error cores.

And by the way
Surely we aren't pissing into the wind,
Because you caught them with your face and reply every single comment we made.

CrackongI have no interest to pay for useless error cores.

And by the way
Surely we aren't pissing into the wind,
Because you caught them with your face and reply every single comment we made.

So the extra 2/4 cores would be no use either then i guess

i await your reply as you also can't help but reply to me :)

Anyway enough discussing when you are just trolling me now. bye /ignore

TiggerAnyway enough discussing when you are just trolling me now. bye /ignore

Funny conclusion from the one who brought the 'piss' word to the discussion in the first place.

I guess you do realize logic isn't on your side.

Press M for Mayday :D

Max(IT)my 12700K with E cores surely is better than those CPUs with no E cores, even in gaming. So I still dont see the point here.
Hybrid architecture is an improvement.


what value you post has, since you didn't link any useful reference ?


absolutely not.
In gaming my 12700K need less power than my 5800X, with better performance.

Needs you mean uses less? Would you mind showing me some evidence about that cause I could not find any direct comparisons. Maybe it depends on the game but I looked over (TPU), found the 12900K and system power usage while gaming was around 570W for the whole system with PL1=125W PL2=125W so I assume 12700K would have used a bit less. 5800x used 364W but with a 2080 TI not 3080 like Intel processor so you can chop 60W off. So this is what I have found so not sure how you have calculated it uses less power while gaming and thus my surprise. If it is less indeed, how much less is it? From my search here on TPU, with taking some logic into perspective, it's hardly possible that the 12700K used much less power than what the 12900K did.

I didnt find any direct comparisons here on TPU about gaming vs power consumption except the 12900K. @W1zzard any reason for that? Gaming comparison is valuable info I'd say.




BTW.
Is Hybrid arch an improvement? It is an improvement but.
Intel did the e cores to have more cores available and that (in my opinion) is due to Intel being unable to get more P cores in there to tackle the core count with AMD counterparts. Kinda marketing scheme to be fair.
Ecores dont have hyperthreading Pcores does but you can fit more ecores since these are smaller. I think 1 pcore is equivalent to 1.5 Ecore in size. Correct me if I'm wrong.
So improvement is there but i can bet it would have been better if you replaced Ecores with Pcores. I think, it would have been better for Intel to put more Pcores instead of Ecores. As you know, Pcores are much faster and have HT but the core count would have been lowered significantly and from the marketing perspective would not look as good with lower core count in comparison to AMD's offerings. The other aspect is power consumption. Ecores use less power than Pcores and thus makes them efficient by a significant margin.
Is it an improvement? You get more cores so in a way it is. Is it the right direction for the improvement? You can argue here if you think strictly about performance.
So when you say your 12700K is better than CPUs without ecores sure you are right but if you had the choice to have a 12700K without Ecores but instead you would get more Pcores, which one would you pick? 12700K as it is now or 12700K with, lets say, 2 more Pcores instead Ecores?
Also, from a manufacturing side, Ecores are smaller so easier to make them instead of just Pcores. But that is relative.

At least, that is just what I think. I dont doubt people have different point of view.

ratirtGaming comparison is valuable info I'd say.

It'll be part of the next round of CPU reviews. For this round I didn't measure it on most comparison processors, so I'm lacking the data and don't have time to setup 20 systems just for this one data point

W1zzardIt'll be part of the next round of CPU reviews. For this round I didn't measure it on most comparison processors, so I'm lacking the data and don't have time to setup 20 systems just for this one data point

I figured that is the challenge here but didnt wanted to assume things. A lot of stuff and things change. Retesting everything is a hard and time consuming task. :)

Dr_b_do you notice a performance boost in games with the E Cores on in your cpu? You make a claim about it performing better, even in gaming, when most games only use at most 6 to 8 cores. And that is the point here, E Cores share cache with P cores, and E Cores are not as performant. Thread director is also not perfect, OS can dump jobs mistakenly to E Cores, then your perf can be reduced.

intel/comments/qq7q6w
"In E-Off, E-Cores were disabled and AVX512 were enabled via BIOS. Arma 3 saw the biggest improvement, with six siege seeing a minor but measurable increase. Disabling E-core also did not reduce performance in any scenario (again, small sample)."

the analysis in here, in this website.



with Windows 11 there is a little regression disabling E-cores while gaming.
I dont care about random reddit users...

CrackongThanks for the advice

But I bought 12400
I did not waste money on the Error cores
And I did not waste time to disable them / nor paid beta testing Windows 11

ok now everything is more clear to me.
You settled for an inferior CPU and now your are claiming E cores are useless.

Great for you.

ratirtNeeds you mean uses less? Would you mind showing me some evidence about that cause I could not find any direct comparisons. Maybe it depends on the game but I looked over (TPU), found the 12900K and system power usage while gaming was around 570W for the whole system with PL1=125W PL2=125W so I assume 12700K would have used a bit less. 5800x used 364W but with a 2080 TI not 3080 like Intel processor so you can chop 60W off. So this is what I have found so not sure how you have calculated it uses less power while gaming and thus my surprise. If it is less indeed, how much less is it? From my search here on TPU, with taking some logic into perspective, it's hardly possible that the 12700K used much less power than what the 12900K did.

I didnt find any direct comparisons here on TPU about gaming vs power consumption except the 12900K. @W1zzard any reason for that? Gaming comparison is valuable info I'd say.




BTW.
Is Hybrid arch an improvement? It is an improvement but.
Intel did the e cores to have more cores available and that (in my opinion) is due to Intel being unable to get more P cores in there to tackle the core count with AMD counterparts. Kinda marketing scheme to be fair.
Ecores dont have hyperthreading Pcores does but you can fit more ecores since these are smaller. I think 1 pcore is equivalent to 1.5 Ecore in size. Correct me if I'm wrong.
So improvement is there but i can bet it would have been better if you replaced Ecores with Pcores. I think, it would have been better for Intel to put more Pcores instead of Ecores. As you know, Pcores are much faster and have HT but the core count would have been lowered significantly and from the marketing perspective would not look as good with lower core count in comparison to AMD's offerings. The other aspect is power consumption. Ecores use less power than Pcores and thus makes them efficient by a significant margin.
Is it an improvement? You get more cores so in a way it is. Is it the right direction for the improvement? You can argue here if you think strictly about performance.
So when you say your 12700K is better than CPUs without ecores sure you are right but if you had the choice to have a 12700K without Ecores but instead you would get more Pcores, which one would you pick? 12700K as it is now or 12700K with, lets say, 2 more Pcores instead Ecores?
Also, from a manufacturing side, Ecores are smaller so easier to make them instead of just Pcores. But that is relative.

At least, that is just what I think. I dont doubt people have different point of view.

It is not a point of view.
it is a matter of fact confirmed by someone who owns 12700K, 5900X and 5800X in my house



and by the way Igorslab already compared gaming power consumption of those CPUs:





12700K is far more efficient than 5800X.
My tests at home absolutely confirmed what Igorslab reported (actually in Far Cry 6 I have a lower power usage, but I'm playing at 1440P, so it is normal).
I don't know why people are in denial here. Maybe too many years of Intel 14++++++ ....

TiggerSo the extra 2/4 cores would be no use either then i guess

i await your reply as you also can't help but reply to me :)

Anyway enough discussing when you are just trolling me now. bye /ignore

it sounds like sour grapes to me... ;)

Max(IT)ok now everything is more clear to me.
You settled for an inferior CPU and now your are claiming E cores are useless.

Great for you.

Have you check Buildzoid's analysis?
You asked, I delivered, but where is your follow-up ?
He did series of studies for P-cores and Error-cores

Those Error-cores are like the shortest plank of the barrel.
When they are enabled, your poor CPU have to downclock its components to match them, thus hindering performance.
Although they are called "Efficient cores" but they are not that efficient and eats up the power budget which could translate to more P-core performance if correctly ultilized.
Not to mention the totally unnecessary scheduler introduces compatibility problems.

There are reasons why Intel themselves chose to retain a pure P-core design in their Sapphire Rapids lineup.
They certainly knew their hybrid architecture would be disaster for any mission critical applications.

CrackongHave you check Buildzoid's analysis?
You asked, I delivered, but where is your follow-up ?
He did series of studies for P-cores and Error-cores

Those Error-cores are like the shortest plank of the barrel.
When they are enabled, your poor CPU have to downclock its components to match them, thus hindering performance.
Although they are called "Efficient cores" but they are not that efficient and eats up the power budget which could translate to more P-core performance if correctly ultilized.
Not to mention the totally unnecessary scheduler introduces compatibility problems.

There are reasons why Intel themselves chose to retain a pure P-core design in their Sapphire Rapids lineup.
They certainly knew their hybrid architecture would be disaster for any mission critical applications.

and again, I don't care about Buildzoid analysis (the same guy making 1+ hours video analyzing Z690 boards just looking at the manufacturer webpage instead of actually test motherboards ‍:kookoo:).
The articles about E cores performance are here in Techpowerup website, and I trust this website very much.

www.techpowerup.com/review/alder-lake-windows-10-performance/

sure, software and scheduler has to be improved, and it is happening over the time.

Max(IT)and again, I don't care about Buildzoid analysis (the same guy making 1+ hours video analyzing Z690 boards just looking at the manufacturer webpage instead of actually test motherboards ‍:kookoo:).
The articles about E cores performance are here in Techpowerup website, and I trust this website very much.

sure, software and scheduler has to be improved, and it is happening over the time.

So
Did you watch the deep analysis?

Do you know what components in your CPU will be downclocked when error cores are enabled ?
Do you know how much the error cores are affecting your CPU's interconnect speed, and IMC frequency ?

The problem is there, blindfolded yourself won't solve the problem.

CrackongSo
Did you watch the deep analysis?

Do you know what components in your CPU will be downclocked when error cores are enabled ?
Do you know how much the error cores are affecting your CPU's interconnect speed, and IMC frequency ?

The problem is there, blindfolded yourself won't solve the problem.

so, did you read Techpowerup test ?

Keep calling them "error cores" because you don't have it, doesn't make your point valid.
On Windows 11 a 12900K with E cores enabled is 1.2% FASTER than with E cores disabled.

Max(IT)so, did you read Techpowerup test ?

Keep calling them "error cores" because you don't have it, doesn't make your point valid.
On Windows 11 a 12900K with E cores enabled is 1.2% FASTER than with E cores disabled.

There is a tradeoff though, several other reputable sites/tubers show its a wash, some improvement in some games, others a regression, and running win11, which is even by MS standards not great. id still rather not have the E Cores, a smaller die, or more P Cores. I agree with another poster about E Cores being a marketing stunt to fit more cores into a CPU, can't believe so many are convinced having efficiency cores that are less performant in a non-power constrained desktop makes any sense (oh look 1.2%!!!11oneleven, and you had to have 8 of them to get that performance "advantage"), and are happily running win11, this is surely the end of civilization.

Max(IT)so, did you read Techpowerup test ?

Keep calling them "error cores" because you don't have it, doesn't make your point valid.
On Windows 11 a 12900K with E cores enabled is 1.2% FASTER than with E cores disabled.

Ofc I read TPU review.

And I see you are unable to understand our argument.
My first reply in the post already stated I want more P cores over the Error cores.
Each Error core cluster could fit at least one P core
So the "Pure P core CPU" would always have 2 more P cores than the hybrid ones.
That basically means a 12900k/12700k with Error cores disable vs a 12600k
The "Pure P core CPU" will always have 25-33% more P-cores, Higher boost clocks, faster interconnects, able to run higher frequency memory, and no compatibility issues.
That will certainly be a win of the "Pure P core CPU".

(Edit: by the time of the 13th gen, Intel would double the amount of Error cores
Since the die area is around 1.5x for a cluster vs P-core, 4 cluster = 6 P-cores
So the "Pure P core CPU" would have 75-100% more P-cores if the die spaces are fully utilized,
But all is lost when they decided to go the Error core route)

That's why I suggested you to watch Buildzoid, He deep tested everything and listed every calculations he did.


Yet you are still blindfolded, only comparing Error cores on vs off , obviously you really don't understand.

Intel went hybrid, in my view, to control temperature. The E cores in combination with the P cores allow a level of performance while keeping temperatures in check. I suspect there is a reason why there isn't an all P core 12900 as some seem to be proposing.

Max(IT)the analysis in here, in this website.



with Windows 11 there is a little regression disabling E-cores while gaming.
I dont care about random reddit users...


ok now everything is more clear to me.
You settled for an inferior CPU and now your are claiming E cores are useless.

Great for you.


It is not a point of view.
it is a matter of fact confirmed by someone who owns 12700K, 5900X and 5800X in my house



and by the way Igorslab already compared gaming power consumption of those CPUs:





12700K is far more efficient than 5800X.
My tests at home absolutely confirmed what Igorslab reported (actually in Far Cry 6 I have a lower power usage, but I'm playing at 1440P, so it is normal).
I don't know why people are in denial here. Maybe too many years of Intel 14++++++ ....


it sounds like sour grapes to me... ;)

Very good post.

mamaIntel went hybrid, in my view, to control temperature. The E cores in combination with the P cores allow a level of performance while keeping temperatures in check. I suspect there is a reason why there isn't an all P core 12900 as some seem to be proposing.

If they had dropped the E cores and added more P cores, the 12700/900k would be even more powerful than they are, maybe they did want to test the hybrid structure and knew the P cores where powerful enough so did not need to add more. Who knows

mamaIntel went hybrid, in my view, to control temperature. The E cores in combination with the P cores allow a level of performance while keeping temperatures in check. I suspect there is a reason why there isn't an all P core 12900 as some seem to be proposing.

The E-cores are there to bump up the multicore benchmarks only.
If we draw a line at the PL2 of the 12900k ~240W , a 8+8 CPU will have better MT benchmark performance vs a theoretical 10 P-core CPU (5~8%)
However is that worth all the trouble caused by the hybrid design and windows 11?
Personally I think not.

Max(IT)It is not a point of view.

It kinda is a point of view of you consider, that statement is about Ecores and Pcores not power consumption. Too bad you have missed it.

Max(IT)it is a matter of fact confirmed by someone who owns 12700K, 5900X and 5800X in my house



and by the way Igorslab already compared gaming power consumption of those CPUs:





12700K is far more efficient than 5800X.
My tests at home absolutely confirmed what Igorslab reported (actually in Far Cry 6 I have a lower power usage, but I'm playing at 1440P, so it is normal).
I don't know why people are in denial here. Maybe too many years of Intel 14++++++ ....

Nice. What is weird, is that the 12600K is using more power than the 12700K.
I'm gonna have to check my 5800x with Horizon Zero Dawn what it will show. I have the 6900xt so it will be apples to apples.

Dr_b_There is a tradeoff though, several other reputable sites/tubers show its a wash, some improvement in some games, others a regression, and running win11, which is even by MS standards not great. id still rather not have the E Cores, a smaller die, or more P Cores. I agree with another poster about E Cores being a marketing stunt to fit more cores into a CPU, can't believe so many are convinced having efficiency cores that are less performant in a non-power constrained desktop makes any sense (oh look 1.2%!!!11oneleven, and you had to have 8 of them to get that performance "advantage"), and are happily running win11, this is surely the end of civilization.

another drama queen about Windows 11 ...

I saw this happening in any Windows new release :rolleyes:
Windows 11 is just fine and it is improving update after update.

ratirtNice. What is weird, is that the 12600K is using more power than the 12700K.

in some games is like that, in the other the 12700K is using more... it could just be better silicon in the 12700K in that specific case.
BTW the point was the efficiency of Alder Lake.

ratirtI'm gonna have to check my 5800x with Horizon Zero Dawn what it will show. I have the 6900xt so it will be apples to apples.

I checked my 5800X in APEX Legends, a game I have on both PCs, and it's using 5 to 10W more than my 12700K, on average.

CrackongThe E-cores are there to bump up the multicore benchmarks only.

PC world is not only benchmarks and gaming.
E-cores has a relevant impact on performance in many applications:







THIS website did an extensive analysis about that, but I think it is easier to just look at Buildzoid ranting on an Youtube video, right ?
For a "professional overclocker" computer usage is just running a benchmark.
For me it is something different.

Max(IT)PC world is not only benchmarks and gaming.
E-cores has a relevant impact on performance in many applications:
THIS website did an extensive analysis about that, but I think it is easier to just look at Buildzoid ranting on an Youtube video, right ?
For a "professional overclocker" computer usage is just running a benchmark.
For me it is something different.

"PC world is not only benchmarks and gaming."
Nice sentence, but do you really understand ?

Backwards compatibility is the most important factor in x86
Being an hybrid architecture and relying on the beta testing windows platform automatically fails in any mission critical applications.

I've mention this in previous reply but clearly you didn't read.
There are reasons why Intel themselves retain "Pure P core" structure in Sapphire Rapids.
Because no one would consider a beta testing hybrid architecture in production environment.
No one would risk a chance of important applications being randomly thrown into Error cores and suffer performance penalty (or simply crash due to incomplete instruction set)

I am sorry but you simply don't understand.
These hybrid CPUs are exactly the opposite of your sentence.
They can't be trusted to do anything critical besides benchmarks and gaming.
They are nice, but not trust worthy and has to be babysit in order to function "normally".

As for buildzoid,
Once again,
Did you watch that video I posted up there?
Did you find anything "incorrect" in that video?
Can you prove that his data is "wrong" / "not trust worthy" / "just ranting" ?
If not, then his analysis on P+E cores should be considered "correct" and you should listen.

Max(IT)another drama queen about Windows 11 ...

I saw this happening in any Windows new release :rolleyes:
Windows 11 is just fine and it is improving update after update.

A lot of people will disagree with you, its changes for the sake of changes, and poorly done, by a company that has the resources do it right and still failed

Crackong"PC world is not only benchmarks and gaming."
Nice sentence, but do you really understand ?

Backwards compatibility is the most important factor in x86
Being an hybrid architecture and relying on the beta testing windows platform automatically fails in any mission critical applications.

I showed "mission critical applications" here above, with performance improvements and no fails.

CrackongI've mention this in previous reply but clearly you didn't read.
There are reasons why Intel themselves retain "Pure P core" structure in Sapphire Rapids.
Because no one would consider a beta testing hybrid architecture in production environment.
No one would risk a chance of important applications being randomly thrown into Error cores and suffer performance penalty (or simply crash due to incomplete instruction set)

baseless assumptions.
A lot of people are using Alder Lake in production environment, me included.

CrackongI am sorry but you simply don't understand.
These hybrid CPUs are exactly the opposite of your sentence.
They can't be trusted to do anything critical besides benchmarks and gaming.
They are nice, but not trust worthy and has to be babysit in order to function "normally".

again, baseless assumptions.
Alder Lake works without any babysitting. Just using the right operative system.

Max(IT)I showed "mission critical applications" here above, with performance improvements and no fails.

baseless assumptions.
A lot of people are using Alder Lake in production environment, me included.

again, baseless assumptions.
Alder Lake works without any babysitting. Just using the right operative system.

My "baseless assumptions" is accompanied by the fact that Intel's own mission critical flagship Sapphire Rapids NOT having a Hybrid architecture.

Now, your turn, time to prove your baseless assumptions,
Please tell me WHO are using Alder Lake Hybrid CPUs in "Production environment" when Alder Lake Xeon W NOT YET existed in the market.
Please don't tell me throwing a 12900k into a typical Z690 MB is a "Production Environment".
Do you really know what's a "Production Environment" ?

Also, you still misunderstood what I said.
Not all Alder Lake CPUs need babysitting, just the Hybrid ones.

And again and again,
Did you watch the video up there?
Find anything "incorrect" ?
If you don't , then my "assumptions" aren't "baseless"