Valantar.... sooooooo maybe that should tell you that simple tests work for simple products with few variables, while more complex products with more variables might need more complex testing? Just a thought.

Looking forward to seeing your results.

Yes. It's not super light, but it's not a particularly heavy workload. There's a reason why nobody uses CB as a measure for ST power draw.

The part that's not believable is the sheer number of ... well, numbers you keep pulling out of thin air with zero corroboration, whether from documenting your own testing or from other sources. You keep making statements that break significantly with results from other well established and trusted sources, which puts the onus on you to corroborate your data. Instead, you keep making unsubstantiated claims.

Ah, so you weren't lying, but you were lying? Got it. Cool.

Also: you're wrong. Igor tests AM4 systems with PBO auto, inlcuding their 12400 "workstation" (CPU focused loads) review. Unless you're looking at the gaming review, which literally doesn't have any CPU-based power testing, only game testing? I mean ... I shouldn't have to tell you that to test CPU power consumption, you need some kind of controllable load, and that games are not this whatsoever. If you're looking at CPU efficiency, you need to run CPU tests to do so. That doesn't render the gaming tests invalid, but they have too many variables to pin-point the exact reasons for the specific power consumption - is the workload CPU or GPU bound, is there a GPU driver issue loading or keeping the CPU idle or other driver overhead that differs between CPU architecutres, does the game behave differently on AMD or Intel CPUs, does the game run at a higher FPS on one, requiring more CPU power to keep up, etc. You can't control for this in a game - there are too many variables - which means you can't actually test for anything resembling CPU architectural efficiency in games.

Really? As you have said time and time again: there are tons of benchmarks out there. So far I've only seen Cinebench from you? You seem to have the time and resources to do at least some benchmarking, so I'd recommend diversifying that workload a bit.

What numbers are you talking about ive literally no idea. The cbr23 numbers already were corroborated by a review i linked from club365, they hit the exact number i said im getting (23500 to 24500).

I mean at this point im not sure, are we past the point of debating whether tpus review is wrong? Cause it absolutely is and if we can't even agree with that then this is all pointless.

Yes igors tested with pbo auto but in the case of the 5600x it reported a total package power of 90w, which if im not mistaken is the default PPT of the cpu, right? So it was basically running stock afaik. So the efficiency comparison between these 2 cpus (12400 and 5600x) were valid (assuming, again, the default ppt of 5600x is 90). The rest werent, thats why i didnt use them as a comparison point, a thing you have now accused me 3 times, claiming im lying when im absolutely not.

fevgatosDude, are you daft? In that specific test it draws the exact same wattage as the 12900k. So.. At the same wattage the 12900k should slam the 12600k in cbr23. But it didnt. Therefore the test is laughably wrong

And yes, the 12900k underperforms in every power limited test, not just in the 125w. The 100, 75 and 50 are also hilariously wrong.

Stop insulting people with your attitude cause I have had your fanboish craze on this forum literally. It has been said confront this with W1zz and ask him about the specifics of the test or review. Maybe there are details you dont know about. Pinpoint your findings and go there and fight for the righteous cause.
If you still wish to keep talking about how great 12900K and other Intel products are, or maybe specifically about the issues you have with W1zzard's review make a separate thread and stop flooding this, AMD 7000 series CPUs with your Intel stuff over and over since this is getting really annoying. You are in literally every AMD CPU related thread talking about how great Intel 12th gen is.

ratirtStop insulting people with your attitude cause I have had your fanboish craze on this forum literally. It has been said confront this with W1zz and ask him about the specifics of the test or review. Maybe there are details you dont know about. Pinpoint your findings and go there and fight for the righteous cause.
If you still wish to keep talking about how great 12900K and other Intel products are, or maybe specifically about the issues you have with W1zzard's review make a separate thread and stop flooding this, AMD 7000 series CPUs with your Intel stuff over and over since this is getting really annoying. You are in literally every AMD CPU related thread talking about how great Intel 12th gen is.

If you stop replying, ill stop it as well
When i have 3 4 people quoting me, you dont expect me to reply?

fevgatosWhat numbers are you talking about ive literally no idea. The cbr23 numbers already were corroborated by a review i linked from club365, they hit the exact number i said im getting (23500 to 24500).

a) you keep saying things like

fevgatosA 16p core intel cpu at 130W would vastly outperform the 12900k at 240w and the 5950x at its current 125w limit. So power and heat aint an issue at all, its die space. Even at 240w a 16p core would be way easier to cool than the 12900k, and it would score over 36k in cbr23.

fevgatosPersonally I tested 3 12900k at 4 different mobos and they all came back with the same results, 23500 to 24500 at 125w. Nowhere near close to TPUs numbers.

fevgatosYou mean oced? Around 29950, unify X and 6000c30 ram on a u12a cooler

fevgatosAnd i know cause ive tested, 4 motherboards and 3 cpus, all scored 23500 to 24500 at 125watts.

Ask anyone with a 12900k to test stock with 125w power limit, they will all verify what im telling you. They'll score over 23k points

Without providing any sources or corroborating data for those statements. Though you've improved slightly in this regard since being called out on it, you still keep doing the same thing. How hard is it to select the text of your data, press ctrl+k, and paste in the link from the site you're referring to? This is just lazy.
b) I'm a tad dubious about those Club386 numbers simply because of the test setup: they don't differentiate between PL1 and PL2, instead just talking about "TDP" across all platforms, which makes it impossible to know what the actual settings are. That doesn't mean their numbers are wrong, it just means I don't know the power level they were actually achieved at.

There's also the issue of your consistent denial towards there being any kind of issue with extrapolating MSDT GC performance into your theoretical 16c CPU, despite the indisputable fact that this would necessitate major reconfigurations of its internal fabric and other aspects crucial to performance.

fevgatosI mean at this point im not sure, are we past the point of debating whether tpus review is wrong? Cause it absolutely is and if we can't even agree with that then this is all pointless.

I've accepted long ago - explicitly, in posts directed at you - that there might be something wrong with that testing. The problem is, until we know what went wrong and how, and have equally stringently tested data saying otherwise, we're none the wiser. And sadly the reviews you've linked showing different results are much less thorough, or have issues of their own - the Club386 thing above, Igor's Lab running AM4 with an auto OC. The Techspot 12700 review has the same issue as Club386 - I love that they test at both "unlimited" (their board's stock behaviour) and "65" settings, but ... what are those 65W? PL1? PL2? What's Tau set to? Without detailing this, the data becomes a lot more murky.

Test methodology is crucial, and presenting that methodology clearly and with the necessary level of detail can make or break the quality of the conclusions derived from testing. There's a reason why everyone isn't a hardware reviewer: it's difficult, and takes a lot of care and attention, as well as the development of a mode of presentation that maintains this data while still making it readable and understandable to the audience. That's hard.

fevgatosYes igors tested with pbo auto but in the case of the 5600x it reported a total package power of 90w, which if im not mistaken is the default PPT of the cpu, right? So it was basically running stock afaik.

No. PBO boosts all cores higher than stock, pushing voltages and thus core power higher than stock. I've shown above from my own testing how letting Zen3 boost too aggressively will hurt performance, tanking efficiency. Why? Because it doesn't scale higher in terms of clocks, but in terms of thermals it can take more power if you'll let it - you just get nothing in return. That's why you don't test with an auto OC mode enabled.

As for lying: you literally said "I don't use Igor's Lab for efficiency comparisons", then admitted in the next sentence that you had indeed done so. And this is the core of the problem here: you don't consider your words before posting. Heck, I'm not necessarily sure that you're entirely wrong about things here, but you seem fundamentally unable to present things in a reasonable, measured, well-supported way, instead blurting out grandiose statements that you then have to walk back when challenged. Your way of discussing forces people to constantly be correcting you or desperately try to shove some nuance back into your simplifications, which is why we're in this quasi-hostile mode of discussing in the first place. Things would get a lot, lot better if you took some more care with how you phrased things, thought things through a bit more, corroborated your statements with data or sources consistently, and made nuanced arguments rather than black-and-white statements.

Valantara) you keep saying things like




Without providing any sources or corroborating data for those statements. Though you've improved slightly in this regard since being called out on it, you still keep doing the same thing. How hard is it to select the text of your data, press ctrl+k, and paste in the link from the site you're referring to? This is just lazy.
b) I'm a tad dubious about those Club386 numbers simply because of the test setup: they don't differentiate between PL1 and PL2, instead just talking about "TDP" across all platforms, which makes it impossible to know what the actual settings are. That doesn't mean their numbers are wrong, it just means I don't know the power level they were actually achieved at.

I've accepted long ago - explicitly, in posts directed at you - that there might be something wrong with that testing. The problem is, until we know what went wrong and how, and have equally stringently tested data saying otherwise, we're none the wiser. And sadly the reviews you've linked showing different results are much less thorough, or have issues of their own - the Club386 thing above, Igor's Lab running AM4 with an auto OC. The Techspot 12700 review has the same issue as Club386 - I love that they test at both "unlimited" (their board's stock behaviour) and "65" settings, but ... what are those 65W? PL1? PL2? What's Tau set to? Without detailing this, the data becomes a lot more murky.

Test methodology is crucial, and presenting that methodology clearly and with the necessary level of detail can make or break the quality of the conclusions derived from testing. There's a reason why everyone isn't a hardware reviewer: it's difficult, and takes a lot of care and attention, as well as the development of a mode of presentation that maintains this data while still making it readable and understandable to the audience. That's hard.

No. PBO boosts all cores higher than stock, pushing voltages and thus core power higher than stock. I've shown above from my own testing how letting Zen3 boost too aggressively will hurt performance, tanking efficiency. Why? Because it doesn't scale higher in terms of clocks, but in terms of thermals it can take more power if you'll let it - you just get nothing in return. That's why you don't test with an auto OC mode enabled.

As for lying: you literally said "I don't use Igor's Lab for efficiency comparisons", then admitted in the next sentence that you had indeed done so. And this is the core of the problem here: you don't consider your words before posting. Heck, I'm not necessarily sure that you're entirely wrong about things here, but you seem fundamentally unable to present things in a reasonable, measured, well-supported way, instead blurting out grandiose statements that you then have to walk back when challenged. Your way of discussing forces people to constantly be correcting you or desperately try to shove some nuance back into your simplifications, which is why we're in this quasi-hostile mode of discussing in the first place. Things would get a lot, lot better if you took some more care with how you phrased things, thought things through a bit more, corroborated your statements with data or sources consistently, and made nuanced arguments rather than black-and-white statements.

Techspot used the rm1 cooler for the 65w result. Im not sure exactly what your issue is with that one, to me its obvious they set pl2 to 65w. It wouldnt make sense any other way since, the way they phrased it in the review, it would be idiotic to have pl1 at 65 and unlimited pl2. Plus the score would have been higher if that was the case, as demonstrated by their power unlimited test.

I dont see anything wrong with igorslab review either. Yes zen were run with pbo but that's irrelevant, what matters is the 12900k perfromance at 125w. Since in the blender test it matches a pboed 5900x and slaps the 12600k, there is no way in cbr23 it gets matched by the 12600k.

Club365 used xtu to power limit, they even have a picture of their settings in the first page. And since the numbers perfectly match the ones i observed with 3 cpus tested witb 4 different motherboards, i have no reason to doubt them. The cpu is running at 4.3 ghz for the pcores

www.club386.com/wp-content/uploads/2021/11/Power-Limiting1-1068x530.jpg

Today im back, i can show you the 29900 score and all that, but i find it irrelevant cause, how can i show you im running a u12a? I might as well have a custom loop for all you know.

Regarding what went wrong with tpu, fixed voltage is by far the most likely explanation. Msi boards are quite reknown (mine included) for doing shaenaningans, although he uses an asus hero, and as far as my experience with the asus apex goes, it wasnt doing any weird stuff when plimited,but who knows, maybe the hero does.

What i find really weird is how he himself didnt get puzzled with the results. That's by far my biggest surprise.

fevgatosWhen i have 3 4 people quoting me, you dont expect me to reply?

I expect you to stop flooding AMD threads with your fanboish Intel stuff and believe me no one will bother you and quote your comments if these aren't there.
Respect a subject of a conversation in a thread. Fairly simple.

fevgatosGo ahead, I hope he replies. I guarantee you 100% the benchmarks are flawed. Could be a bios thingy or something else, but its most definitely without a shadow of a doubt flawed. Im not the only one saying it, there is a thread on tom's hardware also making fun of that benchmarks, and even in the discussion of that very benchmark there were people doubting the results. That's cause they just don't make any sense, the 12600k cant be more efficient than the 12900k at same wattage, it's hillariously obvious. The flaw is so monumental, imagine if you clock the 5600x to 125w and suddenly it matches the 5950x. Well thats what you are looking at with those numbers...

Ive tested 3 12900k on 4 motherboards at 125w, all scored pretty much the same in CBR23, between 23500 and 24500. TPU scored 18k, lol

Okay, I just found this review on Tom's Hardware. Take a look at "Intel Alder Lake Core i9-12900K and i5-12600K Power Consumption, Efficiency, and Thermals" section. It shows 12600k being more efficient than 12900k at both DDR4 and DDR5 configurations.

That's in "Handbrake power efficiency - x264 renders per hour"




as well as "Handbrake power efficiency - x265 renders per hour"


Here is in the "blender bmw27 power efficiency". Compare the y-axis height of these pairs (12900k vs 5950x) (12700k vs 5900x) (12600k vs 5800x).
Only 12600k beats 5800x in efficiency, other pairs lose by a big margin. This shows both the strengths of 12600k (sweet spot with P cores) and weakness of 5800x (bad energy efficiency compared to the rest of the zen 3 lineup).


as well as the "blender koro power efficiency".
Pretty much the same thing continues as in previous example. In this example 12600k does consume a little less power at the expense of being a little slower than 5800x, which puts their efficiency at about the same level or slightly in 12600k's favour.


Needless to say, in those first 2 examples 5950x (best amd's zen3 efficiency processor) compared to 12900k (one of the worst intel's 12th gen efficiency processor), it comes out with 50% better efficiency.
And this is as you say in the Tom's hardware's own benchmark which supposedly is making fun of TPU's benchmark results.

I'd advise you to put your money where your mouth is. Look at the benchmarks of the site you were citing before making such claims. This site (Tom's review) gave exactly the different results to what you were saying it did.

mahirzukic2Okay, I just found this review on Tom's Hardware. Take a look at "Intel Alder Lake Core i9-12900K and i5-12600K Power Consumption, Efficiency, and Thermals" section. It shows 12600k being more efficient than 12900k at both DDR4 and DDR5 configurations.

That's in "Handbrake power efficiency - x264 renders per hour"




as well as "Handbrake power efficiency - x265 renders per hour"


Here is in the "blender bmw27 power efficiency". Compare the y-axis height of these pairs (12900k vs 5950x) (12700k vs 5900x) (12600k vs 5800x).
Only 12600k beats 5800x in efficiency, other pairs lose by a big margin. This shows both the strengths of 12600k (sweet spot with P cores) and weakness of 5800x (bad energy efficiency compared to the rest of the zen 3 lineup).


as well as the "blender koro power efficiency".
Pretty much the same thing continues as in previous example. In this example 12600k does consume a little less power at the expense of being a little slower than 5800x, which puts their efficiency at about the same level or slightly in 12600k's favour.



Needless to say, in those first 2 examples 5950x (best amd's zen3 efficiency processor) compared to 12900k (one of the worst intel's 12th gen efficiency processor), it comes out with 50% better efficiency.
And this is as you say in the Tom's hardware's own benchmark which supposedly is making fun of TPU's benchmark results.

I'd advise you to put your money where your mouth is. Look at the benchmarks of the site you were citing before making such claims. This site (Tom's review) gave exactly the different results to what you were saying it did.

If you think the graphs you just posted disagree with what Im saying then you either don't understand the graphs or what Im saying.

I never - ever - ever - ever - EVER - EVER EVER suggested that the 12900k at stock is more efficient than the the 12600k or the 5950x in MT workloads. Ever. Never ever. Actually, quite the contrary, I've said multiple times that at stock power limits it's extremely inefficient in these MT workloads. If you understood something different then the problem lies with you

fevgatosIf you think the graphs you just posted disagree with what Im saying then you either don't understand the graphs or what Im saying.

I never - ever - ever - ever - EVER - EVER EVER suggested that the 12900k at stock is more efficient than the the 12600k or the 5950x in MT workloads. Ever. Never ever. Actually, quite the contrary, I've said multiple times that at stock power limits it's extremely inefficient in these MT workloads. If you understood something different then the problem lies with you

I think this is what you said:

Valantar... Except Zen3 cores peak around 20W, while ADL P cores can draw 2-3x that much. More efficient at lower clocks? Depends on the workload. More efficient at stock? Not even close in any CPU heavy task. They do run very well in games though, with most of those being variable, low threaded workloads that let the CPU boost high to race to finish each frame's compute cycle, which suits ADL's high clocks and good IPC nicely. But, crucially, you can't reliably measure a CPUs efficiency in something that isn't a cpu-intensive task. And for anything CPU-intensive, both Zen3 and E cores are vastly more efficient at anything resembling stock power levels.

fevgatosMore efficient at everything. What he is saying is that intel cant fit 16p cores cause of power draw which is absurd, cause we already know a p core outperforms a zen 3 core at same wattage. Therefore a 16p core intel would outperform the 5950x for example at same or lower wattage

It is pretty clearly not, as it is shown by both the TPU's and Tom's benchmarks.

mahirzukic2I think this is what you said:


It is pretty clearly not, as it is shown by both the TPU's and Tom's benchmarks.

Do you understand what "same wattage" means? Of course when you test one CPU at 500watts and the other one at 50w the second one will be more efficient. Do you want me to bold the same wattage part or can you read it? Also - im talking about 8GC cores vs 8 zen 3 cores.

fevgatosDo you understand what "same wattage" means? Of course when you test one CPU at 500watts and the other one at 50w the second one will be more efficient. Do you want me to bold the same wattage part or can you read it? Also - im talking about 8GC cores vs 8 zen 3 cores.

So now it's "More efficient at everything." at some wattage? Some arbitrary wattage? Any wattage? No wattage?

mahirzukic2So now it's "More efficient at everything." at some wattage? Some arbitrary wattage? Any wattage? No wattage?

What do you mean "so now". That's what im saying from the frst post. And yes, at any wattage, from 10w all the way up to 300 watts

fevgatosWhat do you mean "so now". That's what im saying from the frst post. And yes, at any wattage, from 10w all the way up to 300 watts

Yes you are saying that from the first post and it is still 'so now' since you talk about efficiency with wattage you choose is best depending what you are comparing the CPU to not its advertised performance and wattage.
Either way case is closed. If you really wanna talk about 12th gen Intel please make your own thread.

Either way.
I think 7000 series AMD will top the charts with the frequency 5.7Ghz if true. that is a substantial bump.

fevgatosWhat do you mean "so now". That's what im saying from the frst post. And yes, at any wattage, from 10w all the way up to 300 watts

But that's exactly the thing most people (now me too) have been trying to show you. That is just not true. And it's backed up by benchmarks from TPU and the one you mentioned on Tom's Hardware.
It's not more efficient at any wattage.
It's mostly less efficient at any wattage, and in some biased scenarios specifically Intel's Alder Lake undervolted and underclocked vs stock zen3 with PBO on, then intel's AL is more efficient than zen 3.
Then again, that's not really a fair comparison, as zen 3 can also be undervolted and underclocked, which still results in AL being less efficient, which corresponds to the stock values as well.

mahirzukic2But that's exactly the thing most people (now me too) have been trying to show you. That is just not true. And it's backed up by benchmarks from TPU and the one you mentioned on Tom's Hardware.
It's not more efficient at any wattage.
It's mostly less efficient at any wattage, and in some biased scenarios specifically Intel's Alder Lake undervolted and underclocked vs stock zen3 with PBO on, then intel's AL is more efficient than zen 3.
Then again, that's not really a fair comparison, as zen 3 can also be undervolted and underclocked, which still results in AL being less efficient, which corresponds to the stock values as well.

I think you don't understand jack. I invited you to the other thread with results posted

fevgatosI think you don't understand jack. I invited you to the other thread with results posted

I don't own neither the AL nor Zen3 CPU in my home rig, but rather an old 2600k, which I am looking to replace with either Intel's 13th gen or Zen4 depending on how they fare and the reviews.
I have a Intel(R) Core(TM) i9-9980HK CPU on my work Macbook and a Intel(R) Core(TM) i7-12700H CPU in my work Dell XPS laptop.
One is running a Mac OS and the other a Ubuntu linux, so sadly I cant post any scores. Also these being laptop CPUs and as such not comparable to the desktop counterparts.

mahirzukic2I don't own neither the AL nor Zen3 CPU in my home rig, but rather an old 2600k. I have a Intel(R) Core(TM) i9-9980HK CPU on my work Macbook and a Intel(R) Core(TM) i7-12700H CPU in my work Dell XPS laptop.
One is running a Mac OS and the other a Ubuntu linux, so sadly I cant post any scores. Also these being laptop CPUs and as such not comparable to the desktop counterparts.

Well you don't have to post any numbers, you can just check what 8 GC cores do against 8 Zen 3 cores in same wattage. And it's not pretty o_O

fevgatosWell you don't have to post any numbers, you can just check what 8 GC cores do against 8 Zen 3 cores in same wattage. And it's not pretty o_O

Could you please post a link to that thread? I'm very interested in your findings

Arc1t3ctCould you please post a link to that thread? I'm very interested in your findings

There are no findings, nobody with 8 zen 3 cores posted any results yet. And I don't think they will, as I've said a couple of pages ago, whenever I ask someone they just disappear. But here's the link. 8 GC cores at 65 watts score 16500 to 17200 o_O

www.techpowerup.com/forums/threads/cinebench-r23-efficiency-race.297551/#post-4811927

fevgatosMan are you for real? There are power consumption metrics in the review, yes the 12600k consumes 5w less than the 12900k at 125w while it scores the same. Which, as ive repeated multiple times, its impossible


No 12700 isnt the same configuration. It has half the ecores yet at 65w it outperforms the 12900k at 100w,which, again, is absolutely impossible.

There is nothing sketchy about comparing across reviews, cbr23 is a repeatable workload and when tested at similar power limits the cpus should score the same. And i know cause ive tested, 4 motherboards and 3 cpus, all scored 23500 to 24500 at 125watts.

Ask anyone with a 12900k to test stock with 125w power limit, they will all verify what im telling you. They'll score over 23k points

I googled for you some reviews testing at 125w. They all verify what im saying, TPUs review is absolutely wrong. Here you go, 125w = 23500 score

www.club386.com/intel-core-i9-12900k-at-125w/3/

On this link www.club386.com/intel-core-i9-12900k-at-125w/6/ in the concluding section they still have that 5950x at stock has higher efficiency than 12900k at 125w let alone at stock, which is 241w.
Stock for stock, zen 3 in this review is about 50% more efficient than Alder Lake which is consistent with the other reviews (TPU and Tom's).

mahirzukic2On this link www.club386.com/intel-core-i9-12900k-at-125w/6/ in the concluding section they still have that 5950x at stock has higher efficiency than 12900k at 125w let alone at stock, which is 241w.

You realize I posted this link, right?

fevgatosYou realize I posted this link, right?

Yes, a link which clearly shows in it's conclusion that 5950x is more efficient than 12900k at stock (50% more) as well as when power limited to 125w (about 10% more).
Which pretty much show exactly the opposite of what you have been claiming this whole time.

mahirzukic2Which pretty much show exactly the opposite of what you have been claiming this whole time.

What am I claiming? Can you repeat it for me please, cause apparently you haven't got a whiff yet

fevgatosWell you don't have to post any numbers, you can just check what 8 GC cores do against 8 Zen 3 cores in same wattage. And it's not pretty o_O

I don't own Zen 3 nor Adler Lake right now. I have owned CPUs from both manufacturers over the years though, so I have no beef in the game.
All I care is bang for the buck, and lately (last few years) this has been sometimes in AMD's favour, sometimes in Intel's favour.

fevgatosWhat am I claiming? Can you repeat it for me please, cause apparently you haven't got a whiff yet

This:

fevgatos said:
More efficient at everything. What he is saying is that intel cant fit 16p cores cause of power draw which is absurd, cause we already know a p core outperforms a zen 3 core at same wattage. Therefore a 16p core intel would outperform the 5950x for example at same or lower wattage

fevgatosWhat do you mean "so now". That's what im saying from the frst post. And yes, at any wattage, from 10w all the way up to 300 watts

mahirzukic2I don't own Zen 3 nor Adler Lake right now. I have owned CPUs from both manufacturers over the years though, so I have no beef in the game.
All I care is bang for the buck, and lately (last few years) this has been sometimes in AMD's favour, sometimes in Intel's favour.


This:

And how exactly does the review I posted from club365 disprove that? He is testing 8+8 against 16 zen 3 cores.