ratirtYeah you are right. The power consumption literally sucks since 10th gen and it is getting higher every gen slightly.
If Intel put 16 pcores in one CPU it would have been hard to tame for sure. 240watts is literally for 8pcores and 8ecores for the 12900k. with 16pcores, it might have needed more power than a 240w unless Intel would have stopped there.

What you are saying makes no sense whatsoever. I think you don't understand how cpus works. Intel can add 99999 pcores and still have a 240w power limit. Or even a 50w limit.

The 5800x has half the cores of the 5950x but the same tdp. Cores have nothing to do with power consumption. Power consumption is an arbitary
limit that the manafacturer chooses depending on how he wants to place his product.

The 13900k will exists in 3 versions, the t at 35w,the k at 240w and the non k at 65w. So it doesnt matter how many pcores they put, they can choose to run at any power limit thry want

fevgatosWhat you are saying makes no sense whatsoever. I think you don't understand how cpus works. Intel can add 99999 pcores and still have a 240w power limit. Or even a 50w limit.

The 5800x has half the cores of the 5950x but the same tdp. Cores have nothing to do with power consumption. Power consumption is an arbitary
limit that the manafacturer chooses depending on how he wants to place his product.

The 13900k will exists in 3 versions, the t at 35w,the k at 240w and the non k at 65w. So it doesnt matter how many pcores they put, they can choose to run at any power limit thry want

I don't know what you're trying to prove, it's obvious that intel is going to run this i9 at the highest TDP they can to have any chance against the 7950x which has every advantage of a denser and more efficient process.

ZoneDymoin response to article:
Dear lord 380 watts...

that is insane. hard pass for me, especially with rising electricity costs. the 170w AM5 cpu's are as high as I am willing to go. even then I may opt for the 105w AM5 variants.

DenverI don't know what you're trying to prove, it's obvious that intel is going to run this i9 at the highest TDP they can to have any chance against the 7950x which has every advantage of a denser and more efficient process.

Yet according to the leaks amd is also running it at the highest tdp it can achieve no? It seems to be thermally throttling out of the box even with a 360 aio.

fevgatosWhat you are saying makes no sense whatsoever. I think you don't understand how cpus works. Intel can add 99999 pcores and still have a 240w power limit. Or even a 50w limit.

The 5800x has half the cores of the 5950x but the same tdp. Cores have nothing to do with power consumption. Power consumption is an arbitary
limit that the manafacturer chooses depending on how he wants to place his product.

The 13900k will exists in 3 versions, the t at 35w,the k at 240w and the non k at 65w. So it doesnt matter how many pcores they put, they can choose to run at any power limit thry want

That is called product location since there are different market segments and these have something to do with power consumption. Like laptops vs desktop.

Yes 5800x has the same TDP but like I said, TDP is not power consumption. 5950x uses a bit more power and it is only a bit since these are a very efficient cores and are not pushed to hard with frequency.

Yes Intel can increase cores but 9999pcores is absurd. There is a limit to how many pcores can be packed in a die and I guess Intel wants to go with 8 max for obvious reasons. P cores use more power than ecores which is another obvious thing. Each core has to use some power and if you pack too many of them in one die, you would need to increase power consumption since each core needs certain power to operate correctly and it may be too much for coolers to handle. You can lower frequency of the cores to make them consume less power but then the performance will drop as well. This more pcores may not be worth or possible due to constraints and perhaps, that is one of the reasons Intel has ecores.

ratirtThat is called product location since there are different market segments and these have something to do with power consumption. Like laptops vs desktop.

Yes 5800x has the same TDP but like I said, TDP is not power consumption. 5950x uses a bit more power and it is only a bit since these are a very efficient cores and are not pushed to hard with frequency.

Yes Intel can increase cores but 9999pcores is absurd. There is a limit to how many pcores can be packed in a die and I guess Intel wants to go with 8 max for obvious reasons. P cores use more power than ecores which is another obvious thing. Each core has to use some power and if you pack too many of them in one die, you would need to increase power consumption since each core needs certain power to operate correctly and it may be too much for coolers to handle. You can lower frequency of the cores to make them consume less power but then the performance will drop as well. This more pcores may not be worth or possible due to constraints and perhaps, that is one of the reasons Intel has ecores.

You keep showing a fundamental misunderstanding of how cpus work. No pcores do not require more power than ecores. They will both work with as much power as you are willing to give them.

16 p cores will be faster than 8 p-cores with both at 50w. 16p cores will be faster than 8 pcores with both at 200w. No matter the wattage, 16 p cores will be faster. So obviously power is not the problem.

Looks promising - will definitely be interesting to see how the Zen4 v. RPL fight plays out. Also kind of fun to see how hard Intel has clearly been targeting Cinebench as a performance metric in the past two generations, which must be seen as a direct response to AMD using it as one of their major ways of breaking open the entrenched accepted wisdom of "Intel is faster, AMD is cheap" back in 2017. Makes me wonder which benchmark/workload will take this place next - as the one promoted by one party, then used as an explicit optimization target for the other, only for them to "take it back" later.

fevgatosNeither, just performance

For power consumption increases check zen 4, got plenty of that, up to 60% more power.

How so? If you want to assume from leaks at least considered all of them.

fevgatosYou keep showing a fundamental misunderstanding of how cpus work. No pcores do not require more power than ecores. They will both work with as much power as you are willing to give them.

16 p cores will be faster than 8 p-cores with both at 50w. 16p cores will be faster than 8 pcores with both at 200w. No matter the wattage, 16 p cores will be faster. So obviously power is not the problem.

Yes they will be faster because there's 16 of them. Intel literally said that ecores (literally efficiency cores) are using less power because they are clocked lower and have no SMT among other things.
They don't require more power? It is not just power but also performance with a given power budget. Four Ecores, according to Intel, can deliver 80% more performance than 2 skylake cores and still use less power. Obviously Intel went with ecores to boost MT performance and not use as much power. You can cap pcores to 35watts but it will cut performance as well.

ratirtYes they will be faster because there's 16 of them. Intel literally said that ecores (literally efficiency cores) are using less power because they are clocked lower and have no SMT among other things.
They don't require more power? It is not just power but also performance with a given power budget. Four Ecores, according to Intel, can deliver 80% more performance than 2 skylake cores and still use less power. Obviously Intel went with ecores to boost MT performance and not use as much power. You can cap pcores to 35watts but it will cut performance as well.

Again, you dont really know what you are talking about. According to intel, pcores are faster at same power than ecores. Therefore they require less power for the same performance. Therefore 16 p cores would be faster than 16 ecores.

Conclusion is, power isnt the reason they dont have a 16p cpu. Its die size.

1d10tHow so? If you want to assume from leaks at least considered all of them.

So the 13600k is both cheaper and faster than the 7700x. Lol

fevgatosAgain, you dont really know what you are talking about. According to intel, pcores are faster at same power than ecores. Therefore they require less power for the same performance. Therefore 16 p cores would be faster than 16 ecores.

Conclusion is, power isnt the reason they dont have a 16p cpu. Its die size.


So the 13600k is both cheaper and faster than the 7700x. Lol

13600k if priced at current estimations with the leaked performance will definitely be one heck of processor, certainly among the best (bang for buck) for gaming.

I would assume that it would a tad more expensive than 7600x (due to having 50% higher MT) and steal its thunder.

fevgatosYet according to the leaks amd is also running it at the highest tdp it can achieve no? It seems to be thermally throttling out of the box even with a 360 aio.

No. I just saw an engineering sample on a motherboard still with an old bios running with overvoltage.

While it's true that both sides will push CPUs beyond the efficiency curve, Intel is going to do this much more aggressively, including a boost to 440w+

fevgatosYou keep showing a fundamental misunderstanding of how cpus work. No pcores do not require more power than ecores. They will both work with as much power as you are willing to give them.

Yes, but for a given power budget, you can get more cores, often at a higher clock speed (up to the E core clock ceiling of ~4GHz), using E cores vs. P cores - and in a much, much smaller die area, of course.

fevgatos16 p cores will be faster than 8 p-cores with both at 50w.

In an nT workload. In a non-nT workload, this will not be true - and if given the same boost parameters, the 8p CPU will likely be marignally faster than the 16p one in any 1t-8t workload (putting aside the messiness of SMT for now) due to the extra power needs of those idle cores. The difference won't be much, and will most likely fall within error, but it's crucial to note that a 16p CPU will only be faster in >8t workloads.

fevgatos16p cores will be faster than 8 pcores with both at 200w.

Again: in >8t workloads.

fevgatosNo matter the wattage, 16 p cores will be faster.

In >8t workloads.

fevgatosSo obviously power is not the problem.

No, the problem is that P cores take a ton of area and power, which would force low clocks on a theoretical 16p CPU in an MSDT socket, which would in turn make it faster in nT workloads, but likely slower (or at least no faster) in less heavily multithreaded workloads. Meaning the value proposition would be terrible, while the price (due to the large area) would be very high. Using E cores to fill out that nT "deficiency" of an 8p CPU (compared to 16c Ryzen) solves all of those issues at once - massively improving nT performance, without gobbling massive power, letting the P cores clock high still, and not using tons of die area. It's a win-win-win solution, at the cost of needing a heterogeneous core layout-aware scheduler. Which we already have.

ValantarYes, but for a given power budget, you can get more cores, often at a higher clock speed (up to the E core clock ceiling of ~4GHz), using E cores vs. P cores - and in a much, much smaller die area, of course.

In an nT workload. In a non-nT workload, this will not be true - and if given the same boost parameters, the 8p CPU will likely be marignally faster than the 16p one in any 1t-8t workload (putting aside the messiness of SMT for now) due to the extra power needs of those idle cores. The difference won't be much, and will most likely fall within error, but it's crucial to note that a 16p CPU will only be faster in >8t workloads.

Again: in >8t workloads.

In >8t workloads.

No, the problem is that P cores take a ton of area and power, which would force low clocks on a theoretical 16p CPU in an MSDT socket, which would in turn make it faster in nT workloads, but likely slower (or at least no faster) in less heavily multithreaded workloads. Meaning the value proposition would be terrible, while the price (due to the large area) would be very high. Using E cores to fill out that nT "deficiency" of an 8p CPU (compared to 16c Ryzen) solves all of those issues at once - massively improving nT performance, without gobbling massive power, letting the P cores clock high still, and not using tons of die area. It's a win-win-win solution, at the cost of needing a heterogeneous core layout-aware scheduler. Which we already have.

Yes of course im talking about n core workloads, in other cases no cpu really pushes consumption that high to be an issue.

I still disagree, no core takes a lot of power. Every core in existence uses as much power as it is given. The question is if it performs well enough with that power. We know that pcores are more efficient than ecores at everything when it comes to 5w and above per core. Therefore i really fail to see how a big full pcore cpu will have any problems with power draw.

DenverNo. I just saw an engineering sample on a motherboard still with an old bios running with overvoltage.

While it's true that both sides will push CPUs beyond the efficiency curve, Intel is going to do this much more aggressively, including a boost to 440w+

You have no idea whether or not there is an overvoltage issue. Just because someone undervolted his cpu doesnt mean anything. Since he didnt run any actual benchmark with scores, the cpu might as well have been clock stretching with the lower voltages, meaning it actually runs much slower than thr reported freequency.

Also I was able to undervolt my 12900k by 0.17 volts, dropping the cbr23 consumption to 170w. Does that mean it had an overvoltage issue with the bios?

fevgatosSo the 13600k is both cheaper and faster than the 7700x. Lol

Yet to be seen, as previous gen show the opposite even with all "disadvantages".

fevgatosAlso I was able to undervolt my 12900k by 0.17 volts, dropping the cbr23 consumption to 170w. Does that mean it had an overvoltage issue with the bios?

This really depends on the motherboard. Some give with default settings a higher voltage than what is pre-programmed into the CPU.

Example from personal experience: my 12700K at 4.7 GHz requires according to its built-in voltage-frequency curve 1.22V. By default the motherboard (MSI Z690-A Pro) was giving more than 1.3V in MT AVX loads at this frequency, so it was overvolting it. I found it to be stable at 1.15-1.16V (undervolting it).

fevgatosYes of course im talking about n core workloads, in other cases no cpu really pushes consumption that high to be an issue.

I still disagree, no core takes a lot of power. Every core in existence uses as much power as it is given. The question is if it performs well enough with that power. We know that pcores are more efficient than ecores at everything when it comes to 5w and above per core. Therefore i really fail to see how a big full pcore cpu will have any problems with power draw.


You have no idea whether or not there is an overvoltage issue. Just because someone undervolted his cpu doesnt mean anything. Since he didnt run any actual benchmark with scores, the cpu might as well have been clock stretching with the lower voltages, meaning it actually runs much slower than thr reported freequency.

Also I was able to undervolt my 12900k by 0.17 volts, dropping the cbr23 consumption to 170w. Does that mean it had an overvoltage issue with the bios?

It has already been confirmed that there was an overvoltage problem:

."I saw that the 795 can achieve 38100+ and does not require 360 cooling. I think the problem is the BIOS he's using."

DenverIt has already been confirmed that there was an overvoltage problem:

."I saw that the 795 can achieve 38100+ and does not require 360 cooling. I think the problem is the BIOS he's using."

https://twitter.com/i/web/status/1565698175381704704

Your link has some really nice data, like this one

Greymon55 has also posted the multi-threaded performance benchmarks for the two CPUs. The AMD Ryzen 7 7700X allegedly scores around 19800 points while the Ryzen 5 7600X scores around 15100 points. This puts the Ryzen 5 7600X 34% faster than its predecessor but slower than the i5-12600K. At the same time, the Ryzen 7 7700X is around 29% faster than its predecessor but slower than the Core i7-12700K.

So both the new R5 and R7 lose to Intel's old i5 and i7 despite their highly increased TDP's.. But now they have to compete with the new ones. Holy crap

fevgatosYour link has some really nice data, like this one

Greymon55 has also posted the multi-threaded performance benchmarks for the two CPUs. The AMD Ryzen 7 7700X allegedly scores around 19800 points while the Ryzen 5 7600X scores around 15100 points. This puts the Ryzen 5 7600X 34% faster than its predecessor but slower than the i5-12600K. At the same time, the Ryzen 7 7700X is around 29% faster than its predecessor but slower than the Core i7-12700K.

So both the new R5 and R7 lose to Intel's old i5 and i7 despite their highly increased TDP's.. But now they have to compete with the new ones. Holy crap

When you see that you are wrong do you start to change the direction of the discussion?

These CPUs are for gaming, 6-8 physical cores is enough for that, that's the benchmarks we need to see, not cinebench scores.

DenverWhen you see that you are wrong do you start to change the direction of the discussion?

These CPUs are for gaming, 6-8 physical cores is enough for that, that's the benchmarks we need to see, not cinebench scores.

What do you mean I'm wrong? We already know that TDP for the 5950x is 170w and actual power consumption (PPT) is 230w. Are you saying power draw will be lower than the numbers im claiming?

4 cores were enough for gaming 10 years ago yet people constantly blamed intel for not having 9999 cores. What changed...

fevgatosYes of course im talking about n core workloads, in other cases no cpu really pushes consumption that high to be an issue.

Except that Intel's aggressive turbo clocks cause them to draw 50+W per P core while boosting, while Zen3 tops out at ~20W/core. So even while ADL is faster at those clocks, it needs tons of power to get there. That doesn't mean that it can't run more efficiently - but it would be slower than Zen3 in that case.

fevgatosI still disagree, no core takes a lot of power. Every core in existence uses as much power as it is given.

That is a gross oversimplification, and essentially entirely untrue. All CPU cores have a base power level required for them to function properly at all. They also have a lower threshold of clock speed and power where they become able to perform efficiently - where clocks and performance rise high enough to overcome the inherent disadvantage of that base power draw. For most modern CPU cores that clock speed is quite low, but it's still a threshold. And, crucially, all cores have some form of clock/power ceiling. For Zen3 that is ~21W, where scaling essentially stops past this. You can push more power into them, but clocks will stagnate and performance will drop. Exceeding 30W/core for Zen3 is essentially impossible (possibly outside of LN2). For ADL, it's more like 70W, though no stock CPU reaches those levels - but it scales much, much further in terms of clocks and power.

From what I've seen, it seems that Zen3 cores generally scale very well downwards - shown by their use in very low power mobile implementations where base clocks are still decent. Intel P cores seem to have a higher base power requirement, but in desktop CPUs this is offset in low threaded tasks by MCM Zen3's high uncore power (due to through-package IF).

fevgatosThe question is if it performs well enough with that power. We know that pcores are more efficient than ecores at everything when it comes to 5w and above per core. Therefore i really fail to see how a big full pcore cpu will have any problems with power draw.

Do we know that? Also, if that's the crossover point, your "16p > 8p8e @50W" thing seems rather shaky - that would give each of those P cores just 3.125W after all, assuming zero uncore power (so obviously less than that). That seems like it would be cutting things pretty close in terms of which is more efficient if that's the case.

Also: a big core only CPU would have the potential for problems with power draw as it will tend to boost too aggressively, even for small tasks. Most CPU boost algorithms, particularly Intel's ones, are tuned for race-to-finish efficiency. This is great for mobile tasks with moderate boost power, but falls apart for desktop parts with massive per-core boost power budgets. After all, a P core boosting to 50W would need to be 5x faster than an E core at 10W for it to match its efficiency. And we know a P core doesn't come close to 5x the performance of an E core, even at peak boost. This is one of two scenarios where E cores shine: they're fast enough and plentiful enough to handle potentially disturbing background tasks without consuming too much power. If you only have P cores, are running a high performance but low threaded foreground task, and a background task needs some work, one of two things happen: either you're power limited and the new core boosting will force the foreground cores to clock down, or you aren't power limited and the new core will increase power consumption for however long the background task takes to finish.

Is this a massive problem? No. Does it make an all-P core design less efficient outside of nT tasks? Yes. (And, of course, in nT tasks you run into the question of how many E cores you could have gotten for the same silicon cost within the same power envelope as those P cores, and how that would balance out - but that gets complicated very quickly due to boost powers, cache access, etc.)

ValantarThat is a gross oversimplification, and essentially entirely untrue. All CPU cores have a base power level required for them to function properly at all. They also have a lower threshold of clock speed and power where they become able to perform efficiently - where clocks and performance rise high enough to overcome the inherent disadvantage of that base power draw. For most modern CPU cores that clock speed is quite low, but it's still a threshold. And, crucially, all cores have some form of clock/power ceiling. For Zen3 that is ~21W, where scaling essentially stops past this. You can push more power into them, but clocks will stagnate and performance will drop. Exceeding 30W/core for Zen3 is essentially impossible (possibly outside of LN2). For ADL, it's more like 70W, though no stock CPU reaches those levels - but it scales much, much further in terms of clocks and power.

Doesnt make a difference really, this is a desktop cpu. The low threshold is what, just 2-3 watts per core. You ain't going to run a cpu at 35w

ValantarDo we know that? Also, if that's the crossover point, your "16p > 8p8e @50W" thing seems rather shaky - that would give each of those P cores just 3.125W after all, assuming zero uncore power (so obviously less than that). That seems like it would be cutting things pretty close in terms of which is more efficient if that's the case.

Depending on if the task is int or fp, but generally speaking anything above 2-3w per core pcores become more efficient than ecores. So yeah, we kinda know that

ValantarIs this a massive problem? No. Does it make an all-P core design less efficient outside of nT tasks? Yes. (And, of course, in nT tasks you run into the question of how many E cores you could have gotten for the same silicon cost within the same power envelope as those P cores, and how that would balance out - but that gets complicated very quickly due to boost powers, cache access, etc.)

But that is not the point. The point is the main or even only reason we dont see 16p cores is the die size. Wattage is really not a problem. I dont see how it could be, since we already know that pcores are more efficient than e cores in everything unless you drop down to a couple of watts per core, which you wont in a desktop.

Stay on topic.
Stop the insults.

PilgrimThis means that it's still going to be around 30% slower than 7950X

50% more cores than 12900 and it amazes with 30% more performance. Poor result IMO.

7950X3D will be a beast, although for me the 7900X3D is my sweet spot this gen.

Minus Infinity50% more cores than 12900 and it amazes with 30% more performance. Poor result IMO.

7950X3D will be a beast, although for me the 7900X3D is my sweet spot this gen.

33% more threads for 30% more performance. Great result actually