What is most realistic all core overclock I could get on a Ryzen 7 7700X and 7950X or 13900K/13700K with e-cores off using Noctua NH-D15

[AusWolf]

Cyberpunk is probably one of the worst games you could use as a reference. It's horribly optimized. But that aside, e-cores off impacting your frames might be a you problem.
An 11900K (8c16t) has no problems running any AAA titles. In fact, going based off Gamers Nexus, any 6c12t processor is well more than adequate to play AAA titles at GPU bottlenecks, with minimal frame drops.
To further iterate, a 5950X or 7950X has no problem operating with 50% (CCX) disabled. Margin of error for frame rate impact.

So either your frame rate issue (E-core off) is either and limiting factor of the chip, or issues with your build.
If I had to guess, Windows casuing it. Are you using Windows 10 or Windows 11?

I don't know much about Alder Lake and Raptor Lake, but I have to say, you have a point. If an 8-core AMD CPU, or an 11900/11700K can run the game, then why does dropping the E-cores hurt performance?

 

[JustBenching]

Of course. But it's to show how disabling cores, regardless of what processor you have, will allow for higher overclocks and lower voltage. And this is still true of disabling e-cores.
I think we can move on from that.



I mean, if all you play is DotA, and nothing else, technically a 13900K running on 2 P-cores at max frequency would yield highest frame rates. ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯



So 100mhz to 150mhz on my old R5 3600 was the difference between 40fps and 50fps for Mario Odyssey on Yuzu. You would be very surprised just now much it matters.
For your run of the mill Call Of Duty, not likely a noticeable difference.

Cyberpunk is probably one of the worst games you could use as a reference. It's horribly optimized. But that aside, e-cores off impacting your frames might be a you problem.
An 11900K (8c16t) has no problems running any AAA titles. In fact, going based off Gamers Nexus, any 6c12t processor is well more than adequate to play AAA titles at GPU bottlenecks, with minimal frame drops.
To further iterate, a 5950X or 7950X has no problem operating with 50% (CCX) disabled. Margin of error for frame rate impact.

So either your frame rate issue (E-core off) is either and limiting factor of the chip, or issues with your build.
If I had to guess, Windows casuing it. Are you using Windows 10 or Windows 11?



So I do have to ask, in the event I've missed something. But what exactly has been updated regarding Windows 10 and its big.LITTLE style management? I've checked very recently, and I can't seem to come up with anything other than Microsoft focussing on the issue for Windows 11.

Cyberpunk is horribly optimized? I stopped reading there, obviously you have no idea. Also, a 100 mhz taking you from 40 to 50 fps is not happening. 5% frequency gave you 25% performance?

 

[chrcoluk]

My advice would be keep e-cores on.

But use process hacker to lock your games to the p-cores on cpu affinity settings (at least for older games which probably have no scheduling logic).

You could also lock svchost.exe, explorer.exe and browsers to e-cores and that might give really good results.

 

[JustBenching]

I don't know much about Alder Lake and Raptor Lake, but I have to say, you have a point. If an 8-core AMD CPU, or an 11900/11700K can run the game, then why does dropping the E-cores hurt performance?

Silly question. Just because the game can use more than 8 cores, doesn't mean it won't work with less than 8 cores. There are areas of the game with lots of npcs that are really demanding, ,and your 11900 will struggle as hell there. Im talking below 60 fps, probably less. A 6 core 7600x hits 100% usage in this area, maxing out every single core

Check this, a 7950x droping to 60s

 

[AusWolf]

Check this, a 7950x droping to 60s

Still, CPU usage is around 35-50% which means that there's no need for more than 8 cores. Faster cores are generally better than more cores for gaming.

 

[JustBenching]

Still, CPU usage is around 35-50% which means that there's no need for more than 8 cores. Faster cores are generally better than more cores for gaming.

That has nothing to do with what you asked. Of course faster cores are better - duh, but 6 and 8 core current CPUs struggle at that area. Turning ecores off also struggles

 

[AusWolf]

That has nothing to do with what you asked. Of course faster cores are better - duh, but 6 and 8 core current CPUs struggle at that area. Turning ecores off also struggles

How do they struggle when CPU usage doesn't jump above 50% on a 16-core CPU?

 

[JustBenching]

How do they struggle when CPU usage doesn't jump above 50% on a 16-core CPU?

That's not how CPU usage works. You can't just read the % number.. Say you have a CPU with one core + HT. If the game uses the actual core at 100% and the HT at 0% cpu usage will show as 50% even though the CPU is almost maxed out

 

[AusWolf]

That's not how CPU usage works. You can't just read the % number.. Say you have a CPU with one core + HT. If the game uses the actual core at 100% and the HT at 0% cpu usage will show as 50% even though the CPU is almost maxed out

A fair point. Based on the video, CP77 doesn't seem to be using HT that much.

 

[JustBenching]

A fair point. Based on the video, CP77 doesn't seem to be using HT that much.

Check a 12900k and a 13900k with ecores on at the same area. CPU usage is nuts,

Turning ecores off destroys the performance at this area. Power consumption on the 13900k is 140 to 160w in this area, while the 12900k sits at 90 to 100 watts. I wanna sit those people that turn off ecores show us some performance and power draw in this area

 

[Vario]

Check a 12900k and a 13900k with ecores on at the same area. CPU usage is nuts,

Turning ecores off destroys the performance at this area. Power consumption on the 13900k is 140 to 160w in this area, while the 12900k sits at 90 to 100 watts. I wanna sit those people that turn off ecores show us some performance and power draw in this area

For these videos, Windows 10 or Windows 11?

 

[JustBenching]

For these videos, Windows 10 or Windows 11?

11

 

[Wolverine2349]

11

Windows 11 performance drops with e cores off because of thread director. Turn one e core on and no issue. In Windows 10 turn all e cores off and no such issue.

Core i7-12700K review revealing that E-cores hurt P-cores, but without them, everything is only worse

A review of the Core i7-12700K reveals that the presence of E-cores negatively impacts the performance of P-cores, but without them, the overall performance is even worse.

fox-laptop.com

Overclocking Raptor Lake - 13900k,13700k, 13600k etc...

www.overclock.net

Falkentybe discovered this and I confirmed their findings and happens only in WIN11 with ecores all off

 

[JustBenching]

Windows 11 performance drops with e cores off because of thread director. Turn one e core on and no issue. In Windows 10 turn all e cores off and no such issue.

Core i7-12700K review revealing that E-cores hurt P-cores, but without them, everything is only worse

A review of the Core i7-12700K reveals that the presence of E-cores negatively impacts the performance of P-cores, but without them, the overall performance is even worse.

fox-laptop.com

Overclocking Raptor Lake - 13900k,13700k, 13600k etc...

www.overclock.net

Falkentybe discovered this and I confirmed their findings and happens only in WIN11 with ecores all off

Why would I do that since ecores on performs better in games? LOL, im not a lunatic,

 

[Zero.exe]

Cyberpunk is horribly optimized? I stopped reading there, obviously you have no idea. Also, a 100 mhz taking you from 40 to 50 fps is not happening. 5% frequency gave you 25% performance?

So you don't believe it, so it must not be true. Right?
You claim you didn't read the rest, but ultimately that's an excuse to not answer to my points. Cause you know full well 8c16t processors are actually overkill for 99.99% of games. So disabling e-cores on 13700k/13900k, as my previous post guessed, is a you problem. You're projecting the poor performance of your system as a fact for everyone else's.
There is absolutely ZERO reason for E-core disabling to affect frame rates.

 

[JustBenching]

So you don't believe it, so it must not be true. Right?
You claim you didn't read the rest, but ultimately that's an excuse to not answer to my points. Cause you know full well 8c16t processors are actually overkill for 99.99% of games. So disabling e-cores on 13700k/13900k, as my previous post guessed, is a you problem. You're projecting the poor performance of your system as a fact for everyone else's.
There is absolutely ZERO reason for E-core disabling to affect frame rates.

It's not whether or not I believe it. It's literally IMPOSSIBLE for a 5% frequency to give you a 25% performance boost. It just CANNOT happen. It can't. You need to recall your statement, cause that's blatantly false / you are lying.

My system is perfectly fine, im running ghost windows and everything is tuned to the bone. As you can see from the posted videos. The game requires more than 8 cores at those heavy areas, so obviously it performs better with ecores on. Nothing more, nothing less.

 

[Zero.exe]

Why would I do that since ecores on performs better in games? LOL, im not a lunatic

He literally explained the issue to you.
But here. A credible source showing you you're wrong.

As many have commented on these YouTube videos, the video uploaders tend to forget with E-cores off you can up the ring ratio to 47 which will truly show the advantage of E-Cores off.

It's not whether or not I believe it. It's literally IMPOSSIBLE for a 5% frequency to give you a 25% performance boost. It just CANNOT happen. It can't. You need to recall your statement, cause that's blatantly false / you are lying.

Ryzen 3600 stock boosts to 4.2Ghz. At stock, it would typically hover around 40fps. Sometimes 42. Sometimes 39.
Setting a manual 4.35Ghz netted me minimum 46fps, and up to 50fps.

Yuzu is heavily clock dependant. Having a static clock above boost clock will net you a fair bit more frames.
Emulation doesn't work like a video game LOL.

As you can see from the posted videos. The game requires more than 8 cores at those heavy areas, so obviously it performs better with ecores on. Nothing more, nothing less.

You posted the same game over and over. Cyberpunk, which is the most memed game for it's glitches and poor optimization. Are you for reel?

Anyways I gave you Hardware unboxed video who tested in multiple games for fair comparison.
And, I don't think he even pushed the ring ratio which would further the discrepancy the between E-core On/Off LMAO

 

[Wolverine2349]

It's not whether or not I believe it. It's literally IMPOSSIBLE for a 5% frequency to give you a 25% performance boost. It just CANNOT happen. It can't. You need to recall your statement, cause that's blatantly false / you are lying.

My system is perfectly fine, im running ghost windows and everything is tuned to the bone. As you can see from the posted videos. The game requires more than 8 cores at those heavy areas, so obviously it performs better with ecores on. Nothing more, nothing less.

Requires more than 8 cores LMAO. No game actually requires more than 6 let alone 8. Heck the recommended system requirements do not even state more than a quad core with HT:

Cyberpunk requirements for PC - Can I Run It? - PCGameBenchmark

Cyberpunk 2077 system requirements 2024 - can your PC run Cyberpunk 2077? What is your PCGameBenchmark rating? #pcgaming

www.pcgamebenchmark.com

Not to say that 4 cores is enough for best experience, but still. There are 0 games that require more than 6 cores. Now whether they can benefit is another matter, but required lol.

Very few if any benefit from more than 8 cores and even those that do the benefit is tiny at best and almost margin of error.

 

[JustBenching]

Requires more than 8 cores LMAO. No game actually requires more than 6 let alone 8. Heck the recommended system requirements do not even state more than a quad core with HT:

Cyberpunk requirements for PC - Can I Run It? - PCGameBenchmark

Cyberpunk 2077 system requirements 2024 - can your PC run Cyberpunk 2077? What is your PCGameBenchmark rating? #pcgaming

www.pcgamebenchmark.com

Not to say that 4 cores is enough for best experience, but still. There are 0 games that require more than 6 cores. Now whether they can benefit is another matter, but required lol.

Very few if any benefit from more than 8 cores and even those that do the benefit is tiny at best and almost margin of error.

K, go get more 60 fps in toms dinner with a 6 core cpu. Good luck. Even the 5800x 3d drops to the 50ies in that area. Why don't you shows us the results with your E cores off 13900

He literally explained the issue to you.
But here. A credible source showing you you're wrong.

As many have commented on these YouTube videos, the video uploaders tend to forget with E-cores off you can up the ring ratio to 47 which will truly show the advantage of E-Cores off.


Ryzen 3600 stock boosts to 4.2Ghz. At stock, it would typically hover around 40fps. Sometimes 42. Sometimes 39.
Setting a manual 4.35Ghz netted me minimum 46fps, and up to 50fps.

Yuzu is heavily clock dependant. Having a static clock above boost clock will net you a fair bit more frames.
Emulation doesn't work like a video game LOL.


You posted the same game over and over. Cyberpunk, which is the most memed game for it's glitches and poor optimization. Are you for reel?

Anyways I gave you Hardware unboxed video who tested in multiple games for fair comparison.
And, I don't think he even pushed the ring ratio which would further the discrepancy the between E-core On/Off LMAO

Hwunboxed tests with rt off. Rt is incredibly heavy on the cpu. Anyways, it's obvious you havent played the game so you are just guessing, no point arguing with you. Go test it first and then see what happens in heavy areas when you turn off ecores.

 

[Wolverine2349]

K, go get more 60 fps in toms dinner with a 6 core cpu. Good luck. Even the 5800x 3d drops to the 50ies in that area. Why don't you shows us the results with your E cores off 13900


Hwunboxed tests with rt off. Rt is incredibly heavy on the cpu. Anyways, it's obvious you havent played the game so you are just guessing, no point arguing with you. Go test it first and then see what happens in heavy areas when you turn off ecores.

Did you ever test it on Windows 10 with all e-cores off?

It is known turning off all e-cores in WIN11 hurts performance because WIN11 uses thread director and thread director is messed up with all e-cores off and no way to shut it off in WIN11 with 12th nor 13th gen. WIN10 does not use thread director so no such issue all eocres off.

If you have only used WIN11, better to test with 1 ecore on and the rets off rather than all off as WIN11 and thread director messes up P core performance and hyper threading when no e-cores are active. Turning on 1 ecore could test whether the game really uses any ecores or not as 1 ecore is not going to really help much with anything.

 

[JustBenching]

I just tested warzone 2 on a 12900k.

Ecores off 5.4 all core / 4.9 cache, 1% lows were between 140 - 170

Ecores on stock cpu 1% lows = 153 - 186

Ecores on 5.4ghz / 4.2 cache / 4.0 ecores 1% lows = 170 - 230

When you run with only 8 cores the game sends threads to the HT hurting performance. With ecores on game doesn't use HT at all, boosting 1% lows substantially.

There you go lads, don't turn off your ecores

 

[Zero.exe]

K, go get more 60 fps in toms dinner with a 6 core cpu. Good luck. Even the 5800x 3d drops to the 50ies in that area. Why don't you shows us the results with your E cores off 13900

Well thank goodness I don't play cyberpunk, so I'll never have to worry about frame drops with e-cores off .

Hwunboxed tests with rt off. Rt is incredibly heavy on the cpu. Anyways, it's obvious you havent played the game so you are just guessing, no point arguing with you. Go test it first and then see what happens in heavy areas when you turn off ecores.

HWU tested fairly. Multiple games and scenarios. You keep bringing up the one game on Win 11.
Further, you didn't test like Wolverine2349 asked you to.
I'm done with this. Enjoy your CPU to the fullest and your game. Cheers.

 

[Wolverine2349]

I just tested warzone 2 on a 12900k.

Ecores off 5.4 all core / 4.9 cache, 1% lows were between 140 - 170

Ecores on stock cpu 1% lows = 153 - 186

Ecores on 5.4ghz / 4.2 cache / 4.0 ecores 1% lows = 170 - 230

When you run with only 8 cores the game sends threads to the HT hurting performance. With ecores on game doesn't use HT at all, boosting 1% lows substantially.

There you go lads, don't turn off your ecores

I would agree now that it is now pointless to disable e-cores especially since you can just get a 7700X or 7800X3D if you just want 8 cores with no e-cores.

Though is the Intel Thread director supposed to work seamlessly and make e-cores work well with all secondary threads in games so they work just as well if not better than if there were additional but less P cores. Like if a game thread beyond 8 (HT off) would use 10-30% intermittently of an additional P core, does thread director know how to make it use e-core the same performance cause e-cores are weaker so instead it makes it use say 60% of e-core and no slow down or waiting for process. Does it work without developers having o developed with hybrid arch in mind cause thread director takes care of everything or at least in 99% of cases but maybe some rare edge cases?

Back to original topic though. My question about overclocking headroom on Intel CPUs. With the recent issues it appears there is none and when it appears stable down the road its not. ANd even at stock with 13th and 4th Gen there are random Intel stability issues and there are theories about fast degradation, PCH being too weak to handle high end parts, voltage and current applied too much out of box or even with limits enforced though not as bad but still an issue.

Which does lead me to interest in Arrow Lake to get all P and e-cores on same tile for good latency and hopefully its process node 20A and/or TSMC will have all stability issues fixed.

Cause games are becoming more threaded when paired with high end GPU espeically even at 4K. Though 8 cores still plenty for most that could be starting to change. And AMD has the dual CCD design which is not good for games and 8 cores max per CCD/CCX not good with cross latency penalty and that not changing with Zen 5. and maybe even Zen 6 as well at least on mainstream platforms. SO Arrow Lake is where it may be when it comes out especially if you wee to shut HT off anyways and hipe for the perfect P and e core system for threaded stuff. ANd hopefully lower power consumption too.