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

[Zero.exe]

Hope I'm not too late to the party, but wanted to add some thoughtful points:

@fevgatos look at what advantages do exist:
-Dropping the e-cores does account to lower power consumption (package total).
-Possibility of running P cores at lower vcore, as @GerKNG said: "P Cores needed around 1.33V (Load Vcore) and without the E-Cores the P Cores were stable at 1.30V"
-Possibility to OC P cores to 6.0Ghz with vcore headroom.
-Not having to use Windows 11

So what advantage is there to all that? Emulation for starters. Yuzu/RPCS3 will heavily favour higher clocks. 13900K reigns king in those two specifically for it's higher clocks (waiting for Ryzen vChache CPUs to see if there's any change there).

As for games, it will be very much per game basis. DotA for example absolutely will favour a higher P-Core clock. Many games still use 4 cores. For 99.99% of games in existence, having 16 threads with just P cores is more than enough to run the game, mitigate background tasks, run all drivers including audio, and even stream the game with additional software without having a meaningful impact to frame rates.

For the most part, we are not at the point where having 32 cores makes a difference in Gaming. But it does make a difference in rendering programs or video encoding with handbrake.

I totally agree with you @Wolverine2349 on how you run your system as it suits your needs. I am currently contemplating upgrading from a my 5600(nonX) (OCd to 4.8Ghz all core btw for those who say OCing isn't viable). I would personally benefit from the maximum sustainable OC of 4-8 cores, so 13900k has been on my radar.
But 7950X also achieves above 5.5Ghz, and with vCache arriving, this also is a solid choice. It's all about those clocks!

 

[JustBenching]

@fevgatos look at what advantages do exist:
-Dropping the e-cores does account to lower power consumption (package total).
-Possibility of running P cores at lower vcore, as @GerKNG said: "P Cores needed around 1.33V (Load Vcore) and without the E-Cores the P Cores were stable at 1.30V"
-Possibility to OC P cores to 6.0Ghz with vcore headroom.
-Not having to use Windows 11

1) Ι wish it did drop power consumption, it doesn't - not in games. Unless you are talking about MT workloads, but then you are losing way more performance then you are dropping power consumption

2) That's entirely based on your specific cpu. If your P cores are better binned than your E cores, sure, but that's also easily fixable by running your ecores 100mhz lower. No reason to turn them off

3) You can't OC them. I mean, ok you can, but the CPU already draws an insane amount of power during games, trying to go over the stock clocks is just suicide. 6.0ghz is of course completely out of the question, but even 5.8 - you are going to be drawing 200+w in GAMES. Plus, whether you have ecores on or off doesn't change your possibility of ocing the pcores.

4) Windows 10 work just fine

 

[Zero.exe]

@fevgatos
Point 1) I think is up for debate. I don't have a 13XX cpu, otherwise I would conduct my own test. I do know modern game engines will offset loads across the chip (at least on Win 10) so I would hazard that there still is more power usage. I could be wrong here. Perhaps someone with said chips could try running a AAA title and some background apps (streaming, discord etc) with e-cores on and off to verify.

Point 2) As others have stated, in overclock scenarios, disabled e-cores allows lower vcore usage on P-Core. Deny it all you want. I don't think @GerKNG or @Wolverine2349 have a reason to lie about it.
Lowering your e-core clock still has active cores on the package. Until more data is provided, it certainly seems disabling them provides higher flexibility for P-Core vcore.

Point 3) You can overclock anything. 6.0 GHz is doable with delidding + LM + LF2 420, on a top 3% bin. Otherwise 5.8 is doable. 5.9 with enough effort. Games don't max all cores so you just need to be stable enough. Don't have to be prime95 stable. As long as you're not BSODing or doing sensitive work with potential data loss.

Point 4) Win 10 still prioritizing e-cores. There hasn't been a fix for it. I'll pass on Win 11.

 

[JustBenching]

@fevgatos
Point 1) I think is up for debate. I don't have a 13XX cpu, otherwise I would conduct my own test. I do know modern game engines will offset loads across the chip (at least on Win 10) so I would hazard that there still is more power usage. I could be wrong here. Perhaps someone with said chips could try running a AAA title and some background apps (streaming, discord etc) with e-cores on and off to verify.

Point 2) As others have stated, in overclock scenarios, disabled e-cores allows lower vcore usage on P-Core. Deny it all you want. I don't think @GerKNG or @Wolverine2349 have a reason to lie about it.
Lowering your e-core clock still has active cores on the package. Until more data is provided, it certainly seems disabling them provides higher flexibility for P-Core vcore.

Point 3) You can overclock anything. 6.0 GHz is doable with delidding + LM + LF2 420, on a top 3% bin. Otherwise 5.8 is doable. 5.9 with enough effort. Games don't max all cores so you just need to be stable enough. Don't have to be prime95 stable. As long as you're not BSODing or doing sensitive work with potential data loss.

Point 4) Win 10 still prioritizing e-cores. There hasn't been a fix for it. I'll pass on Win 11.

I tested it in cyberpunk, power draw is similar but performance tanks with ecores off at specific areas of the game.

What do you mean "deny it all you want". Ecores being active doesn't change the behavior off the pcores. I have both a 12900k and a 13900k, it doesn't make any difference to ocing.

So not only are you fine with running unstable configurations, but you are also fine with drawing over 200w for GAMES. Lol man

 

[phanbuey]

I tested it in cyberpunk, power draw is similar but performance tanks with ecores off at specific areas of the game.

What do you mean "deny it all you want". Ecores being active doesn't change the behavior off the pcores. I have both a 12900k and a 13900k, it doesn't make any difference to ocing.

So not only are you fine with running unstable configurations, but you are also fine with drawing over 200w for GAMES. Lol man

+1

I can vouch for this. Performance is less consistent with e cores off.

RTX 4090 & 53 Games: Core i9-13900K E-Cores Enabled vs Disabled Review | TechPowerUp

"We would still recommend that the E-cores be left untouched, for the simple reason that the P-cores don't appear to be gaining from the freed up power budget of the processor to sustain boost clocks better, as hypothesized in the article's introduction. Leaving the E-cores enabled benefits games, but the OS itself can dump low-priority background processes (think Antivirus, firewall, etc.,) as well as audio and networking stacks, to the E-cores (which have plenty of compute power to deal with them)"

 

[Wolverine2349]

Of course you're free to disable E-cores for higher P-core OC, but E-cores are here to stay whether you like them or not.

The bottom line, E-cores are free cache and Intel's lifeline to staying competitive in gaming performance. Uncore portion of each Gracemont cluster includes the same 3MB L3 slice that each P-core has - remove them from the die entirely and you're down to a 24MB 8-core and it's Rocket Lake vs. Comet Lake all over again, a massive core doesn't mean much in games if it's starved for cache.

And yet you can disable all e-cores on 13900k or KS and the 8 p cores still have all 36MB L3 cache free to them. Perfect for cached starved games. Why it makes sense to get 13900k over 13700k even if disabling all e cores because of extra L3 cache. It's intel wafer design that forces e cores to be there more more L3 cache despite the fact you can disabke e cores with no loss of L3 cache to P cores. In fact better as p cores have full access to L3 cache themselves.

+1

I can vouch for this. Performance is less consistent with e cores off.

RTX 4090 & 53 Games: Core i9-13900K E-Cores Enabled vs Disabled Review | TechPowerUp

"We would still recommend that the E-cores be left untouched, for the simple reason that the P-cores don't appear to be gaining from the freed up power budget of the processor to sustain boost clocks better, as hypothesized in the article's introduction. Leaving the E-cores enabled benefits games, but the OS itself can dump low-priority background processes (think Antivirus, firewall, etc.,) as well as audio and networking stacks, to the E-cores (which have plenty of compute power to deal with them)"

Only with lots of background tasks and in WIN11. I do not run heavy background tasks. What I gave is an intel equivalent to 7800X3D both 8 cores. Intel has better IPC and clock speeds. AMD has more cache. Both great for gaming. Though intel can use well tuned DDR4 4000MHz dual rank CL15 Samsung b.die rock stable. Where as AMD need buggy XMP DDR5 which I have had nothing but trouble with.

And almost 0 games benefit at all from more than 8 cores. Plus I debloat Windows 10 so little background tasks so e cores do nothing to help games.

 

[Vario]

Windows 10 is fine, I use it with my 12900KS and no issues. E cores are fine. I like having them. Seems to improve multithread performance, no downside.

Answering OP's question, I think about 5.2 all core was doable with my 12900KS on the NHD15S, I have mine arbitrarily limited to 5.0GHz all core, 1.2Vcore with NHD15S. This keeps my core temperatures in games at 65C max. typically mid 50s on average. Games such as GTAV, Space Engineers.

 

[phanbuey]

And yet you can disable all e-cores on 13900k or KS and the 8 p cores still have all 36MB L3 cache free to them. Perfect for cached starved games. Why it makes sense to get 13900k over 13700k even if disabling all e cores because of extra L3 cache. It's intel wafer design that forces e cores to be there more more L3 cache despite the fact you can disabke e cores with no loss of L3 cache to P cores. In fact better as p cores have full access to L3 cache themselves.


Only with lots of background tasks and in WIN11. I do not run heavy background tasks. What I gave is an intel equivalent to 7800X3D both 8 cores. Intel has better IPC and clock speeds. AMD has more cache. Both great for gaming. Though intel can use well tuned DDR4 4000MHz dual rank CL15 Samsung b.die rock stable. Where as AMD need buggy XMP DDR5 which I have had nothing but trouble with.

And almost 0 games benefit at all from more than 8 cores. Plus I debloat Windows 10 so little background tasks so e cores do nothing to help games.

The debloat works wonders tbh -- it's amazing how one can spend 100's of hours FAFOing in the bios then leave steam overlay, VBS, game bar, and a bunch of stuff in the background and just lose 20% performance.

I'm too lazy to close down my workbench so for me it's an effort decision but I would love to have a proper very aggressive "game mode" where the os just goes into hardcore minimal mode during a game.

 

[Wolverine2349]

Well I got a 13900ks now. Can do 5.6GHz all core with 1.26v LLC6 5GHz ring e cores off. Passes all stress tests except did not try Y crunchet set nor Prime95 with AVX on as those beyond torture CPUs and will cook it to throttling. This on be quiet dark rock pro 4 fans at only 1000 rpm. Temps never exceed high 80s.

I could also do 5.6GHz on regular k, but needed 1.34v LLC6 and temps on tougher tests got to mid to high 90s. The better binning helps for sure on ks.

 

[phanbuey]

mine takes like 1.44v for 5.7 LOL ... In an ITX case.

 

[JustBenching]

E cores on if you want to game. E cores off if you want to pretend you gain anything from turning them off

 

[ir_cow]

E cores on if you want to game. E cores off if you want to pretend you gain anything from turning them off

100%

 

[D007]

13900k, Hmm all Ecores off under air?
Only one way to find out. CPU contact frame and Liquid Metal will help but that Air cooler is shooting you in the foot big time.
I’d say 5.7 tops. But that depends. If you stress test it you’re going to thermal throttle instantly. In gaming it would likely be pretty stable as long as ambient temps are good.
But would disabling the Ecores actually get you better performance in the desired applications? I’d test both configurations to find out.
That’s part of the fun of overclocking anyway

 

[Wolverine2349]

Hmm all Ecores off under air?
Only one way to find out. CPU contact frame and Liquid Metal will help but that Air cooler is shooting you in the foot big time.
I’d say 5.7 tops. But that depends. If you stress test it you’re going to thermal throttle instantly. In gaming it would likely be pretty stable as long as ambient temps are good.
But would disabling the Ecores actually get you better performance in the desired applications? I’d test both configurations to find out.
That’s part of the fun of overclocking anyway

Under air e cores off with Dark Rock Pro 4. 5.6GHz all core on 13900ks at 1.26v LLC6 ring 5GHz. Both fans on Dark Rock Pro 1000 RPM as well as all 4 case fans.

Ambient room temp is 19.5C

Temps For CPU under Cinebench R23 mid 80s max.

Passes OCCT Large Data Set

Passes all Y cruncher except ommited SFT

Passes Prime95 all FFT with AVX OFF

Passes multiple Linpack Xtreme runs

I omitted Prime95 AVX as that runs power and heat through roof especially on intel CPUs

Also ommited Y cruncher SFT as that beyond runs power and heat through roof on any CPU.

If I pass all other tests I call it stable.

Though linpack XTREME and y cruncher BBPNalso very rough they are much more tame than ommited tests.

And it runs great and no stupid e cores to worry about. Would not have any chance 5.6GHz all p cores to run with respectable thermals and that low 1.26vcore. I never wanted them so they stay off and I gain thermal and oc headroom for the strong cores.

It's the best 8 core CPU in existence.

Hmm all Ecores off under air?
Only one way to find out. CPU contact frame and Liquid Metal will help but that Air cooler is shooting you in the foot big time.
I’d say 5.7 tops. But that depends. If you stress test it you’re going to thermal throttle instantly. In gaming it would likely be pretty stable as long as ambient temps are good.
But would disabling the Ecores actually get you better performance in the desired applications? I’d test both configurations to find out.
That’s part of the fun of overclocking anyway

Under air e cores off with Dark Rock Pro 4. 5.6GHz all core on 13900ks at 1.26v LLC6 ring 5GHz. Both fans on Dark Rock Pro 1000 RPM as well as all 4 case fans.

Ambient room temp is 19.5C

Temps For CPU under Cinebench R23 mid 80s max.

Passes OCCT Large Data Set

Passes all Y cruncher except ommited SFT

Passes Prime95 all FFT with AVX OFF

Passes multiple Linpack Xtreme runs

I omitted Prime95 AVX as that runs power and heat through roof especially on intel CPUs

Also ommited Y cruncher SFT as that beyond runs power and heat through roof on any CPU.

If I pass all other tests I call it stable.

Though linpack XTREME and y cruncher BBPNalso very rough they are much more tame than ommited tests.

And it runs great and no stupid e cores to worry about. Would not have any chance 5.6GHz all p cores to run with respectable thermals and that low 1.26vcore. I never wanted them so they stay off and I gain thermal and oc headroom for the strong cores.

Though there is a hard wall at 5.6GHz on both 13900k and KS. Could never get 5.7GHz fully stable at any reasonable voltage e cores off of course. Even 1.33v temps got a bit high bit still within spec and thought it was stable after it passed lots of tests so thought I was good. Come back and run Cinebench again and WHEA. Back down to 5.6 and NEVER WHEA and lower vcore and temps by far.

It's the best 8 core CPU in existence

 

[D007]

Under air e cores off with Dark Rock Pro 4. 5.6GHz all core on 13900ks at 1.26v LLC6 ring 5GHz. Both fans on Dark Rock Pro 1000 RPM as well as all 4 case fans.

Ambient room temp is 19.5C

Temps For CPU under Cinebench R23 mid 80s max.

Passes OCCT Large Data Set

Passes all Y cruncher except ommited SFT

Passes Prime95 all FFT with AVX OFF

Passes multiple Linpack Xtreme runs

I omitted Prime95 AVX as that runs power and heat through roof especially on intel CPUs

Also ommited Y cruncher SFT as that beyond runs power and heat through roof on any CPU.

If I pass all other tests I call it stable.

Though linpack XTREME and y cruncher BBPNalso very rough they are much more tame than ommited tests.

And it runs great and no stupid e cores to worry about. Would not have any chance 5.6GHz all p cores to run with respectable thermals and that low 1.26vcore. I never wanted them so they stay off and I gain thermal and oc headroom for the strong cores.

It's the best 8 core CPU in existence.



Under air e cores off with Dark Rock Pro 4. 5.6GHz all core on 13900ks at 1.26v LLC6 ring 5GHz. Both fans on Dark Rock Pro 1000 RPM as well as all 4 case fans.

Ambient room temp is 19.5C

Temps For CPU under Cinebench R23 mid 80s max.

Passes OCCT Large Data Set

Passes all Y cruncher except ommited SFT

Passes Prime95 all FFT with AVX OFF

Passes multiple Linpack Xtreme runs

I omitted Prime95 AVX as that runs power and heat through roof especially on intel CPUs

Also ommited Y cruncher SFT as that beyond runs power and heat through roof on any CPU.

If I pass all other tests I call it stable.

Though linpack XTREME and y cruncher BBPNalso very rough they are much more tame than ommited tests.

And it runs great and no stupid e cores to worry about. Would not have any chance 5.6GHz all p cores to run with respectable thermals and that low 1.26vcore. I never wanted them so they stay off and I gain thermal and oc headroom for the strong cores.

Though there is a hard wall at 5.6GHz on both 13900k and KS. Could never get 5.7GHz fully stable at any reasonable voltage e cores off of course. Even 1.33v temps got a bit high bit still within spec and thought it was stable after it passed lots of tests so thought I was good. Come back and run Cinebench again and WHEA. Back down to 5.6 and NEVER WHEA and lower vcore and temps by far.

It's the best 8 core CPU in existence

I’m running 6.1 when 2 are active, 6.0 on 4, 5.9 on 4-8 PCore active.
EKWB velocity v2. Ecores at 4.6 and stable.
0 AVX offset.
In testing at 100% it throttles to 5.8-5.7 but that’s due to TVB, not thermals reaching 100c. It doesn’t hit 100c.
However what OS are you running and are you sure what you are doing increases performance? Did you test the difference? I’m curious.

If you bump the SA voltage to 1.3, you could probably hit 5.1 on the ring. Might see if it gives you more performance without much thermal increase.
Maybe raise the PL4 ring offset to +15 mv.

Did you check HWinfo64 performance limiting reasons to see if anything is holding you down?
Also in testing does it drop the core frequency at all or stay at 5.6 all core?
Might consider trying what I did. 5.9 if 2 active, 5.8 if 4, 5.7 if 6, 5.6 if 8. You can play with those numbers a bit.
worth a shot.
You could get lower temps and extra OC headroom if you try setting the Vcore voltage mode to adaptive and set 6-10 to -.040mv. See if it’s still stable, if it is, drop it to -.050, -.060 until unstable then step back.

 

[Zero.exe]

What do you mean "deny it all you want". Ecores being active doesn't change the behavior off the pcores. I have both a 12900k and a 13900k, it doesn't make any difference to ocing.

Two members here testify that e-cores being inactive allows for lower vcore on p cores. This doesn't surprise me.
I am being as clear and concise as I possibly can. You seem to gloss over that and re state your point.
Again, for greater clarification, by disabling E cores, you can lower P core voltage.

Do you know how the world record overclocks are achieved?
They (LN2 clockers) used to use Intel i3's and Disable All but One core.
Why do you suppose that is?

Answer: It's a shared package. The less cores you have active, the greater clocks you can achieve with greater stability. And it's still done that way today.

 

[JustBenching]

Two members here testify that e-cores being inactive allows for lower vcore on p cores. This doesn't surprise me.
I am being as clear and concise as I possibly can. You seem to gloss over that and re state your point.
Again, for greater clarification, by disabling E cores, you can lower P core voltage.

Do you know how the world record overclocks are achieved?
They (LN2 clockers) used to use Intel i3's and Disable All but One core.
Why do you suppose that is?

Answer: It's a shared package. The less cores you have active, the greater clocks you can achieve with greater stability. And it's still done that way today.

First of all, there is stability in ln2 overclocks. Those systems are obviously not stable, you can just take a screenshot and validate the result, that's it.

Disabling Pcores can also give you higher stable overclocks. For example my core 3 and 4 are the worst cores requiring the most voltage for a set frequency Should I disable them as well? Should I only keep the 2 preferred cores active for the maximum clocks? Long story short, I haven't noticed any difference in ocing with ecores on or off, but even if it's there it's just not worth it. Youll get what, 100mhz higher clocks, and youll sacrifice a ton of performance even in games. Cyberpunk, in heavy areas of the game, ecores off drops framerate substantially for example

 

[Wolverine2349]

I’m running 6.1 when 2 are active, 6.0 on 4, 5.9 on 4-8 PCore active.
EKWB velocity v2. Ecores at 4.6 and stable.
0 AVX offset.
In testing at 100% it throttles to 5.8-5.7 but that’s due to TVB, not thermals reaching 100c. It doesn’t hit 100c.
However what OS are you running and are you sure what you are doing increases performance? Did you test the difference? I’m curious.

If you bump the SA voltage to 1.3, you could probably hit 5.1 on the ring. Might see if it gives you more performance without much thermal increase.
Maybe raise the PL4 ring offset to +15 mv.

Did you check HWinfo64 performance limiting reasons to see if anything is holding you down?
Also in testing does it drop the core frequency at all or stay at 5.6 all core?
Might consider trying what I did. 5.9 if 2 active, 5.8 if 4, 5.7 if 6, 5.6 if 8. You can play with those numbers a bit.
worth a shot.
You could get lower temps and extra OC headroom if you try setting the Vcore voltage mode to adaptive and set 6-10 to -.040mv. See if it’s still stable, if it is, drop it to -.050, -.060 until unstable then step back.

Mine stays at 5.6GHz all core in HWInfo64. I do a static all core all the time overclock. No AVX offset. 5.6GHz with AVX2 on

If I left all at stock, running CInebench R23, it throttles right away as it hits 100C and that is with P cortes clocked at like 5.4GH. Even OCCT Large Data set variable tmeps get close to throttling.

With e-cores off and manual voltage 5.6GHz all P cores, temps are in 60s and peak in 70s in OCCT. Temps average and peak in low 80s in Ceinbench R23. Plus ring clock can hit 5GHz easily.

E-cores on not a chance at 5GHz ring. Heck I cannot hit 5.1GHz ring with e-cores off even for full stability.

And I set manually static vcore. At 5.6Ghz, at 1.25v, it was close to fully stable, but BSOD on a random Linpack XTREME run of 10 tests liker maybe test 3 or 4. Upped it to 1.26 and perfectly stable after multiplke runs of all those tests,

Only skipped AVX in Prime95 by checking the Disable AVX check box and skipped Y Cruncher SFT as those tests are so beyond murderous for CPUs. Even Y Cruncher BBP and Linpack XTREME are rough and CInebench R23 is somewhat tough and puts out more heat than some parts of the others, but few are anything like Linpack XTREME or Y Cruncher BBP let alone the moderate to far worse than even them Prime95 with AVX or Y Cruncher SFT.

 

[D007]

Mine stays at 5.6GHz all core in HWInfo64. I do a static all core all the time overclock. No AVX offset. 5.6GHz with AVX2 on

If I left all at stock, running CInebench R23, it throttles right away as it hits 100C and that is with P cortes clocked at like 5.4GH. Even OCCT Large Data set variable tmeps get close to throttling.

With e-cores off and manual voltage 5.6GHz all P cores, temps are in 60s and peak in 70s in OCCT. Temps average and peak in low 80s in Ceinbench R23. Plus ring clock can hit 5GHz easily.

E-cores on not a chance at 5GHz ring. Heck I cannot hit 5.1GHz ring with e-cores off even for full stability.

And I set manually static vcore. At 5.6Ghz, at 1.25v, it was close to fully stable, but BSOD on a random Linpack XTREME run of 10 tests liker maybe test 3 or 4. Upped it to 1.26 and perfectly stable after multiplke runs of all those tests,

Only skipped AVX in Prime95 by checking the Disable AVX check box and skipped Y Cruncher SFT as those tests are so beyond murderous for CPUs. Even Y Cruncher BBP and Linpack XTREME are rough and CInebench R23 is somewhat tough and puts out more heat than some parts of the others, but few are anything like Linpack XTREME or Y Cruncher BBP let alone the moderate to far worse than even them Prime95 with AVX or Y Cruncher SFT.

Ok but does it perform better?

 

[Wolverine2349]

Ok but does it perform better?

Do you have e-cores on or off?

As for using adaptive voltage, does it work well keeping a static clock speed. Almost afraid to try it on air as it was such a pain to get a stable system to begin with with reasonable temps.

 

[D007]

Do you have e-cores on or off?

As for using adaptive voltage, does it work well keeping a static clock speed. Almost afraid to try it on air as it was such a pain to get a stable system to begin with with reasonable temps.

If you’re happy with how it is, it’s really up to you lol. But did you test before and after with Ecores on and off to see if there is a performance improvement? If not, you may of actually made things worse for yourself. Worth checking.

 

[Wolverine2349]

If you’re happy with how it is, it’s really up to you lol. But did you test before and after with Ecores on and off to see if there is a performance improvement? If not, you may of actually made things worse for yourself. Worth checking.

I have tested 12th gen with and without e-cores and e-cores on it was choppy in many games. With them off smooth as silk.

13th gen is no different. Just faster P cores and more e-cores. And more power consumption so would not stand a chance on air with high clocked P cores and also e-cores on anyways.

Many games do not like hybrid arch at all. Have to user WIN11 or if using WIN10, need to use Process Lasso or the wrong stuff can get assigned to e-cores as WIN10 does not park e-cores but parks the P cores. WIN11 parks them all I have seen.

 

[D007]

I have tested 12th gen with and without e-cores and e-cores on it was choppy in many games. With them off smooth as silk.

13th gen is no different. Just faster P cores and more e-cores. And more power consumption so would not stand a chance on air with high clocked P cores and also e-cores on anyways.

Many games do not like hybrid arch at all. Have to user WIN11 or if using WIN10, need to use Process Lasso or the wrong stuff can get assigned to e-cores as WIN10 does not park e-cores but parks the P cores. WIN11 parks them all I have seen.

A lot has changed with the way windows 10 handles Ecores and there is a great review here with 52 games included that shows from my recollection that you generally hurt performance in most cases disabling Ecores. I can’t recall if it was Win 10 or 11 but Worth a look at. That’s why I was curious about your befire and after performance differences but from the sound of it, you did not test it and seem not to want to lol. Which is fine but I was trying to figure out if I should try this in Win 11, OC the hell out of the PCores, maybe leave 2 Ecores enabled for some scheduling needs of the OS and see what happens. All good. Enjoy

 

[Wolverine2349]

A lot has changed with the way windows 10 handles Ecores and there is a great review here with 52 games included that shows from my recollection that you generally hurt performance in most cases disabling Ecores. I can’t recall if it was Win 10 or 11 but Worth a look at. That’s why I was curious about your befire and after performance differences but from the sound of it, you did not test it and seem not to want to lol. Which is fine but I was trying to figure out if I should try this in Win 11, OC the hell out of the PCores, maybe leave 2 Ecores enabled for some scheduling needs of the OS and see what happens. All good. Enjoy

Windows 10 leaving any e-cores on does nothing and disabling does not hurt gaming at all and in fact could help it a bit or makes no difference
.

In Windows 11, turning them off can actually hurt performance of P cores with hyper threading on because the thread director not knowing how to deal with physical and logical cores per link below

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

And there is no way to disable thread director in WIN11 so you need at least 1 e-core on or pe4rformance is erratic with P cores as I once tested with CPU-Z and score all over the place with all e-cores off. In WIN10 21H2 no such issue as it is not thread director aware so it is effectively disabled and CPU-Z single thread score consistent proving the point as WIIN10 not being able to use thread director on 12th and 13th Gen means it treats a 13900K with e-cores off as a normal 8 core 16 thread CPU.

Though some say e-cores work fine along with P cores on WIN10 and that may be true because they technically do not need thread director to still perform well, but you will not get best experience with it., Need WIN11 for that, though does not mean WIN10 is broken with it.

But WIN10 with them off great performance or with them on still great performance and better for infinite threaded apps if cooling can handle it.

WIN11 best for stock 12th and 13th gen. With all e-cores off and HT on a problem unless you leave one e-core enabled due to link above. Unless there comes a switch to disable thread director on 12th and 13th Gen or WIN11.

I am staying with my WIN10 21H2 2021 LTSC and all e-cores off. Best 8 core chip there is and no hybrid stuff to deal with.

 

[Zero.exe]

First of all, there is stability in ln2 overclocks. Those systems are obviously not stable, you can just take a screenshot and validate the result, that's it.

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.

Disabling Pcores can also give you higher stable overclocks. For example my core 3 and 4 are the worst cores requiring the most voltage for a set frequency Should I disable them as well? Should I only keep the 2 preferred cores active for the maximum clocks?

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⁠)⁠_⁠/⁠¯

Long story short, I haven't noticed any difference in ocing with ecores on or off, but even if it's there it's just not worth it. Youll get what, 100mhz higher clocks, and youll sacrifice a ton of performance even in games. Cyberpunk, in heavy areas of the game, ecores off drops framerate substantially for example

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?

A lot has changed with the way windows 10 handles Ecores and there is a great review here with 52 games included that shows from my recollection that you generally hurt performance in most cases disabling Ecores.

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.