Sorry I do not have a better answer or preference ,all I can say good job and I am sure someone who likes getting that extra couple points in Cinebench or Latency in AIDA64 will help out more.
Don't worry, really appreciated!
I wouldn't focus too much on AIDA64 or Cinebench or any else benchmark on Windows 11... for me keeps being very unreliable.
It's already tons better than the first release but not yet mature.
Gaming is probably the best thing to run with Windows 11.
I have sometimes a big variance but in general seems to be very enjoyable.
I guess when DirectStorage will get more traction it'll become more interesting for me.
Damm I just notice I forgot CPU Package power in the RTSS on screen display.
That's something nice to see but better to measure with CapframeX.
You can analyze the total power usage during the run and get even an fps/Watt metric to compare, really nice!
But what would be the best is to measure and log data with a smart plug.
If someone cares about power consumption, buy a Tasmota plug and a Raspberry Pi.
The whole is extremely cheap and will tell you for real what is the power consumption.
The power plan changes the whole system draw, not only the CPU.
For example with my 5950X when the package goes up 80W in reality means 100W to 150W of AC draw.
Intel Core Ultimate LowPower Win11 v1 got the same results as Bitsum Highest Performance.pow ,so that is a good thing.Not sure why it was not boosting higher,it did work.So I am 100% fine with that.
In theory is doing what it should, being LowPower
Probably it wasn't necessary to clock that high. So in theory is a good thing!
I didn't see a boost clock behavior change with the 5950X but well, it does not clock ever remotely that high...
What matters is if the game fps is close enough or same as the other Ultimate plans.
If you or anyone else is going to run a decent in-game benchmark, like SotTR, would be nice to have just a screenshot of the result with different power plans.
Nothing big or time consuming.
These 3 were all good and I have no problem using them and have no preference.I am just a PC Gamer so I do not try to find the lowest latency in AIDA64 or the best Ram timings to test PC Games at lowest settings 1080p.
High Performance Win11
Ultimate Performance Win11
Intel Core Ultimate Performance Win11 v1-By Mannix
I wouldn't run any other plan other than mines with Win11 or Win10
Not because of pride of course, the reason is very simple; the default settings for the P-Cores are... plain rubbish.
Which I suspect is the main reason they so fondly suggest to use Win11 instead of Win10.
Win11 is just discarding or adapting power plans settings at its will. While Win10 doesn't.
They must have realized too late that the P-Cores to boost needs parking.
And it's quite the opposite with every other CPU.
In the power plan there is a set of settings for cores with Efficiency Class 1.
Those settings are set by default for less powerful cores, with parking enabled, and should be applied to the E-Cores in this case.
The E-Cores in the CPUSet from the system have Class 1 and the P-Cores are 0, like all other CPU cores.
I guess that Intel and Microsoft decided would have been too risky to have those new CPU depending on a specific power plan to work properly.
Bad reviews, bad user feedback because of running with the wrong power plan. A no-go for sure.
AMD has it for the 3000s but they still work pretty fine with a standard plan.
So they invented this smart trick of flipping the P-Cores ad E-Cores efficiency class.
When both P-Cores and E-Cores are enabled in BIOS the Windows Scheduler is internally swapping the Efficiency class.
I couldn't test it myself but I think when the E-Cores are disabled, the P-Cores are using the "normal", Efficiency Class 0 settings.
This is very likely the reason why Gaming in general is better with the E-Cores disabled.
Unfortunately I can't fine-tune these settings, the only Alder Lake I have access to is a 12600K not OCed.
I can't test properly the boost behavior, made some changes but it could probably be better.
The mistake form Microsoft & Intel is that they didn't change the Efficiency Class 1 settings to match a Performant processor, they kept it as it is.
Not sure why, maybe Microsoft is planning to support ARM CPUs or something else.
So the result is the P-Cores have boosting policies set for efficiency instead of performances...
The initial P-state after un-parking is 50%, the policy is IdealAggressive instead of Rocket, it doesn't clock immediately to the higher P-State.
The latency sensitivity hit is 50% instead of 97-99% so they are much slower to react to user input, etc, etc.
That's the reason why some games, like the Division 2, are stuttering hard with middle-end Alder/Rocket-lake CPUs when the E-Cores are enabled.
With my power plan I've been told it doesn't happen; I will check with the 12600K if I can replicate.
Obviously when you have highly clocked CPUs, the issue is less evident.
Win11 is doing its own magic to compensate but on Win10 there's no magic sparkle on top, the plan limitations are going to be more evident.
That's why they warmly suggest to only use Win11.
And again thanks a lot for the video with all this editing effort!
