Overclock.net banner
1,481 - 1,500 of 2,993 Posts
  • Rep+
Reactions: Veii and MrTOOSHORT
  • Wow
Reactions: Veii and MrTOOSHORT
With the majority of win 10 results showing superior performance compared to win 11, what is the consensus to just running win 10 only? Is it mainly the end of support in Oct?

Going to tag @yzonker on this question too as I see you posting a lot of win 10 benchmarks as well.
 
  • Love
Reactions: Veii
Have you guys tried 22h2 11?
It's a bit maybe better than 24H2, but still falls behind W10 by a lot.
Tbh I don't know what Microsoft is doing, but leaving a lot of performance on the table.
 
Yah crazy nowadays you need like 8 os versions for benching 🤣
It's ridiculous.
My daily is latest W11 Insider dev, but that's too heavy for benching. But it can't bring these points even after fresh install.
 
  • Wow
Reactions: igor_kavinski
A bit better one:
View attachment 2693189
Have you tested how 25m looks like for you ?
Get under 200ms with your bench system ?
For you i'd guess 185-188ms, maybe sub 180 if Win10 does good on clock ramping

The difference to balanced powerplan is approx. 25ms
Its ramps slow even if PPM should have refined that.

EDIT:
Neevermind.
Well done :) exactly on point.
Screenshot

But was that Win11 or which 10 build ?
 
Have you tested how 25m looks like for you ?
Get under 200ms with your bench system ?
For you i'd guess 185-188ms, maybe sub 180 if Win10 does good on clock ramping

The difference to balanced powerplan is approx. 25ms
Its ramps slow even if PPM should have refined that.

EDIT:
Neevermind.
Well done :) exactly on point.
View attachment 2693206
But was that Win11 or which 10 build ?
This is my W10, but on the 9266 profile. I will check it on my daily windows later.
This happened on balanced PP actually.
 
  • Rep+
Reactions: Veii
Daily is 26100.2448+ ?
Use High Perf if you can.
Nope, it's the Insider Dev. I will check it soon.

Edit:

26120.3000
 
  • Rep+
Reactions: Veii
@Nickshih
It's still happening on 2.26.AS03

Completely dead past:
CAS change, RCD change, XSR too low & at same time BoostMax change ~ single attempt
Started till 55h, crashed, dead.
Aaand it's dead again, now replicable.
ROM + profiles are dead & it's version specific.

The way it dies is failed memory OC, but requirements need to be met:
~ System has to always cold boot cleanly
~ MRC warm training off, fastboot off.
After some time normal boots of reboots, trigger coldboot, it resets, trains, hang on 55h, shut down.
The next boot then is normal
A long duration cold-boot (night) will do the same, start train freeze at 55h shutdown, boot after 2-3 sec successfully

Now if you memory OC
Any change generally is following
~ Boot, 55h, stop, reboot, post
If this memory OC for some reason triggers failsafe reset

It will execute
~ boot, 55h, fail/freeze, execute coldboot, stop because no boot
PSU off and psu on then will fail the boot and ROM (+ME?/PCH) is in coma-state.
Buttons don't function, does nothing, it's dead.

Requires Clean flashback flash
But AS03 is broken.

Unfortunately I don't exactly know what starts this state of double reboots.
Is it the Extreme power settings + PL1/2 edit, is it the per-core overrides or is it the disabled MRC warmtraining bugs.

Whatever triggers it
The way it full fails is a different issue. Once you start to double reboot, if you trigger failed memory boot, you brick it.
It is a guarantee now :)

I'll swap ROM again and see if i get it bricked on other than 2.26~
Or maybe reflash it, with slow ME and see if I get that to brick the same way.
Maaaybe it's something in Peter's ME blob or the way it gets flashed, but unsure yet.
CPU VID table was fine this time.
Just decided to die, but replicable now.
 
Maaaybe it's something in Peter's ME blob or the way it gets flashed, but unsure yet.
Found it. 100% not public ROM related.
It is 1854v2.2 & user issue (me being dumb and cross-flashing)
File having wrong target address.

For some reason worked till platform-halt reset.
Once it fails mem training, init won't function. Kind of logical.

Same if you just hex edit swap, init won't function.
If you patch ROM, init does function as also MEI updater, because address gets corrected.
Need to make my own, can't just reuse and hope its compatible.

Any case target for update
DCOD 63 49
PCOD 63 49
ACOD 132 130.1
Probably some more things I can't decipher blindly.

I'm just stupid, that's the truth.
Sorry for the ping Nick.
Reminded me of same launch problem.

And guys don't flash it, I'll take the old link down.
Till I learn how to make my own comboME too for fun,
well or new Bioses releases
Unfortunate about latency improvements.
 
What are you waiting for, nothing will change in terms of performance with new bios, the thing is finished! When properly optimized, the performance is also perfectly ok.
You sure about that? I am not.
Full silence a month ago, probably we will get a heavily revised bios.
 
1,481 - 1,500 of 2,993 Posts