No its a flaw in 1903 happens on intel and amd based systems. No one knows what causes it other than everyone has tried everything under the sun to resolve and the issue remains only microsoft can fix it at this point its a inherited issue with the 1903 os itself.
I used to get some nasty spikes, upwards of 3800 to over 6000, or even more. And the spikes happened often, and I didn't have to wait long. Until I applied the latest microcode and mitigation patch that MS just released. Now I get results like this: The ntoskernal.exe spike that you see was the only one after a 1/2 hour of runtime, and it wasn't so brutal as before. Also before current latency was usually 170+, and as you can see I now get under 100 easily.
Yes but when you claimed the microcode update solved the issue that was incorrect. The kernel should still not spike to 900 or 600 or even 300. That is an issue with 1903. ntoskrnl.exe spiking should not occur like it does in 1903 and there is a flaw in 1903 somewhere that only microsoft can fix.
You are not taking into consideration that any kernel or ACPI call is routed through the mitigation software.
I doubt very much it has anything to do with micro-code, the best performance is no micro-code or OS mitigations enabled.
Ok, let's think this out a bit, the microcode is added code on boot for "fixes", the CPU has been running for there lifetime, so could be XP, WIn7, Win10 15xx, 160xx, etc to 1903. The issue only pops up in later versions, so I don't see how it could be a direct cause. Now maybe the new micro-code compared to old might improve, which it is suppose too. I think the cause is more in the core of OS, kernel level IMO.
No worries. I used to be a active member there @NVIDIA , but that guy put me off for good.. been a long time member there since 2007 way before he showed up and he made a big stir there.. he got banned here at guru3d eventually then started his crusade there.. he sometimes has a ok post, but mostly just to provoke and salt on others brains.. xD Anyway sorry for off topic. That said, does this new patch mild this ntskernel bug or not? Personally I didn't get any stutter because of it while gaming, ran that latency monitor when I Played destiny2 and it didn't spike , nvidia driver had ~1200us spike, but when in idle and listening to music I would sometimes get like a stretched sound stutter.. and I saw this ntskrnel spike to 6000-9000 . Max one time was 41000 lol I lowered it when I reinstalled all intel drivers etc.. maybe I should try xfi hd driver reinstall again.. is this newest beta the newest or did they re-release a newer version now (final)?
What patch are you referring too? The last official CU's are from 6/11/19 https://support.microsoft.com/en-us/help/4503293/windows-10-update-kb4503293 https://support.microsoft.com/en-us/help/4497935/windows-10-update-kb4497935 and micro-code https://support.microsoft.com/en-us/help/4497165/kb4497165-intel-microcode-updates If your referring to this one it was only released to preview ring. https://www.tenforums.com/windows-1...ndows-10-v1903-build-18362-207-june-24-a.html My guess will be released soon for public release. I would see if any new drivers get released with 19H1 (1903) specific version listed as might have been optimized better.
Yeah my bad that preview ring. Thought it was released to public by now. Wasn't behind pc for a good week I already have latest microcode.. newer then in that patch though.
https://www.tenforums.com/windows-1...ndows-10-v1903-build-18362-207-june-24-a.html Installed and no changes, still having spikes.