tried to flash both files with the original video sections (Skylake CPU).
first one (all four microcodes together) went well.
second one (split microcodes) did run ok but was not possible to switch the PC off: it was simply disconnecting from windows.
with Skylake.cannot shut it down. any "soft" control just disconnects windows.
switched it off keeping the on/off button pressed for many seconds.
restarted but had the same problem.
with the first BIOS it is fine
@elisw ,
Can you verify with HWInfo that the Cpu uCode is loaded? Sounds like the microcode not load issue maybe. Well i guess go with the first one then…
strange.
no matter which microcode is in the BIOS HWInfo always shows BA.
tried even removing the battery to clear CMOS, still BA is displayed.
Thats weird. Two in one day. Try another HWinfo version? Care to dump your bios to make sure it flashed properly? Dont remember what version should have loaded?
Will do once I receive the programming tool.
Maybe there is some anti roll back feature, BA is the first mc with the ht fix.
Windows can overwrite a microcode update issued by the BIOS during boot.
You may check it in Windows Registry Editor.
I just became aware of a possible issue one could have modding their firmware if you have an extra microcode for your cpuid with a revision of ‘FF’.
It will look like this:
I asked CodeRush what he thought and he told me that if you have an extra ‘FF’ revision that could be for either a specific microcode meant for overclocking
or to block any other microcodes past that version of ‘FF’ or what ever it really is to be loaded by the OS. Such that if you have a specific board with
unlocked overclocking capability for a certain cpu this may be a problem. Could try to replace the ‘FF’ Revision with another microcode and alter the revision
to ‘FF’ in a hex editor but not sure how that could affect overclocking capabilities. The simplest solution would be just to remove it.
If anyone has this kind of microcode pattern please let us know the effects of either removing it, leaving it, or altering it to FF.
Thanks
Yes that is known. The microcode is actually 74, which is the last version to allow overclocking on non-K SKL CPUs. From 76 that got blocked. Modders & OEMs changed the version to FF so that it won’t be replaced at OS boot with a newer one which would block “unsupported” overclocking. This is a specific case with SKL and there was a similar one with HSW. Changing other microcodes to FF won’t do anything.
Thats interesting. Nothing like a mystery solved. Thank you. Was wondering though for those guys with non K cpus is there a way to unlock overclocking on a newer microcode? Thanks.
The other night stumbled upon this article describing how to analyze the header of a cpu microcode. The blogger wrote a script that used along the 010 Editor app would print out the microcode specifics and details. I tried it out and it worked out nicely. Heres the link to the guide:
https://blog.fpmurphy.com/2016/11/examin…re-updates.html
sorry but it was a red herring. Found the culprit of this and not related to the bios
I reinstalled Windows from scratch, cleared CMOS, flashed the same (old, original) BIOS twice in a row, still got BA.
I’m guessing Windows is pushing the BA version onto your system. Only way to know for sure is to run a Ubuntu live Cd Usb and check.
@elisw
I wonder if changing the revision number to ‘FF’ would solve that problem? Or would it break the intended microcode’s checksum?
As I understand it if you have a microcode with revision ‘FF’ windows is forced to take it…
@davidm71
it is not really a problem for Skylake (unless I intend to overclock the CPU) as long as Windows accepts the Coffee Lake microrcode with that CPU installed.
@elisw ,
So just to clear up any confusion it says ‘BA’ revision is loaded with a 506E3 Skylake CPU installed? Or is this with the Coffee lake cpu you were experimenting with?
So far Coffee lake hasn’t gone over revision 80 and I don’t see any reason why not it wouldn’t accept the correct microcode for a coffeelake cpu if one is installed and the MC file has a correct header.
@davidm71
with Coffee all is good and microcode shown is 72.
A (positive) side effect of your "cleaning" in the first BIOS you provided is UBU can now update the LAN section without errors (there was a size issue).
