What controls Turbo Core in Xeons?

[mtgben]

kjboughton any idea

 

[fabiosun]

yep sttingns to 37 for core you can have this..but it is not working in real life
have you renamed windows32 dll?

 

[polyx]

Hi folks,
oh guys you are all so great, many BIOSes worked just fine till now! Thank you so much!!!

Oh, I´am searching for another modded BIOS. I found one for the "mod_Z10PE-D8-WS-ASUS-3407.zip", but is there a one for the Z10PE-D16 WS, maybe?! I use 2x e5-2686 V3 QS, it would be so great to unlock all core of these two also.

Best whishes
Chris

 

[kjboughton]

This says you have EIST disabled.

 

[mtgben]

This says you have EIST disabled.

but eist and turbo enable....

 

[mtgben]

yep sttingns to 37 for core you can have this..but it is not working in real life
have you renamed windows32 dll?

system 32 ...? which files???

 

[mtgben]

system 32 ...? which files???

work all core turbo x37 for you real...?upload hwinfo64 image...

 

[polyx]

...or can I use the Z10PE-D8 WS ModBios also for the D16, I guess not, right?!

 

[kjboughton]

work all core turbo x37 for you real...?upload hwinfo64 image...

I wonder if the problem is *no* microcode applied. Perhaps the system will not properly turbo without something applied. Have you tried applying a microcode at boot following UEFI programming?

Start here:
https://labs.vmware.com/flings/vmware-cpu-microcode-update-driver

Also...
New Broadwell-E/EP RC9 release build to set OC lock bit following programming. Please try this:
https://github.com/freecableguy/v3x4/releases/tag/v3x4-0.10b-i406f1-rc9

 

[Cata40]

can it be from microcode in system32, mcupdate modified or deleted?

 

[kjboughton]

...or can I use the Z10PE-D8 WS ModBios also for the D16, I guess not, right?!

ASUS Z10PE-D16 WS modded for v3 (Haswell-E/EP) unlock

 

[foureight84]

but eist and turbo enable....

on a random note, was EIST green before the flash? If you don't know or it wasn't, might want to check the cpu power pins as well. Happened with my dual v4. They wouldn't turbo and i couldn't figure out why. I thought it was because I had a bad set of QS CPU's. Months later I was cleaning my computer and saw one of the two CPU power connectors was not connected.

 

[mtgben]

can it be from microcode in system32, mcupdate modified or deleted?

i modified this...but nothing...

 

[mtgben]

I wonder if the problem is *no* microcode applied. Perhaps the system will not properly turbo without something applied. Have you tried applying a microcode at boot following UEFI programming?

Start here:
https://labs.vmware.com/flings/vmware-cpu-microcode-update-driver

Also...
New Broadwell-E/EP RC9 release build to set OC lock bit following programming. Please try this:
https://github.com/freecableguy/v3x4/releases/tag/v3x4-0.10b-i406f1-rc9

i have cpumcupdate2.1 but i havent broadwell file for example: 0x27 - 0x38 -0x39 for hoswell...please help for this and i checking..
i test this v3x4-0.10b-i306f2-rc9_no_ucode_lock_oc.efi

all is wery well to programming and in hwinfo64 multiplier all core is x37 but when i stres test cpu-z or rendering cpu show x28 also normal without turbo all
just core 1 is x37 ...!
?

 

[evilr00t]

i have cpumcupdate2.1 but i havent broadwell file for example: 0x27 - 0x38 -0x39 for hoswell...please help for this and i checking..
i test this v3x4-0.10b-i306f2-rc9_no_ucode_lock_oc.efi

all is wery well to programming and in hwinfo64 multiplier all core is x37 but when i stres test cpu-z or rendering cpu show x28 also normal without turbo all
just core 1 is x37 ...!
?

if you're using an ES1 chip that uses microcodes with a mask of 0x80000000, you almost certainly won't find it in the cpumcupdate dll unless you add it yourself. That should be able to be extracted from your BIOS if it has a 0x80000000 mc update, which it almost certainly does.

 

[mtgben]

if you're using an ES1 chip that uses microcodes with a mask of 0x80000000, you almost certainly won't find it in the cpumcupdate dll unless you add it yourself. That should be able to be extracted from your BIOS if it has a 0x80000000 mc update, which it almost certainly does.

i dont no..help me for that ...?my chip is aptio v

 

[mtgben]

kjboughton dont be tired...
kjboughton in now time all is ok just one problem is cpumcupdate2.1 for broadwell and all turbo multiplier core show x37 in aida but in cpu-z ,hwinfo64,other win software(3dsoftware) all core turbo to x28)!!!

in aida all core x37 and stres cpu-z to x28..i also set bios to turbo charger...but i dont know >>(C 0,c1 ,c2 , c6 state cpu ??)

i also modified mcupdate file to cancel update microcode (mcupdate_GenuineIntel ,mcupdate_AuthenticAMD ,mcbuilder)

 

[wiseguyuk]

Somebody post a couple of V4s to @kjboughton. He will find a way. That in my opinion would be the fastest method.

 

[wiseguyuk]

https://www.tweaktown.com/news/59905/intel-cpus-serious-bug-vulnerable-remote-attacks/index.html

E5-2696 V3 may be vulnerable

 

[fabiosun]

kjboughton dont be tired...
kjboughton in now time all is ok just one problem is cpumcupdate2.1 for broadwell and all turbo multiplier core show x37 in aida but in cpu-z ,hwinfo64,other win software(3dsoftware) all core turbo to x28)!!!

in aida all core x37 and stres cpu-z to x28..i also set bios to turbo charger...but i dont know >>(C 0,c1 ,c2 , c6 state cpu ??)

i also modified mcupdate file to cancel update microcode (mcupdate_GenuineIntel ,mcupdate_AuthenticAMD ,mcbuilder)

Hi mtgben
to have all core in Aida @x37 is only a bios result. in my case I set my bios to syncallcores and put value to 37. so for now we have no evidence we can use old V3 finding
Yesterday I have tested all microcode available with UBU for my cpu/motherboard
With old microcode my system boots fine but I have no result using @kjboughton script
With latest system does not boot..
Default one for my motherboard is like xxx0C
I think some result we can achieve finding some value in Cpu power limit
Also, when using AVX in test I achieve different result in different OS

MS windows 10 pro 64bit/ cinebench15 MAX TDP reaches 137
all cpu @2.8 Ghz with a result of 3170 MAX TDP reaches 124

OSX High Sierra
all cpu @2.6 Ghz with a result of 2950

@kjboughton
Thank you for your effort

 

[fabiosun]

E5-2696 V3 may be vulnerable

2696 V4 is ok instead...

 

[Cata40]

Tool Started Fri 24.11.17 8:58:10 AM
Name: WINDOWS-SC6I3FN
Manufacturer: ASUS
Model: All Series
Processor Name: Intel(R) Xeon(R) CPU E5-2696 v3 @ 2.30GHz
OS Version: Microsoft Windows 10 Pro
Engine: Intel(R) Management Engine
Version: 9.1.37.1002
SVN: 1
Status: This system is not vulnerable.
Tool Stopped

 

[Dufus]

E5-2696 V3 may be vulnerable

Relax, ME is to do with the chipset and nothing to do with the CPU. Your results are probably due to ME being FUBAR or lacking MEI driver.

Systems from ME 11.0 and upwards are quoted as vulnerable which means chipsets with Skylake and later.

 

[foureight84]

Somebody post a couple of V4s to @kjboughton. He will find a way. That in my opinion would be the fastest method.

What should I post? More than willing to do this.

 

[mtgben]

yes thank you kjboughton
with kjboughton EFFORT (efi cpu programmer) set all turbo ok...(x37 for 2696 v4)
only one problem is turbo maximum in windows lock to x28..an dont up to x37....!!!

after EFFORT programing and exit to windows all core is x28 and when i use cpumcupdate2.1 all back to x12..
i think we have a problem in windows to block turbo for x37.and i think with all friend working (specialy kjboughton)we colud unlock full turbo speed multiplier and remove microcode from bios for v4...

only one problem and it is block xeon v4 2696 cpu turbo speed to x28 (in WINDOWS) !!!!