• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

What controls Turbo Core in Xeons?

Page 13 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

lucien_br

Member
Feb 13, 2017
37
7
41
My X99 E4 refuses to recognize this image as a bios - "no image file detected".
The diff shows a lot of changes from the 3.60 bios on the web. Can you describe what you did to this and how you load it?
The USB Stick should be formated in FAT32!
 

lucien_br

Member
Feb 13, 2017
37
7
41
As each cpu unit type has an unique id there shouldnt be a switch in system. However the Turbos are written by Dufus and I dont know how he wrote them for 2696v3. You can check that by using ru.efi after you injected dufus v3.efi, as you can read out your turbo in cpu msrs than.
By myself I wish Dufus gave us a little bit more information about how he created or modify v3.efi, that we could make our own adaptions.
Also there is a way to fake cpuid's. But Im not fond of this either, but would love to try it if someone would know more. Clover FakeCPUID for expample uses a kind of bootloader to fake cpuid (which even results in different clocks) before the general system starts. Mostly its used for increasing support on OS X, but clover also supports Win10.

Still Im struggle to increase my cpu ratio above 28x. Doesnt matter what I try my best performance I have clearly without any microcode and only injected v3.efi. I tried now microcodes 27, 29, 2D, 36, 38 and 39. Its at least all I found. Would like to try any pre-27 microcodes, but don't know where to get them. Anyway they all performed by me exactly the same.

Quite confusing for me that we all got the same MB and as I could see from pictures have even the same stepping of our 2686v3, but got such different results. Thats why I still believe that there is something else we didnt noticed which has influence. Sadly Lucien didnt share yet his setup configs (voltage, c states, turbos, if he used dufus efi or an own, etc.)



Intel download center is your source, but as I said, I didnt found any pre-29 mc, but the others I mentioned are there definately.
My Bios config! There is nothing special! I Swap CPU to Xeon 2669v3 to test but the bios config are the same! The last image the storage config with raid0 port 1 and 5.







 
Last edited:

cekim

Member
Mar 6, 2017
87
19
41
The USB Stick should be formated in FAT32!
Oh, it definitely is. I am able to flash .ROM files from this drive from the BIOS.

In fact, the only way I've been able to flash the BIOS on this board was:
1. via internet to get to 3.10
2. take the A1512.ROM that doing so put on my FAT32 drive and run it through UBU.dat
3. flash A1512.ROM (modified by UBU) via the bios.
I put the X99E43.60 file on the same 4G FAT32 USB stick and you can see it searching through the files on that stick, but it does not recognize it as a .ROM file.
 
Last edited:
  • Like
Reactions: lucien_br

Zladimir

Member
Apr 14, 2011
34
3
71
Hi! Simply I used v3.efi, ucode 0x27 or 0x39 (make no difference), Bios Modded:
https://drive.google.com/file/d/0B_ZQtaxTWt-QdGNGSnA5U0UtMkU/view?usp=sharing
, unpark all cores, speed-step off, vcore offset +0,02, BClk 103,2, C-States on (default) and nothing special configuration.
Try any thing above!
With HT on x29 all 18 cores and HT off x35 all 18 core. See my video with BF1 and general tests for more details and CPU-Z prints!
Thanks for the feedback!
The results for HT off with 35x on allcore I can repeat.
But whatever I try about HT on allcore I never pass x27 with MC 0x27 or 0x39, even with copying your settings fully.
Makes me a bit desperate, cause I know the cpu can do more, but I dont get how.

It works perfectly - 30x on both CPUs
CPU-Z: Single: 1361 and Multi: 40178
Thank you very much.

Aside from benching, my "real-world" app (vray rendering) ends up at 29x across all cores at full load.
Quite nice results.
Is it on two 18c chips ? If yes I wonder why you didnt got close to 50.000 on multi, especially with x29 ratio.


My best results are here: http://valid.x86.fr/2g92bt
18core 1627 single and 24621 multi with x28 ratio on no(0x00) microcode.


@Welsper did you changed something on maximum turbo ? Or did you put like Dufus all on max setting ? Just for curiousity.

Anyway is there any difference in v3.efi to v3x1.efi except support for wider list of processors ? Which die you use Lucien_br ? First or second .efi ?

Still would like to try any pre-27 microcodes, but sadly I cant find any microcode.dat with an earlier version ..
 

Welsper

Member
Mar 5, 2017
25
18
51
Works the v3x2.efi also for me? Asus Z10PA-D8 , BIOS 3107.CAP , Win10 . My First try to disable Microcodes and renamed mcupdater_GenuineIntel.dll.old wont work. In this Case Windows stop here. http://www.abload.de/image.php?img=mobile.16d7sad.jpg
Did you disable using C3/C6 states in BIOS ?
My EP2C612 WS stops in same way unless I do it.
(It is on V3.efi. I forgot to test it on v3x2 when both CPU "unlocked", I'll test it in some hours later.)
 

mtnew10gm8

Junior Member
Mar 6, 2017
8
2
41
Did you disable using C3/C6 states in BIOS ?
My EP2C612 WS stops in same way unless I do it.
(It is on V3.efi. I forgot to test it on v3x2 when both CPU "unlocked", I'll test it in some hours later.)
Yes, C-States disabled.
 

mtnew10gm8

Junior Member
Mar 6, 2017
8
2
41
Thanks for the feedback!
The results for HT off with 35x on allcore I can repeat.
But whatever I try about HT on allcore I never pass x27 with MC 0x27 or 0x39, even with copying your settings fully.
Makes me a bit desperate, cause I know the cpu can do more, but I dont get how.



Quite nice results.
Is it on two 18c chips ? If yes I wonder why you didnt got close to 50.000 on multi, especially with x29 ratio.


My best results are here: http://valid.x86.fr/2g92bt
18core 1627 single and 24621 multi with x28 ratio on no(0x00) microcode.


@Welsper did you changed something on maximum turbo ? Or did you put like Dufus all on max setting ? Just for curiousity.

Anyway is there any difference in v3.efi to v3x1.efi except support for wider list of processors ? Which die you use Lucien_br ? First or second .efi ?

Still would like to try any pre-27 microcodes, but sadly I cant find any microcode.dat with an earlier version ..
No, I have 2 x E5-2683 v3 here, so 2 x 14c. No microcode used so far.
 

mrpiggy

Member
Apr 19, 2012
193
12
81
Oh, it definitely is. I am able to flash .ROM files from this drive from the BIOS.

In fact, the only way I've been able to flash the BIOS on this board was:
1. via internet to get to 3.10
2. take the A1512.ROM that doing so put on my FAT32 drive and run it through UBU.dat
3. flash A1512.ROM (modified by UBU) via the bios.
I put the X99E43.60 file on the same 4G FAT32 USB stick and you can see it searching through the files on that stick, but it does not recognize it as a .ROM file.

I had a similar issue with my Extreme 6 where I had to rename my modded BIOS file to the same name/number format as the ASRock downloaded BIOS's names or it would not recognize or see the ROM file on the USB stick using the BIOS instant flash utility.
 
Last edited:

cekim

Member
Mar 6, 2017
87
19
41
I had a similar issue with my Extreme 6 where I had to rename my modded BIOS file to the same name/number format as the ASRock downloaded BIOS's names or it would not recognize or see the ROM file on the USB stick using the BIOS instant flash utility.
I tried that, but I've just realized the real problem here. This is an X99M Extreme4, not an X99 Extreme4 which evidently has a rather different BIOS and explains why my system is behaving differently than others here.

Sorry for my mistake there, I hadn't looked closely at this board, it was something I had around to sanity check processors, so I never looked very closely at it.

This MB was always a stop-gap experiment. My plan was to migrate this mod to either a dual socket system or a different single depending on performance relative to other setups I have. I'll have access to a Taichi tomorrow. I've tried to hand-modify the A1215.ROM for this board (X99M), but I get the same result (pretty consistent code 79/hang on reboot after successful shutdown that requires CMOS clear to fix).

What is odd about that is that the 0x27, 0x38 or 0x39 uCode have been loaded (in different experiments) by the time I shut down (because I've loaded them in Linux). So, you'd think that whatever the MB needed to shut down cleanly would have been done by that time, but I guess whatever is going wrong has gone wrong before the OS loads and updates the uCode and just does cause anything to fail while its running.

It seems to be a corruption of NVRAM, so that would actually be consistent.
 

custom90gt

Member
Feb 9, 2017
38
7
41
I tried that, but I've just realized the real problem here. This is an X99M Extreme4, not an X99 Extreme4 which evidently has a rather different BIOS and explains why my system is behaving differently than others here.

Sorry for my mistake there, I hadn't looked closely at this board, it was something I had around to sanity check processors, so I never looked very closely at it.

This MB was always a stop-gap experiment. My plan was to migrate this mod to either a dual socket system or a different single depending on performance relative to other setups I have. I'll have access to a Taichi tomorrow. I've tried to hand-modify the A1215.ROM for this board (X99M), but I get the same result (pretty consistent code 79/hang on reboot after successful shutdown that requires CMOS clear to fix).

What is odd about that is that the 0x27, 0x38 or 0x39 uCode have been loaded (in different experiments) by the time I shut down (because I've loaded them in Linux). So, you'd think that whatever the MB needed to shut down cleanly would have been done by that time, but I guess whatever is going wrong has gone wrong before the OS loads and updates the uCode and just does cause anything to fail while its running.

It seems to be a corruption of NVRAM, so that would actually be consistent.
I have a modified X99M Extreme 4 bios if you need it.
 

cekim

Member
Mar 6, 2017
87
19
41
Here is a link:
https://drive.google.com/open?id=0ByAHxCrW_caRdVp6c3hKMFhhMDg
I renamed it as X99MX43.20 so that it would update on mine, it works with my 2683 V3.
That flashes and boots to the BIOS without issue. Thank you.

Seems they have removed a few things:
1. Voltage control of the processor is gone?
2. the "load EFI shell from device" option has now disappeared from the BIOS "Exit" page. I had been using that to run the bcfg commands from within shellx64.efi

It hangs while bringing up linux, I'll have to experiment further to see what that is. In previous attempts with a modded 3.10 BIOS, I had to crank Vcore up a little after loading V3.EFI, but that is not being done here, so I didn't expect the OS to hang.
 

custom90gt

Member
Feb 9, 2017
38
7
41
That flashes and boots to the BIOS without issue. Thank you.

Seems they have removed a few things:
1. Voltage control of the processor is gone?
2. the "load EFI shell from device" option has now disappeared from the BIOS "Exit" page. I had been using that to run the bcfg commands from within shellx64.efi

It hangs while bringing up linux, I'll have to experiment further to see what that is. In previous attempts with a modded 3.10 BIOS, I had to crank Vcore up a little after loading V3.EFI, but that is not being done here, so I didn't expect the OS to hang.
Sadly I've never used anything but the latest bios since I had a v4 before... You can control the voltage of the cpu but it's under the advanced tab I believe. I've only booted into efi shell via a usb device and pressing F11 on startup.
 

vercety

Junior Member
Feb 9, 2017
16
10
51
One question, if I have microcode 27 instaled, how can I go back to have no mircocode (00) in windows? I am trying to get past x27 multiplier but there has been no luck yet (E5 2686 v3 + x99 deluxe)
 

lucien_br

Member
Feb 13, 2017
37
7
41
Oh, it definitely is. I am able to flash .ROM files from this drive from the BIOS.

In fact, the only way I've been able to flash the BIOS on this board was:
1. via internet to get to 3.10
2. take the A1512.ROM that doing so put on my FAT32 drive and run it through UBU.dat
3. flash A1512.ROM (modified by UBU) via the bios.
I put the X99E43.60 file on the same 4G FAT32 USB stick and you can see it searching through the files on that stick, but it does not recognize it as a .ROM file.
This is strange! I never see this case! With me works well on all fat32 usb stick original or modded Bios with the X99 Taichi.
 

lucien_br

Member
Feb 13, 2017
37
7
41
Thanks for the feedback!
The results for HT off with 35x on allcore I can repeat.
But whatever I try about HT on allcore I never pass x27 with MC 0x27 or 0x39, even with copying your settings fully.
Makes me a bit desperate, cause I know the cpu can do more, but I dont get how.
With HT on, my chip does x29 in all cases. See my bios prints and try.
 

lucien_br

Member
Feb 13, 2017
37
7
41
One question, if I have microcode 27 instaled, how can I go back to have no mircocode (00) in windows? I am trying to get past x27 multiplier but there has been no luck yet (E5 2686 v3 + x99 deluxe)
In the folder "cpumcupdate2.1" run as admin "uninstall.bat".
 
  • Like
Reactions: vercety

Zladimir

Member
Apr 14, 2011
34
3
71
Does anyone know if it is possible to force/overwrite somehow the p-states or number of cores which are in turbo used ?

Lets say I want him for some time constantly in 4core turbo, how I can enforce this despite calls from other sources. How is that achievable ?
 

cekim

Member
Mar 6, 2017
87
19
41
This is strange! I never see this case! With me works well on all fat32 usb stick original or modded Bios with the X99 Taichi.
Turned out to be a silly error on my part. I hadn't realized/didn't remembered this board was an X99M Extreme4 not an X99 Extreme4 and that their BIOS is different. Once I load/mod the right BIOS I'm able to get further, but the 3.2 (latest) BIOS) has (for me at least) removed some critical voltage control. So, progress is limited.

Going to try a Taichi for now and see where that gets me. Hopefully, that will get something stable enough that I can do some real world testing to figure out how best to use these chips. Interestingly, with non-ECC I can't increase the clocks, but I can drastically improve the timing.

stressapp shows an 8-9% speed up in memory throughput so far over stock timing (CAS11 1T). Will have to see if that translates to my compute jobs which are very memory intensive and tend to show a 1:1 relationship to increases in memory throughput. Can't use that on a dual board, of course, but 2 singles might do more work than one dual for me. We'll see.
 

ASK THE COMMUNITY