Page 2 of 30
#16 RE: CPU Microcode BIOS modding questions/problems by plutomaniac 22.10.2014 00:14

avatar

I did some searching at cpu world and basically

06A1-06A5 are Nahalem CPUID's --> useless for Z77
06A6 is only for mobile i7-2720QM SB ES/QS cpu --> useless for Z77
06A8 is only for some mobile IVB ES/QS cpus --> useless for Z77

So I will leave only 06A7_28 and 06A9_19



Alternatively, this is done automatically by UBU (leaves only 06A7 and 06A9).

#17 RE: RE:CPU Microcode BIOS modding questions/problems by MonarchX 06.11.2014 23:44

Zitat von MonarchX im Beitrag #3
Zitat von killkernel im Beitrag #2
CPU Micro-code Rev. 1B for Ivy Bridge and 29 for Sandy Bridge may oblige you to increase the VCore of some step to maintain the level of stability achieved in an overclocked CPU with the older revisions that in this case are the Rev. 19 for Ivy Bridge and 28 for Sandy Bridge and for this reason in UBU is indicated "For Overclockers".
As i work in a laboratory that repair PCs and 'ive the possibility to test a lot of mainboards i've checked this situation and is confirmed even on ASUS motherboards. ;)
So the older Rev. 19 may require less voltage to achieve the same overclock for some CPU steppings, correct? How much less voltage? 0.01v? or 0.1v? 0.1v would be a BIG deal!

Maybe the reason it does that is because it increases voltage without telling you and Rev. 1B shows true voltage? ASUS motherboards tend to do that on their own. When I set voltage to 1.30v, real voltage is 1.31v. Maybe its the same with Rev. 19. I am just curious if the Rev. 19 could possibly improve my overclock that is already high @ 4.8Ghz @ 1.31v. Is there any info on which steppings the Rev. 19 benefits the most??? What about temperatures? Does Rev. 19 run hotter than Rev. 1B?

Here's my CPU info in extreme detail - is it one of those that can benefit from Rev. 19?



Thank you fro such prompt feedback and support!!!

EDIT by Fernando: Unneeded quoted text deleted (to save space and performance)


Here's the post I lost! I did not realize it was split from the main topic! So, back to the same question. Would the 6A9-19 CPU Microcode affect my Ivy Bridge CPU overclock, based on the GPU-Z info I provided? I never got that long PM...

#18 RE: RE:CPU Microcode BIOS modding questions/problems by killkernel 07.11.2014 09:44

Zitat von MonarchX im Beitrag #9
Thank you! Can't wait to read it!


@MonarchX

Hello, as you have read Fernando have redirected to a dedicated thread for the topic of your interest about CPU Micro-code.
As stated in the previous post, at the moment I'm very busy with jobs that have upcoming deadlines but as soon as I have the right spare time necessary to deepen the questions you posted, i'll respond in a thorough and exhaustive as possible and compared to my knowledge and experience.
Thanks in advance for your patience. :)

Chee
Zitat von MonarchX im Beitrag #17
Zitat von MonarchX im Beitrag #3
Zitat von killkernel im Beitrag #2
CPU Micro-code Rev. 1B for Ivy Bridge and 29 for Sandy Bridge may oblige you to increase the VCore of some step to maintain the level of stability achieved in an overclocked CPU with the older revisions that in this case are the Rev. 19 for Ivy Bridge and 28 for Sandy Bridge and for this reason in UBU is indicated "For Overclockers".
As i work in a laboratory that repair PCs and 'ive the possibility to test a lot of mainboards i've checked this situation and is confirmed even on ASUS motherboards. ;)
So the older Rev. 19 may require less voltage to achieve the same overclock for some CPU steppings, correct? How much less voltage? 0.01v? or 0.1v? 0.1v would be a BIG deal!

Maybe the reason it does that is because it increases voltage without telling you and Rev. 1B shows true voltage? ASUS motherboards tend to do that on their own. When I set voltage to 1.30v, real voltage is 1.31v. Maybe its the same with Rev. 19. I am just curious if the Rev. 19 could possibly improve my overclock that is already high @ 4.8Ghz @ 1.31v. Is there any info on which steppings the Rev. 19 benefits the most??? What about temperatures? Does Rev. 19 run hotter than Rev. 1B?

Here's my CPU info in extreme detail - is it one of those that can benefit from Rev. 19?



Thank you fro such prompt feedback and support!!!

EDIT by Fernando: Unneeded quoted text deleted (to save space and performance)


Here's the post I lost! I did not realize it was split from the main topic! So, back to the same question. Would the 6A9-19 CPU Microcode affect my Ivy Bridge CPU overclock, based on the GPU-Z info I provided? I never got that long PM...

Hello MonarchX! Let's start with a basic information about your configuration e.g. CPU-Micro Code because from your screenshot is not available and to know it you can download, execute and take a screenshot of HW Info (System Summary): http://www.hwinfo.com/download.php

Cheers,

KK

#19 RE: RE:CPU Microcode BIOS modding questions/problems by CodeRush 07.11.2014 20:20

avatar

Just a small addition to CPU microcode modding on new BIOSes for Haswell/Broadwell/Skylake platforms: all microcode binaries are now referenced in a special Firmware Interface Table (read more here, it's the only open document I can found), which can be found by searching for non-Unicode text "_FIT_" in the BIOS file.
If you remove, move, replace a microcode with a file of different size and somehow other tinker with microcodes on a BIOS with FIT, you need to calculate and adjust addresses there. And, if FIT has entries of any other types then Microcode (0x01), you are pretty much done here, because such BIOS is not moddable anymore. I can't say more because of NDA, but please read the document linked above, while it's still available.

#20 RE: RE:CPU Microcode BIOS modding questions/problems by MonarchX 07.11.2014 22:53

OK, here is the CPU screenshot from HWiNFO (latest beta):

HERE is the saved report from HWiNFO (latest beta).

OK, so far I used the latest CPU Microcode for BIOS, NOT the one for overclocking. All I would like to know whether the CPU Microcode for overclocking can provide any overclocking benefit for my specific CPU.

#21 RE: RE:CPU Microcode BIOS modding questions/problems by killkernel 08.11.2014 16:30

OK MonarchX, as you stated you are using CPU Micro-code Rev. 1B for Ivy Bridge and from my experience on different motherboards of different producers, the Rev. 19 is better because at the same level of OC with the same HW, cooling system and BIOS code it is needed less VCore to maintain stability achieved.
An example: One of my i7 3770K is stable @4,5Ghz on a Gigabyte Z77X-UP5 TH with 1,15V with F12J BIOS and CPU Micro-Code Rev. 19 but when i've modded the F12J with the Rev. 1B i lost stability and to regain it i've upped the VCore to 1,155V.
I've tested Gigabyte, ASUS, MSI and ASRock motherboards on Z77 chipset and all the tested MoBos, except ASRock, with BIOSes with Rev. 19 of CPU Micro-code performed better than the same BIOSes but with Rev. 1B of CPU micro-code.
ASRock is the only producer that with modded BIOSes with Rev. 1B of CPU micro-code worked better than Rev. 19...
The only way for your rig is to try the same BIOS version modded with Rev. 19 of CPU Micro-code and check the stability with less VCore e.g. 1,305V.
The little difference of 0.005V may not seems a big vantage but when you're @ the limit of the OC by the temperature or VCore that you feel safe for you system, every little drop (temperature or voltage) may help.

Try it and let us know if you'll benefit or not.

Have a nice week-end!

Cheers,

KK

#22 RE: RE:CPU Microcode BIOS modding questions/problems by MonarchX 08.11.2014 20:52

Thanks for the explanation! I will definitely try and let you know. I shall de-lid my CPU and get Swiftech H240-X AIO WC (very low temps) soon to try to reach 5Ghz. This CPU Microcode may be the key!

#23 RE: RE:CPU Microcode BIOS modding questions/problems by MonarchX 09.11.2014 20:10

Zitat von killkernel im Beitrag #21
OK MonarchX, as you stated you are using CPU Micro-code Rev. 1B for Ivy Bridge and from my experience on different motherboards of different producers, the Rev. 19 is better because at the same level of OC with the same HW, cooling system and BIOS code it is needed less VCore to maintain stability achieved.
An example: One of my i7 3770K is stable @4,5Ghz on a Gigabyte Z77X-UP5 TH with 1,15V with F12J BIOS and CPU Micro-Code Rev. 19 but when i've modded the F12J with the Rev. 1B i lost stability and to regain it i've upped the VCore to 1,155V.
I've tested Gigabyte, ASUS, MSI and ASRock motherboards on Z77 chipset and all the tested MoBos, except ASRock, with BIOSes with Rev. 19 of CPU Micro-code performed better than the same BIOSes but with Rev. 1B of CPU micro-code.
ASRock is the only producer that with modded BIOSes with Rev. 1B of CPU micro-code worked better than Rev. 19...
The only way for your rig is to try the same BIOS version modded with Rev. 19 of CPU Micro-code and check the stability with less VCore e.g. 1,305V.
The little difference of 0.005V may not seems a big vantage but when you're @ the limit of the OC by the temperature or VCore that you feel safe for you system, every little drop (temperature or voltage) may help.

Try it and let us know if you'll benefit or not.

Have a nice week-end!

Cheers,

KK



I have an idea about this CPU Microcode and how it may possibly be good for overclockers, but I would need to use Multi-Meter to measure voltages and also know where to measure them. Sometimes voltage settings in BIOS are not real voltages. For example, every time I select 1.31v in BIOS, CPU-Z reports 1.32v. In reality, voltage can be as high as 1.33v. We don't know unless we measure. Maybe with CPU Microcode 19 a Voltage offset of 1.31v in BIOS = 1.320v in CPU-Z and 1.330v in real life (multi-meter). With Microcode 1B a Voltage offset of 1.31v in BIOS = 1.320v in CPU-Z and 1.320v in real life (multi-meter). Is that a possibility? I know ASUS does this a lot, but you said that other brands also get the same effect from Microcode 19. Maybe Microcode 1B is just more accurate/precise and reports correct voltage, while Microcode 19 is less accurate and reports -0.01v than what actual voltage is. It just makes less sense for Intel to release CPU Microcode that performs worse UNLESS the one before was providing incorrect readings.

When you have tested Microcodes - have you used Multi-Meter to see if that voltage was truly the same? I am curious to know.

#24 RE: RE:CPU Microcode BIOS modding questions/problems by killkernel 11.11.2014 11:37

Yes, i have measured voltages with DMM on motherboards that offered point of measurement.

#25 RE: RE:CPU Microcode BIOS modding questions/problems by HAON 22.11.2014 15:20

How do I inject a old Microcode in an GA-H97-D3H F4 Bios for my Xeon E3 1230 V3 CPU?
Thank you!

Link to the Bios:
http://www.gigabyte.de/products/product-...x?pid=4962#bios

#26 RE: RE:CPU Microcode BIOS modding questions/problems by Fernando 22.11.2014 15:57

avatar

HAON:
Welcome at WinRAID Forum!

Zitat von HAON im Beitrag #25
How do I inject a old Microcode in an GA-H97-D3H F4 Bios for my Xeon E3 1230 V3 CPU?
Why do you want to an old CPU Microcode for your CPU?

Regards
Fernando

#27 RE: RE:CPU Microcode BIOS modding questions/problems by HAON 22.11.2014 16:23

For Multicore Enhancement. So I can run all cores on 3,70 GHz. This function was deactivated by newer microcodes because of Intel. They didnt want you to do that. They want to sell their K-Cpus und Z-Boards.

#28 RE: RE:CPU Microcode BIOS modding questions/problems by Fernando 22.11.2014 19:54

avatar

@ HAON:

Since I am not an expert regarding the CPU Microcode up- and "downgrade", I hope, that someone else will help you.

#29 RE: RE:CPU Microcode BIOS modding questions/problems by HAON 30.11.2014 10:16

nobody?

#30 RE: RE:CPU Microcode BIOS modding questions/problems by mbk1969 30.11.2014 16:04

Zitat von HAON im Beitrag #25
How do I inject a old Microcode in an GA-H97-D3H F4 Bios for my Xeon E3 1230 V3 CPU?
Thank you!


You can load BIOS image to MMTool.exe and replace all CPU MCUs with needed MCU.

Edit: On "Replace" tab select file with GUID "17088572-377F-44EF-8F4E-B09FFF46A070" then load module with needed CPU microcode to "Module file" field and press "Replace" button. Then go to "CPU Patch" tab and check version of microcode.

Edit2: Or you can place module file with needed microcode to the "Modules\CPUI" subfolder of UBU tool with the name of one of its original module files - "LGA1150.ffs", "LGA1155.ffs" - to replace it with your one (choose correct one with appropriate socket and family). And then use UBU tool to do the job.

Xobor Forum Software von Xobor
Datenschutz