X99 OC Formula 3.1 - issues with BIOS update 3.20

Hi,


First, let me post the specs:

- X99 OC Formula 3.1
- Intel 5820K @ 4.5/1.23V - Cache 4.0/1.25V - CPU Input 1.9V
- 16GB Avexir Platinum 2400Mhz @ 2666Mhz - Stock Volt

Edited by bshagen - 11 May 2016 at 1:35am

Bump

bshagen wrote: Hi, First, let me post the specs: - X99 OC Formula 3.1 - Intel 5820K @ 4.5/1.23V - Cache 4.0/1.25V - CPU Input 1.9V - 16GB Avexir Platinum 2400Mhz @ 2666Mhz - Stock Volt - 400GB Intel 750 OS-disk - Samsung 512GB 850EVO (2x) for Steam/Origin - Samsung 256GB 830 for Backup - Titan X SLI - EK/BP watercooled, 3x480XTX radiators with EK-pump, temperatures are not an issue  - Peripherals etc.. Issues started a few weeks ago, when sometimes I couldn't reboot without having to reset the PC (hard reboot) and the Dr. Debug would give me all kinds of different codes each time (A0: SATA - Ad: don't know - 66: Chipset - AA: all good/boot) Without knowing where to troubleshoot, I wanted atleast to have the latest BIOS installed, as the version I was on was the 1.20 and the current is 3.20 (BW-E refresh) I backed up the 1.20 to BIOS B first, then I used the UEFI-function to download and install, and it would download/install successfully, however on reboot would revert back to 1.20 again. I had to try 3 times before the flash was actually installed. This was strange to me, never had a problem with this before, and I I have updated my old OC Formula several times this way. Anyway, the flash was good and I started to set all my old settings again - exactly as above, after verifying a good stock CPU boot. This would prove to be difficult. 4.5GHz (Cache and RAM still stock) was tried @ 1.23 - 1.25 - 1.28 - 1.30 - 1.31, not working. Continued to go down 4.4GHz-4.3GHz-4.2GHz @1.31V, still no go. Disabled C-states bu no idea if this would do anything, still no boot. Figured something was wrong with the flash, so re-flashed 3.20, but still same results. Went back to stock settings and booted fine, gave up and went for food.  The crash: The boot goes past POST, and Debug shows AA (good), but crashes on the Windows loading screen, actually it just shuts down the power, like it was a automatic fuse. This in my book usually indicates a bad overclock, but 4.2GHz @ 1.31V and I used to get 4.5GHz @ 1.23V? Another peculiar thing also, everytime I would alter a overclock setting and then save-exit, the Chassis Fan 1 would go to default setting (100%) on my EK-pump when the system restarts. Sidenote tho.. I opened up my chassis and swithed back to 1.20 on BIOS B, applied my old settings and the system boots with no issue., Debug shows AA and all the overclock settings are applied with good temperatures (60 degrees full load) Million$ question: What could be wrong with my system? DISCLAIMER:  I am very open to the fact that this might be my own fault, no bashing on Asrock here, like your products very much. Any help is deeply appreciated :)

IMO, you have two problems that are unrelated, that blended into what seems to be the same, (simplified) problem, something is wrong with your system, as you put it.

The first one is the sporadic reboot failure, which I don't understand.

Do you mean if you have the PC booted, and then do a Windows Restart, the PC would then fail to boot? You would then see a POST code in the Dr Debug display (only codes that remain displayed and don't change have significance), as you described? At that point you have a black screen? Would using the Restart to UEFI feature result in the same thing? No idea what OS you are using, if Win 7, you can't use Restart to UEFI.

Next is the UEFI/BIOS update, which adds a large amount of complication, and more changes/variables.

First is the use of the backup BIOS function. Let's review how it works, from your board's manual:

This motherboard has two BIOS chips, a primary BIOS (BIOS_ A) and a backup BIOS (BIOS_B), which enhances the safety and stability of your system. Normally, the system will work on the primary BIOS. However, if the primary BIOS is corrupted or damaged, just flip the BIOS Selection Switch to ?œB?? then the backup BIOS will take over on the next system boot.

After that, use ?œSecure Backup UEFI??in the UEFI Setup Utility to duplicate a working copy of the BIOS files to the primary BIOS to ensure normal system operation. For safety issues, users are not able to update the backup BIOS manually. Users may refer to the BIOS LEDs (BIOS_ A_LED or BIOS_B_LED) to identify which BIOS is currently activated.

Did you remember to switch back to the main A BIOS before trying to update using Internet Flash?

When using Internet Flash, it is recommended to insert a USB flash drive into one of the USB 2.0 ports on the board's IO panel. Internet Flash will then copy the downloaded UEFI/BIOS file to the USB flash drive in case the UEFI update is interrupted. That will allow the ASRock Crashless BIOS feature to continue the UEFI update. Not doing this is one danger of using Internet Flash.

The failure to update the UEFI problem you had could have several causes, including not setting the BIOS switch back to the main/A position. That problem may also be related to the 3.20 UEFI update, which is NOT a typical UEFI update.

To review, you did get the 3.20 UEFI version installed, correct?

Now my explanation about what the problem is when you use the 3.20 UEFI/BIOS.

NOTE: The following is my theory about the use of the Broadwell-E processor compatibility UEFI update for X99 mother boards, regardless of the manufacture. This theory has yet to be verified, but my past experience with this type of UEFI/BIOS update tells me what was required in the past will also apply in this instance.

X79 boards are able to use Sandy Bridge-E and Ivy Bridge-E HEDT processors. Intel 6-series chipset boards (P67, Z68, etc) can use mainstream Sandy Bridge and Ivy Bridge processors. Both of these platforms required a UEFI/BIOS update to provide compatibility with Ivy Bridge processors. We now see the same thing for X99 boards, a UEFI/BIOS update to allow compatibility with the new Broadwell-E processors.

The Ivy Bridge compatibility updates for X79 and Intel 6-series chipset boards did not simply have a UEFI/BIOS update to provide that compatibility. They both had corresponding updates to the Intel Management Engine (IME) software.

The UEFI/BIOS download pages for those updates also contain links to the required updated version of the IME software. For example:

http://www.asrock.com/mb/Intel/Fatal1ty%20X79%20Professional/?cat=Download&os=BIOS

http://www.asrock.com/mb/Intel/Fatal1ty%20Z68%20Professional%20Gen3/?cat=Download&os=BIOS

Given this, why would X99 boards not also need a new version of the IME software for use with the Broadwell-E compatibility update? That applies even if we are not using a Broadwell-E processor in our X99 boards. That is true for X79 and Intel 6-series chipset boards.

The UEFI/BIOS compatibility updates contain a new IME firmware version, that is part of the UEFI/BIOS file, like an Option ROM is part of a UEFI/BIOS file. The new IME firmware requires the new IME software, since they work together.

The IME software and firmware allows communication between the internal CPU registers that contain data about over clocking, the UEFI itself, and the OS. That is a very simplified statement, but is the basic reason over clocking fails if the IME software and firmware are not matched.

That is exactly what you are experiencing now, and what other experienced in the past when the IME software was not updated along with the UEFI/BIOS processor compatibility update.

So where is the new IME software for X99? It comes from Intel. Intel has not released it yet, since Broadwell-E processors are not yet available. I've looked, as of the date of this post, no mother board manufactures have updated versions of the IME software for X99. Intel does not as well.

Why release UEFI/BIOS updates for Broadwell-E if the IME software is not available? Because one mother board manufacture released the UEFI updates, and if others don't, owners of those boards get worried, start complaining, get angry, and criticize those manufactures.

Again, all of this is my theory. Everything I've seen so far with X99 and the Broadwell-E compatibility UEFI updates is identical to what happened with X79 and Intel 6-series chipset boards. Not using the correct version of the IME software after the processor compatibility UEFI/BIOS update? Then any OC will fail, along with other strange side effects.

If I'm wrong, then every X99 UEFI update for Broadwell-E has many problems when used with Haswell-E processors. Either way, they aren't working right... yet.

Your short term fix? Don't use UEFI 3.20 yet. I have an X99 Extreme6/3.1 board. Have I updated to the Broadwell-E compatibility version? No chance I would do that... yet.

I had a similar experience to what's described in this post so i'd like to confirm and agree that the 3.x BIOS versions have some issues. Here are my findings.