Z820 e5-2600 v2 ivy bridge upgrade - thread reactivation

[SDH]

I'll post a duplicate of this in the Z620 similar thread up nearby now also.  This relates to an article from England by an author who was personally present at the HP new releases meeting early in September 2013 shortly before the Ivy Bridge Z420/Z620/Z820 version 2 upgrade was officially announced (with the press embargo lifted 9/10/13).  This is the same date that HP did its own press releases announcing this improvement.  Of interest, the author talked personally with  HP's Ron Rogers, the top dog for HP workstation R & D.  The "chips" referred to are the Ivy Bridge processor upgrades and the RAM mentioned refers to those processors allowing the higher speed 1866 MHz memory sticks to work at full speed.  HP and Intel are in close communication well before specific advances are known to the tech writers and the public.  Read this, and especially what I have bolded/underlined:

"While the chips and RAM are new, the rest of the workstation is the same as before. At the event last week,
HP's Ron Rogers – who runs the company's R&D for workstations – told me that they were aware of the V2
range when designing the Z420, Z620 and Z820.   From the outset, the motherboards used supported both V1
and V2 Xeon E5-2600 chips and haven't needed to be upgraded for the V2s.   It's therefore possible for owners
of older Z420, Z620 and Z820 models to swap out their V2 chips and 1,600MHz RAM for V2 chips and
1,866MHz RAM."     I believe the author meant to state "...to swap out their V1 chips and..." in that last part.

Remember that this is a quote from the article's author, and not from Mr. Rogers directly.  The engineers may have found some issue that led them to back off from that approach, or maybe this is a misunderstanding by the author.

HERE  is the quote's source, and it is still active 1/20:

https://www.digitalartsonline.co.uk/news/creative-hardware/hp-adds-ivy-bridge-xeon-e5-v2-chips-with-...

[DGroves]

I've seen and read that article,.....

i agree HP and Intel work hand in hand and HP had advance samples of the newer cpu's well before the z820 launch date so  the time between the v1 and v2 cpu's was under 10 months so saying HP had not fully qualified the newer xeon's when the z820 was released is a bit odd but the only reason i can think of

i do know of at least four z820 board revisions from 1.00 to 1.04 and i have seen 1.01-1.03 boards as v1 or v2 listed on ebay (with bios screen posts) so the board revision is not a factor, was HP reworking the spare boards to the current bios release and perhaps updating the bootblock for warranty replacements?

did hp update the boards and also perhaps chipsets?, but these updates made no difference to the boards ability to run the newer lines of xeons... if so then the bootblock is/was a arbitrary decision hp made until intel officially released the newer line of cpu's to the public

the z800 is a different beast, on this model HP absolutely did have a I/O chipset revision to allow for  the 56xx line of cpu's and these boards that are 56xx capable have a board revision label that ends in xxxxxxx-003

[HashCollision]

Hi all,

Short update on my experiments with a v2 CPU in a v1 motherboard, upgraded to a 2013 boot block. In brief, it works!

More specifically:

With a 2013 boot block and v7 ME firmware, using the v2 CPU works if AMT is disabled by BIOS setting or motherboard jumper or both. Otherwise, with AMT enabled in BIOS + Xeon E5 v2 + v7 ME, the PC will try to boot, ramp the fans, turn off, and then try again until you pull the power and put the ME/AMT jumper into AMT disabled position or switch back to a v1 CPU.

I then upgraded to ME v8 firmware by taking the ME firmware section of the v3.85 BIOS update and placing it into a dump of my Z620's BIOS, then flashing the modified BIOS back to the workstation. The system now reports as having a v8 ME version. I didn't get the v2 CPU and AMT enabled working at the same time, but I think that was more to do with some other ME update steps needing to be performed (a reset of the ME, probably).

So, ME firmware of a Z620 (and presumably Z420, Z820) CAN be upgraded v7->v8, and as I mentioned in my guide there are several ways to do this. However, you don't even need to bother, as you can simply upgrade your boot block and keep AMT disabled. I used my v2 CPU to go into BIOS, boot into DOS, boot into Linux with a GPU and do some tasks, and even boot into my server environment (but I dared not start any VMs as my v2 test CPU is an E5-2603 v2, 4core@1.8GHz).

I have also done a little research and found out that HP, Lenovo, etc originally intended to upgrade the Sandy Bridge workstations to work with Ivy Bridge - boot block, microcode, ME firmware, everything; but Intel decided they didn't like it. So, broadly speaking, HP Zx20 v1 were meant to work with Ivy Bridge CPUs, and you can do the minimum to get v2 CPUs working (upgrade boot block, disable AMT) or go with something like what the official update procedure would have been (upgrade boot block, upgrade ME version, perform the appropriate ME maintenance procedures).

[SDH]

I think HP is going to want to hire you!  Great work..... here is a link for others to your guide:

https://github.com/SuperThunder/HP_Z420_Z620_Z820_BootBlock_Upgrade

Click in that page on the PDF link to get to it......

A quick question..... a person can set ME to disabled in the BIOS but if they then reset BIOS to defaults ME would  be reactivated, and that sounds like it opens a can of worms or could cause significant confusion.  Do you believe using the jumper approach  to turn off ME is better, and if so what header/pin bridging would you suggest?

The flashing to the 2013 later v2 boot block date would likely be beyond the skills of most of us here.  I even found the concept of swapping in a v2 used motherboard was too much for many, but I've done that for all our old v1 Z620 workstations and have the time down to about 30 minutes if I'm in a competition with one of our IT guys.  The trick is to use a fair amount of good tape to position all the cables up and out in the exact correct order as you uninstall the v1 motherboard...... 

[SalSimp]

Indeed he has done a great job but I think he will have more to tell about the ME soon enough, just stay tuned.

[HashCollision]

Good question about what happens if the BIOS is reset to defaults. If the default is to have ME/AMT enabled, then that would result in a non-booting Zx20 if it had ME v7 but a v2 CPU installed. The header is guaranteed to work, and AMT is a very rarely used feature (and one with several vulnerabilities), so I see no reason not to have the ME always overriden. I've been running my Z620 with the ME override header activated for several months now.

The header to use is "E1 ME/AMT Flash Override". It's a 3-pin header where the setting is chosen by which pin the center pin is shorted with. In my guide I show the location on Z420/Z620 and Z820. On Z420/Z620 the default is to be in the position away from the row of SATA ports, and the override position is to place the header closer to the SATA ports. On early Z820 it's in the bottom right corner and horizontal; the default positon appears to be to be righwards. On later Z820, it is vertical and located on the underneath the lower right bank of RAM slots. The default is the upper position as with Z420/Z620.

This winraid thread has some good discussion on Z820 ME upgrades:

https://www.win-raid.com/t1826f39-HP-Z-Unable-to-detect-ME-Firmware-amp-Management-Mode-in-Bios.html

It may be possible to make a system that would do the boot block and ME upgrades from Windows, if the user activates the E14 and E1 headers. But it would be a very risky program to run without the clamp or desoldering + programmer backup, as interrupting the flashing process leaves you in all probability with a corrupted BIOS.

30 minutes for a board replacement is not a bad time at all - it can take that long just to get a good clamp on the BIOS flash chip.