HP Z8-G4 New BIOS Problems w/ PixInsight

[DGroves]

sorry but no,...... you have not confirmed it was/is a HP bios issue  as in reality all you have done is done a clean OS and software install on a same system without installing the latest bios on this system you have no way to actually confirm it's the issue

and i do understand you not wanting to update as doing a rollback to a very early bios appears to be not practical

what you can do however is swap the drives between the two HP systems, and see if the new OS install has the same issue on the original HP system with the later bios

and also see if the drives with the issue when on the later bios is now problem free on the new system you bought with the earlier bios

[silverlitz]

I did a clean OS install and the problem was still there.  I have swapped the drives on both systems, and in every case the Z8 with v2.95 has the problem.  I have tried with different CPUs with v2.95, and there are still problems.   When reverted back to v2.86, with the exact same OS, it works.  What I have yet to try is whether the very recent v2.96 works.

The one thing that could support your thesis, is that this was the 1st time I ever did a BIOS update using Windows Update.

[silverlitz]

Yesterday, I ran PixInsight's WBPP on a large group of subs on the "old" Z8 G4 with latest v2.96 BIOS,  and it ran as it should.  Either the new v2.96 BIOS fixed the problems w/ v2.95 or somehow Windows corrupted the v2.95 install (???).

Can Windows interact with the BIOS flashing causing BIOS corruption?

[DGroves]

yes, flashing within a OS can have the potential for bad bios updating this is why many windows bios updaters actually extract the necessary files needed for the bios update within windows  to the boot drive then execute the update script from dos apon windows restarting once the update finishes the script restores the normal windows startup this method is what HP most likely uses for your workstation as earlier workstation models such as the z800/z820/z840 use this method to update

i did previously check  PixInsight's user forum and found no user reporting issues with a z8 G4 due to a bios update/revision

you might as a user of PixInsight's software ask their support dept of any known issues with your model computer

i still lean towards a software related issue rather than a bios related one if you read HP's description of the changes between 2,95 and 2.96 (and earlier) it will describe all changes

https://www.cloudynights.com/topic/922643-very-disappointed-by-pixinsight-support/

[silverlitz]

After working fine with WBPP for a few days, including a very large dataset that took ~5 hours (w/ drizzle), the "old" Z8 G4 with the "new" Gold 6240 and v2.96 BIOS exhibited the same behavior twice today with a smaller dataset.

I have switched my "fully loaded" setup, w/ 384GB and the many SSD/HDDs to the "new" Z8 G4 with my "old" Platinum 8273CL and the "old" v2.86 BIOS.

I am glad this behavior w/ v2.96 surfaced before changing the BIOS on the "new" Z8 G4.

The motherboard on the "new" Z8 G4 is Rev 1, while the "old" is Rev 0 (some other suffix added). 

[silverlitz]

The PI WBPP dataset that bombed twice today with my "old" Z8 G4 with Gold 6240 and BIOS v2.96 just processed without a hitch with my "new" Z8 G4 with Platinum 8273CL and BIOS v2.86.

[DGroves]

again, stop looking for a non existent hardware/bios problems your issue is most likely a software related issue

from what i've seen on the products support forums there are quite a few user posts on software stability/bugs of this product

[silverlitz]

I still do not think it is a PixInsight S/W issue, as the thing that changed before the issue was Windows 11 Update and Z8 BIOS change.   Prior to this change, it performed for a ~1.5 years without a single issue, and after BIOS change it failed 100% of the time (PI WBPP).  Also, PI WBPP has performed at 100% with every other computer tried: Z840 and Z8 G4 with v2.86 BIOS; all using the exact same dataset as well as different datasets.

Until yesterday, the evidence was hinting at corrupt BIOS update while using Windows Update.  Though, after the initial problem, I used UEFI to drop to v2.94 and back to v2.95, and both times I had the same problem every time running PI WBPP.

With the update of BIOS to v2.96 (on the Z8 w/ v2.95), things seemed to be going fine.  I was running WBPP on a large dataset and adding drizzle (which takes a long time) on both Z8's to get a handle of performance difference of the Platinum 8273CL vs. Gold 6240, as well as 384GB vs. 192GB RAM.  The large dataset worked as expected on both Z8's (v2.86 and v2.96 BIOS).  Then yesterday, running WBPP on a smaller dataset on the v2.96 machine, the problem reappeared.  However, this time was different, as the v2.96 Z8 worked fine for a few days on the larger dataset.

This morning I watched a YouTube video from a computer builder source (focuses on video gaming and GPU performance) and it referred to a relatively new problem of a specific Windows Update that causes serious problems with SSDs with certain controllers (mainly Phison).  

The current boot drive on "new" Z8 with v2.86 BIOS is a Samsung MZVL21T0HCLR-00BH1, which has a Samsung Elpis (S4LV003) controller.   This was the boot drive that was originally on my "old" Z8 when it started having problems.  The current boot drive on "old" Z8 with v2.96 BIOS is a Crucial P310 which has a Phison E27T controller.   This Crucial P310 drive was cloned from the Samsung boot drive after the clean Windows 11 Pro install and subsequent s/w reinstallation.

Yesterday after the problems, I switched the RAM, GPU, and all non-boot drives between the two Z8's.  This left the "old"  with v2.96 BIOS and Crucial P310 boot drive with only 192GB of RAM as my secondary Z8.  I am currently rerunning the PI WBPP on large dataset on this v2.96 BIOS Z8.  I am currently ~3 hours 20mins in this process, and it is working fine.  This is much different behavior than with v2.95 BIOS, which failed fairly early in the process (though at different spots) every time.  This test will also give me data on the differences between the Gold 6240 with 192GB vs. 384GB (384GB is clearly faster on most of the processing steps so far).

So, it is looking like the current problem has a different source than my original, even though the final symptoms are the same.  My working theory is the 1st case was due to bad BIOS install due to Windows Update and the current problem might be related to a specific Windows Update causing serious problems with Crucial P310 SSD with its Phison controller.  

[silverlitz]

I ran PI WBPP on both the very large dataset and the smaller one (that bombed twice a few days ago) on my "old" Z8 w/ v2.96 BIOS, which had its RAM reduced to 192GB.  The "full boat" Z8 is now the "new" Z8 w/ v2.86 BIOS, after the latest problems with my "old" Z8.

This time both datasets worked fine.   Again, the behavior w/ v2.96 is very different from v.2.95, which would bomb every time.

[DGroves]

you simply won't let go your preconceived idea that it's a bios issue and are doing everything you can to make results fit this idea instead of sitting down and doing a correct testing mythology

the proper way is to clone the system's drives with your working setup and then run both systems with identical bios settings on the same data sets 10 or more times or until the new system fails, then take that failed drive  onto the working system and run it 10 or more times while doing the same with the drives from the old working system placed in the new one

your current method is not the correct way to check for problems as it causes non reproducible results if only run once  in slightly different ways on different systems

and again you software programs user forums are filled with people having random crashes on different hardware configurations there's nobody saying they are having hardware or bios revision or bios settings issues just software ones

i'm done with this thread it's your system do with it what you will