News Desperate PC DIYer appeals for help after spending $20,000 on a build that doesn't work despite returning multiple parts – misfortune began with a...

I recently was in a very similar situation.
Eventually I found the culprit to be the memory stick used to update the bios, updating with a corrupted bios image. Switching to another memory stick and reflashing the bios fixed the problem.
 
Q-Code 92 being "PCI Bus initialization is started" - correct?

"The Redditor noticed they needed to update the Asus Pro WS WRX90E-SAGE SE motherboard to fix an incompatibility issue with one of the NVMe SSD drives in the system."

What was the original/specific "incompatibilty issue"?

What NVMe SSD? What other drives are installed? Make, models, configuration?

What about references to the Motherboard's User Guide/Manual:

Motherboard User Guide/Manual

[Verify that I found the applicable User Guide/Manual.]

Just curious.....
 
I’ll happily build my own machine for gaming, but the machine I earn my money with always comes from a supplier with a support contract just to avoid this sort of thing.

Having your pc broken down for days is way WAY more expensive than paying some more for a pre built and then paying some more for a support contract.
 
It sounds like the particular builder is in over their head and panicking. PCs aren’t magic - they’re machines. You simply keep isolating and testing components until you figure out what’s going on. At this point there’s either something on the bus that’s bad or the BIOS flash went wrong. The next thing I’d try is to remove all the drives from the system and see if you can get it to POST. I would not suspect that the main board or anything that’s already been RMA’d is the culprit at this point. Heck, I’d try to POST with a different power supply even - you just never know.
 
I don't understand why he didn't just use the built in BIOS firmware upgrader.

I still always try to stay away from BIOS updates anyways but those built in upgraders are wonderful. No longer having the need for making a boot USB and then typing in a bunch of commands. Yikes. Just go into BIOS and select the file on disk. It's like using Office.

But 3 boards in a row with the same issue is super weird. Must have been one of the RAM sticks cause you gotta have at least one.

And that NVME drive that caused all of the initial problems would've been in the trash by now for me.
 
  • Like
Reactions: MosephV
This is why updating a BIOS isn't always "safe" — things can and do go wrong at times, even for the most experienced among us. I had a Gigabyte (or maybe it was MSI?) board back in the Ryzen 7 2700X days that killed the 2700X when I updated the BIOS. The flash appeared to go fine as well, and swapping in a different AM2 chip showed that everything else was working. I ended up having to replace the CPU.

The biggest failure points, in any build, are usually (in order of potential):
  1. Motherboard
  2. RAM
  3. PSU
  4. CPU
  5. GPU
  6. SSD
That's assuming the problem isn't just "user error," which is actually the most common problem in my experience. But the fact that this builder tried replacing the CPU and GPU before trying different RAM is interesting. If I had to guess, it would be that the BIOS update ended up with an unstable RAM voltage and possibly fried one or more of the sticks — assuming it's the RAM, which at this point seems like the only real possibility remaining.
 
Had this issue before with an ASUS Crosshair III Formula back in the day, except it just stopped working out of the blue. Sent it back 3 times and they said it was fine. Finally yeeted it and bought a new board, everything worked fine. It was then I stopped trusting ASUS.

But to the point of the article: he used "the BIOS tool" and turned his system off during the update, who knows what "the BIOS tool" did.

And I agree with the above comments that I'd you're going to spend a fortune on a proper workstation you spend a little more and get one from a workstation builder. You might skip things like extra drives and RAM, which you add yourself later at a lower price, but those often are sold with on site repair warranty included so downtime is minimal, and some, like from System76, come with elegant custom cases.
 
I’ll happily build my own machine for gaming, but the machine I earn my money with always comes from a supplier with a support contract just to avoid this sort of thing.

Having your pc broken down for days is way WAY more expensive than paying some more for a pre built and then paying some more for a support contract.
Exactly, not to mention double the risk of a melted power connector with two 4090's in there.

At this point, I would try a completely different motherboard model.
 
  • Like
Reactions: bit_user
This is why updating a BIOS isn't always "safe" — things can and do go wrong at times, even for the most experienced among us. I had a Gigabyte (or maybe it was MSI?) board back in the Ryzen 7 2700X days that killed the 2700X when I updated the BIOS. The flash appeared to go fine as well, and swapping in a different AM2 chip showed that everything else was working. I ended up having to replace the CPU.

The biggest failure points, in any build, are usually (in order of potential):
  1. Motherboard
  2. RAM
  3. PSU
  4. CPU
  5. GPU
  6. SSD
That's assuming the problem isn't just "user error," which is actually the most common problem in my experience. But the fact that this builder tried replacing the CPU and GPU before trying different RAM is interesting. If I had to guess, it would be that the BIOS update ended up with an unstable RAM voltage and possibly fried one or more of the sticks — assuming it's the RAM, which at this point seems like the only real possibility remaining.
Yep, RAM would definitely be up there for me in this situation. I've seen bad RAM cause all sorts of weird issues, some not even seemingly related to RAM.
 
Yes I have always built my systems and keep older parts for troubleshooting a new build. When errors pop up like that you go back to basics. My latest board did not like the order of the new memory sticks even though they were the exact same model and speed. 4 - 16GB modules. I followed the recommended slots and order but two sticks did not like their spot. I flipped the sets of sticks and voila it worked but the board was down to one drive, two sticks, the CPU and an older Nvidia card that I knew worked perfectly on my old build a 2080Ti. Then I start adding parts and booting between full shut downs and unpluged powered downs.
 
  • Like
Reactions: bit_user
I'd put my money on the mobo. When dealing with 100s of PCIe lanes, there are 1000s of possible combinations, many of which are probably untested for by the MF, and one of which our "high end builder" may have "accidentally" stumbled upon...
Second guess would be PSU, which may not ne able to deliver stable power to all said "high end" components. Individual power rails must be checked that they indeed can sustain the load without V drops... That would be a best case, worst one being a cable meltdown or right out fire (that's why PC builders always have an extinguisher ready)!
MF=manufacturer... Just in case!
 
Second guess would be PSU, which may not ne able to deliver stable power to all said "high end" components. Individual power rails must be checked that they indeed can sustain the load without V drops... That would be a best case, worst one being a cable meltdown or right out fire (that's why PC builders always have an extinguisher ready)!
Article says a PSU-replacement was tried.

AFAIK, most modern PSUs are now single-rail designs.

One of the very first things I'd have tried was removing RAM. Not necessarily because any of it is defective, but as others have noted, there could be something about which slots it's installed in. For basic setup and installation, you only need one stick. You can try swapping DIMMs and slots, in order to be sure that it's not the RAM. Easy thing to eliminate and a common source of errors.

BTW, on such a build, I hope they used ECC memory. Also, I'd tend to stick with the motherboard's QVL, if I were spending so much on memory. You don't want to spend a couple $G on memory, only to discover the motherboard doesn't like what you bought and now you're stuck trying to sell it and buy some other kind.
 
  • Like
Reactions: JarredWaltonGPU
Pick 2, 4, or 8 sticks of RAM (whatever is the minimum) and verify that they are working in another system. Then put them into bank 2 (or whatever is the first bank) of this system and see if you can get it to boot. Also make sure all your fan headers are all still properly connected especially the CPU fan0 and case fan0 or the bios might not boot (it happened to me.)

I find it helpful to make a copy of the motherboard diagram and outline exactly what connections must be made with colors in ms-paint so i can visually check that my entire build matches the intent.
 
  • Like
Reactions: bit_user
I have similar system like this guy (but twice cheaper, otherwise with 20k price tag it is not worth DIY it ) even the case is the same brand, and plan to move from Genoa to Turin chips some time in the near future. I have Gigabyte MZ73 PCIe5.0 dual processor motherboard, works almost 2 years 24/7 with Linux and Windows simultaneously for number crunching . The scary part is that it will need to upgrade BIOS to accept new gen AMD chips, the appropriate BIOS already exists. Anybody have any experience with BIOS upgrades of server Gigabyte motherboards?
 
Last edited:
I had the exact same issue with a crosshair VII Hero motherboard and a Strix 3090. the issue was the BIOS in the GPU had its CMOS settings screwed up/made incompatible after a motherboard UEFI/BIOS update.

I put the 3090 into a secondary x4 slot, put an OLD GPU into the primary. Everything booted up fine. I was then able to power down the system, put the 3090 into the primary slot and it's booted fine ever since.

basically there's CMOS-style settings on the 3090/4090 series that get messed up, but putting ONE card as a secondary allowed windows to boot and reset those settings to a bootable state once the drivers detected the card.
 
I'd definitely test that with minimum required hardware (just CPU and one RDIMM) then swap in a known working basic video card (preferably one that didn't need PCIe power, but definitely not one that used 12VHPWR). Too bad that board doesn't have an easy way to run graphics bypassing the CPU lanes or that could be an option too.

Pretty sure I'd have bought a different motherboard from a different manufacturer after that first RMA though.

The words BIOS tool make it sound like using OS based software which should never be used unless there's no choice (laptop/handheld). Their workstation boards have a pretty advanced UI in the BIOS for updates these days as well. I highly doubt this is still an issue though since they acknowledged that.
BTW, on such a build, I hope they used ECC memory.
RDIMMs are platform mandatory so there isn't really a choice.
 
  • Like
Reactions: bit_user
I don't understand why he didn't just use the built in BIOS firmware upgrader.

I still always try to stay away from BIOS updates anyways but those built in upgraders are wonderful. No longer having the need for making a boot USB and then typing in a bunch of commands. Yikes. Just go into BIOS and select the file on disk. It's like using Office.

But 3 boards in a row with the same issue is super weird. Must have been one of the RAM sticks cause you gotta have at least one.

And that NVME drive that caused all of the initial problems would've been in the trash by now for me.
Being that this is a brand new computer I doubt it uses BIOS It most likely uses UEFI
 
  • Like
Reactions: frkhrd