I’m confused .. unless this was just hardware failure in real time.
No boot device is recognized?
BIOS is not accessible?
Maybe reflash BIOS again if such a stick might be recognized?
Without access to the BIOS I can’t reflash it. And now when I press CTRL-F5 /to enter the flash mode), I get “BOOTMGR is missing, press any key to restart”. I will get this machine to my vendor, maybe they have an idea. But I will never, ever flash my BIOS unless I have a really good reason for that.
If its as you say and the fault was just the bios update itself somehow killing the mobo then I would have to hand it to you .. or MSI .. it would be the first example I have personally witnessed online or IRL in the last decades. And that has included some dodgy ones too. 
For whatever it is worth the data on the disks should be intact so there is that.
So I just took it to the vendor. They said they tried a Windows ISO and that works just fine. I said “but it says there is no boot device” and they said “well, your Linux died somehow, you have to reinstall that and then you will be able to boot again”
Now what? I remember that when I set up the machine for the first time, I used some magic to add the ESP to the UEFI boot manager. Maybe I just have to redo that?
But I thought neither the BIOS nor any boot device was accessible?
If they are then .. proceed.
If necessary use a live system to then use efibootmgr, ex;
I’ll open a new thread on this.
Suspend standby sleep, never worked proberly on any os, i got 3 systems in my home, 7945hx3d from minisforum with an 9070 xt card 32gb ram ddr5, intel 14900k with an 5080 card and 64gb ram ddr5, Ultra 9 285k with an 4090 card and 96gb ram ddr5.
Also have 5 laptops 1 macbook pro 1 windows and 3 linux laptops 1 from tuxedo computer, specs on the laptops
Windows and Mac have no issues like you have with wierd graphical artifacts on desktop, but work in apps gets corrupted easily, my best advise is save your work and turn off your machine, 30-40 year old issues that might never get resolved.
So in the end the BIOS was not fried at all.
USBs were bootable.
( I still dont understand how this could be confused. But oh well. )
But the boot entries were reset.
So adding the entries again was all that was needed.
( I note that in the other thread you used refind-install which obviously worked fine .. but it might be mentioned that your ESP was not gone and nor were the boot files. So you really only needed to re-add the entry either through the BIOS itself or something like efibootmgr. refind-install happened to do this as well as part of its various functions. )
Hurrah.
But the issue of post-suspend glitches is back too?
PS.
In case this was not highlighted enough..
As mentioned before/above/etc you needed to re-add the entries only.
You know the first item to boot was a generic HD1 when it should be your linux bootloader.
Some of the original images seem to also show secure boot enabled when it needs to be off.
That is all.
So unless there was something else missing then your vendor was wrong about your linux being gone or ‘dead’.
Let me try to recap the things.
So in the end the BIOS was not fried at all.
No, it surely was not. What seemed to happen was that the NVRAM entries were wiped. Until this day, I did BIOS updates on Windows machines only, and it seems as if these BIOS Updates look for Windows Boot Managers only and add them automatically to the NVRAM. I didn’t know this. So the machine had no NVRAM Boot entries, therefore it didn’t “see” the ESP to boot from. Or something like that and a refind-install fixed all that.
USBs were bootable.
( I still dont understand how this could be confused. But oh well. )
Now this is where I am really confused, too. The machine did not boot from a USB stick. Holding F11 during boot brought up the Boot Manager, and when my Ventoy USB stick was attached, the BM showed two entries for the two partitions. Trying to boot the first (==the ESP) just brought me back to the BM, trying to boot the second resulted in this behavior, That’s when I started to panic.
Then I read somewhere that you might have to switch from UEFI to CSM in order to boot from a stick. I did that and then things got even worse, because after that, the BIOS came up in a resolution of something like 800x600 and it was all jumbled, so I could not even change back to UEFI instead of CSM. That was when I took the machine to the vendor.
I still don’t fully understand what happened, but in their shop, they were able to see the BIOS okay. So I took the machine back home again just to see… nothing. Black screens all over. So I decided to detach all drives except for the NVME I actually want to boot from and I also removed the dedicated GPU. This was the moment when things got better. From the integrated GPU, the BIOS showed up okay, so I switched back to UEFI. Then I was also able to actually boot from my Ventoy stick (I think I had to plug it into one of the ports directly on the mainboard, not the front one).
So I was able to run a Cachy Live ISO, chroot into my system (getting some weird Warnings about missing or wrong UUIDs, but I guess they stemmed from the missing hard drives), run refind-install to add the ESP to the NVRAM and all was fine again.
Hurrah.
Indeed 
I’m sorry to have raised such a fuss, I was really panicking…
But the issue of post-suspend glitches is back too?
Frankly, I don’t care at the moment 
I’m just happy that I can continue my current Noita god-run and also Lego Skywalker Saga
As mentioned before/above/etc you needed to re-add the entries only.
Yes, I understand that now. I wasn’t aware of the fact that the BIOS upgrade wiped the NVRAM and seems only to care about Windows Boot Managers, ignoring everything else. I learned a valid lesson yesterday.
So unless there was something else missing then your vendor was wrong about your linux being gone or ‘dead’.
Totally true. They are completely ignorant of Linux and they also told me in the beginning that they only really care for Windows machines. So I don’t blame them.
Some of the original images seem to also show secure boot enabled when it needs to be off.
This is another mystery, because I never set Secure Boot to true, at least not to my knowledge. But of course I can’t really retrace every step that happened.
All in all, I am again very happy for all the nice people here in the forum that spend their precious lifetime helping out noobs like me. I try to give back as good as I can, being well aware of my limits…
A little bit hard to follow, but im having the main issue as well. was the fix to change something with the /tmp directory?
I didn’t try to put that machine to sleep since. Let me do it now…
The problem is the same and worse: this time, “Hibernate” did not even turn the computer off at all and when it came back (after desperately trying CTRL+ALT+F-Keys and the power button, it not only woke up with the graphical glitch, but Firefox was completely frozen, too.
If I need to continue from where I left, I use the “Save Session” functionality of KDE and shut the computer down.
Im still having this issue. I’m starting to think its a wayland thing, and ill test this out later with a live ISO. I’m also experiencing an issue where chromium apps dont seem to detect my GPU, but im 90% sure its my fault
I gave up on this. In KDE, you can save your desktop session, which is what I do instead.
Hi. I’ve had this issue for a while now, hoping that updates will fix it. Did anyone get to the bottom of it? Was it a BIOS issue? Funny how I have almost the same hardware config as the OP.
Not me, no. After learning about the “Save Desktop Session” feature of KDE, I decided that I don’t need the sleep functionality at all.
The B650-Wifi mobo is pretty wide-spread, I guess. You have the same CPU and GPU, too?
I believe the problem stems from Nvidia, and I just don’t have the time to take care of it. A BIOS Update was a disaster, so I went back to the original firmware (although I later learned I should have just done a refind-install)
RTX 3060 and Ryzen 5. MSI motherboard. Would like to get to the bottom of this issue with suspension. Strange this hasn’t been reported anywhere else other than on Cachyos.
I followed the steps from the Arch wiki to fix this. (RTX 3070)
TLDR:
- Verify the kernel module parameter NVreg_PreserveVideoMemoryAllocations=1
- Verify that the services are enabled by default
- Use ext4 as filesystem on /var/tmp and make it 150% (at the very least 105%) the size of my GPU memory.total
Complete entry in the wiki - take extra care reading the Note box: NVIDIA/Tips and tricks - ArchWiki
Thanks @igniztion I try it out and let you know.
While trying to implement @igniztion solution I accidentally killed initramfs (my system uses dracut instead of mkinitcpio) and had to blow away all my stale kernels (inc. fallback), before regenerate a fresh initramfs. This seems to have fixed the wakeup glitches for now. But I have a feeling it will come back, because there is something in my workflow that’s causing it and won’t be resolved until I follow the prescribed solution above. Thanks guys.