Overclock.net banner
1 - 9 of 9 Posts

·
Registered
Joined
·
12 Posts
Discussion Starter · #1 · (Edited)
I can guess on some 'basic' reasons:

Power Outage / System Crash - Enter incomplete and or corrupt values in memory addresses such that the card can't figure out what to do anymore.

Using a Different Card / Maker BIOS - Different order of memory addresses and or different sizes of memory addresses causing nonsense data and or overflowing and corrupting other memory addresses.

Setting some base value so high the card shuts down for protection and or constantly crashes.


1) But beyond that. I don't know. Curious to hear your thoughts / info.


Other Bits:
2) Have you ever 'bricked' a card by starting with that card's own default BIOS and modifying values from that (with non 'totally crazy' values)?
3) What seems to cause the biggest / most frequent issues with your past attempts?
4) What is preventing some cards from ever being re-flashed again?
 

·
Registered
i7-12900K
Joined
·
788 Posts
Technically speaking, you're a hundred percent right. But I'd argue that most real world error comes from human factors.
  • Accidentally hitting the reset button with your chair during a flash.
  • Accidentally placing the wrong .bin file on the flash drive.
  • Thinking "huh, it must be done by now, let me restart" half way through the update. (Did this one last week myself.)
We all have the same dumb monkey brain and it loves to **** us over. :unsure:
 

·
PC Evangelist
Joined
·
48,031 Posts
Most of the time when using the wrong vbios and the tool is not stable or u dont have the official internal tools amd/nvidia uses.
 

·
Registered
Joined
·
356 Posts
Some causes left out:

When your cat decides to lay on your keyboard in the middle of the update.

When your dog decides to give a tug on the power cord.

When your 3 yo asks "what's this?" and pokes the red light on the power strip.

Moral: keep everybody OUT of the room when updating BIOS'.
 

·
Registered
Joined
·
170 Posts
The worst I've done is to cause the driver not to load on a Maxwell card. The card was easily re-flashed with corrected values.
 

·
Tank destroyer and a god
Joined
·
2,790 Posts
I once used bad *.bin. What I did later was "blind" flash in DOS (no image) and i was lucky to restore the card.

Most recently I downloaded UEFI enabled bios by AsRock for my Radeon VII, I executed it expecting some sort of "Do you want to flash new bios". No such message, the bios flashed correctly and enabled me to disable CFM and use full UEFI system.

Just to test i tried to restore original Sapphire bios. I knew it failed for the first time, because GPU-Z wasnt returning any BIOS version, so i did not rebooted the system else the card would be bricked. I tried to flash it again with a different *.bin, this one was correct as GPU-Z and driver returned BIOS version correctly and i rebooted and it was OK.
 
  • Rep+
Reactions: boddole

·
Registered
Joined
·
1,103 Posts
I flashed an R9 280 and a GTX 760 more times than I can remember... A few times the GPU just wouldn't respond to anything or crash etc. Stick it in a different slot, put a known working GPU in the top slot and re-flash. Never had any issues "un"bricking a non working card after a bad BIOS flash.
 

·
Registered
Joined
·
4,276 Posts
Flashing a Vega64 bios to a Vega56 can soft brick the Vega56. The 64 has Samsung HBM. Some 56 have Samsung HBM, some have Hynix HBM. If you flash 64 bios to a 56 with Samsung HBM, you end up unlocking extra cores and memory bandwidth. If you flash 64 bios to a 56 with Hynix HBM, the card won’t boot because the bios is telling the memory to run at a speed it can’t handle. Re-flashing to a 56 bios fixes this.

AMD eventually started releasing drivers that cause stuttering with Vega cards that have non-original bioses. I hope NVIDIA doesn’t lock down the 3090.
 
  • Rep+
Reactions: boddole
1 - 9 of 9 Posts
Top