BIOS Updates

Status
Not open for further replies.

Odigo

Gawd
Joined
Apr 22, 2002
Messages
805
Do you bother to update the BIOS if the only documented change is a CPU Microcode update?
 
I try to update the bios a few months after it is released to give it time to bake. Only thing I am not too keen on is laptop bios updates.
 
Only if I have a problem to fix.

This has been my philosophy.

I flash the latest BIOS when I am upgrading a system (new CPU, etc.) and am going to do stability tests anyway.

I almost never update a BIOS in between hardware upgrades. It's just a pain in the ass. I tweak the hell out of detailed BIOS settings and it is a pain when they reset them all and I have to go back in and tweak them again. That and consumer motherboard manufacturers are very sketchy when it comes to BIOS updates. Sometimes they break things, other times they remove features that used to be there (like ECC capability) etc. etc.

So, my philosophy is, unless I am tearing down a system and setting it up again, I generally never flash the BIOS. This way you avoid surprises. And it was also what they used to recommend back in the 80's and 90's. Never touch the Firmware unless you have a problem you think it might resolve. Though these days they tend to be a little bit more liberal with the BIOS updates. Some laptops even do it as part of Windows updates :eek: I find this very very sketchy, but at least laptops have a built in battery backup to ensure the flash doesn't get interrupted.

The only exception to this was when I needed to mitigate Meltdown/Spectre on my server.


Edit:

Whoops, just realized I necroed this thread. My bad.
 
Do you bother to update the BIOS if the only documented change is a CPU Microcode update?
I always update to the latest official BIOS version available at the time I get a motherboard on the bench or for a system build. After that, I only update BIOS versions as needed. IE, the BIOS offers some feature or improvement I want to take advantage of or there is a fix for something that I'm dealing with. Beyond that I do not unnecessarily update BIOS versions. I am especially reluctant to do so on boards that are so shitty they don't work with default/automatic settings out of the box.
 
I always update to the latest official BIOS version available at the time I get a motherboard on the bench or for a system build. After that, I only update BIOS versions as needed. IE, the BIOS offers some feature or improvement I want to take advantage of or there is a fix for something that I'm dealing with. Beyond that I do not unnecessarily update BIOS versions. I am especially reluctant to do so on boards that are so shitty they don't work with default/automatic settings out of the box.

Yeah, it really sucks when something was working, then you update BIOS and now it no longer does.

I had this happen on a Supermicro server board once.

They had two of the PCIe slots reversed in the BIOS menu where you set PCIe bifurcation. I had figured this out in troubleshooting when I originally set the system up years prior, but had of course forgotten about this by the time I flashed the BIOS.

The new BIOS (which mitigated Spectre/Meltdown) "fixed" the reversed PCIe slots, and suddenly one 8x device was running only in 4x mode, and one of my 16x m2 risers only had one of its 4 m.2 drives visible to the system.

It caused some frustrating downtime.

And that's SuperMicro. They are MUCH better with this stuff than any of the consumer/gamer brands are.
 
Yeah, it really sucks when something was working, then you update BIOS and now it no longer does.

I had this happen on a Supermicro server board once.

They had two of the PCIe slots reversed in the BIOS menu where you set PCIe bifurcation. I had figured this out in troubleshooting when I originally set the system up years prior, but had of course forgotten about this by the time I flashed the BIOS.

The new BIOS (which mitigated Spectre/Meltdown) "fixed" the reversed PCIe slots, and suddenly one 8x device was running only in 4x mode, and one of my 16x m2 risers only had one of its 4 m.2 drives visible to the system.

It caused some frustrating downtime.

And that's SuperMicro. They are MUCH better with this stuff than any of the consumer/gamer brands are.
I've also seen BIOS updates break things like PCIe spread spectrum which is required for things like storage controllers to work. I've also seen it totally break OROM loading with multiple GPU's and storage devices in the system. If it ain't broke, don't fix it certainly applies. People are often wrongly under the impression that BIOS updates will magically add performance or make their systems better overclockers down the line. In the two and a half decades I've been an enthusiast, I have rarely ever seen any significant performance gains from BIOS updates. The few cases I have were with things like the 680i SLI reference boards which didn't work right in the first place and those updates came at a cost. Fried memory modules, broken PCIe device compatibility and so on.
 
People are often wrongly under the impression that BIOS updates will magically add performance or make their systems better overclockers down the line. In the two and a half decades I've been an enthusiast, I have rarely ever seen any significant performance gains from BIOS updates.

Same. I've mostly seen stability and feature fixes, but that's about it.
 
About the only time you really should be consistent with BIOS updating (outside of actual problems) is when a platform is first launched. Whenever Intel's Z790 successor arrives for example, you'll want to do that. With Z690 BIOS updates were really important in the early days because DDR5 compatibility was a shit show. With the upcoming Z790 refresh boards, you won't need to bother with that.
 
Can as well flip a coin. With the Changelogs that we are given with the updates we are blind. There are always lots of changes not in the log.

If it is really just a microcode update I don't need it since Linux loads those fixes at boot.
 
I guess I'm living life in the fast lane like James Dean. I update to the latest bios risking it all lol.
 
There is rarely any benefit to updating your BIOS and there is always that risk of bricking the board.
Agreed.

It comes down to risk/benefit.

There is a small (but not negligible) risk that a flash goes wrong, and you brick it. And even though that can usually be recovered, it is usually a huge pain in the ass, up to and including sending the board to the manufacturer to be re-flashed.

It's a risk that is worth taking if you are having a problem you need to fix, but if everything is working, I'd just leave well enough alone.
 
I do every time I notice a new bios. I never once bricked a board from a flash in 30 years.
 
Last edited:
Back when I was on Intel systems I only rarely did. Mostly when everything was new and RAM compatibility updates were pretty important. With AMD systems I generally jump on most major updates. With the way they roll out new AGESA versions (that often have major changes), it seems to matter more. Luckily I've had zero troubles with BIOS updates in forever. I bricked an ABIT board back in the late 90's, but have had no problems since.
 
I do everyone I notice a new bios. I never once bricked a board from a flash in 30 years.

I bricked 2 Supermicro X58 Xeon boards in a row in what I still think was an incorrect BIOS they put up on their website. But their support that I contacted insisted that they tried the same files and they worked. *Shrug*

I did it because PXE boot was very slow with the old BIOS. There's always something that could be better and the Changelogs are not informative enough to tell you whether there is a chance that they fixed it.
 
I do every time I notice a new bios. I never once bricked a board from a flash in 30 years.

I've literally flashed hundreds of motherboards and I've seen two get bricked. It does happen, though rarely.
 
The last time I had a mishap while flashing a BIOS was with an old Soyo SY6BA+ 440BX motherboard. Thought it was a good idea to flash the BIOS during a thunderstorm, power flickered and killed the board. Surprisingly, I was able to fix that board using a friends ASUS or ABIT board. Hot-swapped the bad BIOS chip into his running board to reflash it, and it worked!

BIOS flashing is ALOT safer today than ever before.
 
There is rarely any benefit to updating your BIOS and there is always that risk of bricking the board.
To be fair I've been doing it for any potential improvement in RAM stability etc (not that I have any issues but just in case I don't know what I would be missing out without the update since the patch notes are not detailed) because I have double sided kit with 64GB kit because I do see in the patch notes stuff about memory tuning etc. But yeah most of the time it is rare to see a benefit unless it's something specific.
 
Agreed.

It comes down to risk/benefit.

There is a small (but not negligible) risk that a flash goes wrong, and you brick it. And even though that can usually be recovered, it is usually a huge pain in the ass, up to and including sending the board to the manufacturer to be re-flashed.

It's a risk that is worth taking if you are having a problem you need to fix, but if everything is working, I'd just leave well enough alone.
Yea doing it preemptively is definitely risky. I don't know why it's so temping to slap the new bios in. It's like if I know the new bios is out, I want it. Maybe it's the false feeling of satisfaction that I've taken "care" of my rig. Not gonna lie though I do get a rush of nervousness when Installing it 😂😰
 
It's flashing a PROM - it's not a perfect operation, and stuff does go wrong. I still stand by "only if something is broken" - and then if my troubleshooting hasn't left me many options. :D
 
Status
Not open for further replies.
Back
Top