I was wondering about the write endurance of current bios chips.
My motherboard has a winbond chip and after looking up their datasheets it says “More than 100,000 erase/program cycles”.
Lately I have been changing some bios settings and turning sata ports on and off as much as 4 times a week. I have 2 OSes on 2 separate drives and I prefer not to deal with dual booting them or being worried about them changing anything in each others drives.
I was wondering if continuing this way could cause issues with the bios chip since I’m writing on it each time?
Or if the 100,000 erase/program cycles could be counted on as the number of times you can write to the chip?
If I remember correctly for a programmable ROM the indication is referring to all the bits of the ROM turned to zero and than programmed.
I don’t think you’re really putting too much wear on the BIOS chip enough to kill it.
What is a bit more concerning is that each time you change settings you need to save them and reboot the system, which is not amazingly healthy due to the first voltage input to start the system.
I think you can make do with setting up the SATA ports as hot swappable so that you can change drives and stop changing settings into the BIOS constantly.
Your BIOS/uefi settings are normally stored in a CMOS chip, not on the bios chip itself. That is part of why motherboards have a battery, which is to keep the CMOS data, because it is volatile. It is also why you can take out the battery to clear the BIOS settings, because they are stored in a separate chip from the actual BIOS.
Thank you for your response. I was changing the drives by hand for the few first times I needed to. But the SATA ports are not as durable themselves and are a bit fragile and prone to wear.
Thank you. This is a really good point. I was focused on the bios chip and forgot about the CMOS battery. So is it safe to say, we can have unlimited writes to CMOS?
Think of the lifespan of the computer as a barrel filled with water. The shortest stave determines when you’ll encounter a problem.
While things like inrush current and electro-migration are a thing, that’s gonna kill your PSU (or cpu) long before your motherboard if it was designed by someone with more than two brain cells to rub together. The real harm from power cycling comes from the thermal cycling that takes place, which can eventually crack the solder connections (which is most often seen In GPUs, which see a high level of this kind of stress). The capacitors on the board also have a lifespan and then there corrosion and possible shorts from dust.
By the way, with a 100,000 write rating, you can completely rewrite the bios every day for 273 years before hitting that rating.
I really really can’t recall on where I red that. Anyone is free to call BS on it and correct me, I don’t want to go around spreading false informations.
In human terms, yes. You are not going to wear it out rebooting and changing settings.
In computer terms, almost unlimited. If you tried, you could probably wear one out by writing to it constantly at max speed, but it would still take a while.
Looks like they can be rated for over a billion writes
It was good information that rebooting does increase wear, although inrush current is probably not the main cause unless you are doing some serious overclocking.
It would be more the load/unload cycles of any hard drives(although power-saving modes are probably worse, see some WD green drives). Also, thermal cycling.
It’s there whenever you have temperature change and different materials attached together, which is why solder types that resist this are studied and developed. As a result it’s like everything else in this thread, for motherboards it’s an “issue” on a timescale outside of a human lifespan and not something to realistically worry about. GPUs on the other hand…
I outright sleep my computer every single day, as do millions of people. I’m gonna pop a cap long before I’m likely to run into other issues.
Eeprom is 10.000 to 100.000 times rewritable, depending on quality and what not. If a cell becomes faulty, it will stop being used, and an empty available instead will be used from here on. So yeah, flash all you want.
Slightly off topic, but you can get one of those bay devices that has switches to disable a data drive. I have used one for toggling between two OSes. It saves time between reboots, not having to go into the bios screen. And no having to unplug cables.
BIOS settings aren’t even stored on flash but in a tiny RAM space in the CMOS clock chip or something. (the reason why pulling the battery or shorting the pins to “reset CMOS” would also erase your settings).
For UEFI board firmware (motherboards produced after 2013 should always be UEFI) I’m not sure, but I still think it’s on a RAM space somewhere (I’ve seen it called NVRAM).
You can test this easily by pulling power plug and the coin battery, then keeping pressed power button for 10 seconds (to be sure any charge is consumed), if this resets the BIOS settings and date to default, you are NOT writing to flash but to a dedicated RAM memory embedded somewhere, and that has no write limit.
If that is the case, your question about flash endurance is irrelevant.
Personally, when I had to do a similar setup I just bought power cables for hard drives, cut and spliced them with a physical switch that would switch power to either go to one side or the other. If the drive is not powered it won’t be detected or usable at all, unsurprisingly.
There are pre-made devices that can do this job too.