Update CPU Microcode

[Ritchie]

Hi

Not sure whether this was the prefereable section to place this topic, but

On Gigabyte's website, I noticed that BIOS updates for some boards I use, such as the Slot 1 686BX and 686BXE boards, include a note in the updates saying Update CPU Microcode.

What does this refer to?

It sounds like updating the main code to run the board or to run the CPU. I was wondering if it even flashed something to the CPU installed to give the CPU more efficient code itself as this is what the description sounds like, although I didn't previously think that a CPU could be flashed (at least not by an end user!).

I recently updated a 686BX BIOS which had a newly installed 40GB HDD in there (hence one main reason for the upgrade) and with 256MB RAM and Windows 2000 and only a 266MHz PII CPU, the system seems to run very fast. I was surprised at the performance after the update, maybe the fastest Windows 2000 on that speed PII that I had seen.

So any enlightment on what people know or think that part of the BIOS upgrade might do would be very interesting to know.

Thanks

[Denniss]

Update Microcode means the Bios knows more Processors and how to load the Microcode-Update for them .
An unknown P2-550MHz will be recognizes as P3-550 for example .

[ajzchips]

The speed improvement you've noticed is basically thanks to the new HDD.

CPU microcodes are an inherent part of the system BIOS, enabling it to recognize and correctly initialize newer CPUs (speed, core, etc.)

Your P2-266 is quite a "classic" already, so most probably no improved support has been added to it.

[Ritchie]

So if I understand correctly, the CPU microcode is only a part of the BIOS that allows it to correctly initialise new CPUs, and when you refer to loading the microcode update this is simply the process of correctly initialising the CPU?

And the CPU microcode has nothing to do with the code in the CPU itself, or the mainboard circuitry or chips on it, and so should have no influence on how efficiently and quickly the system runs? What about the way in which the CPU is initialised, could different intialisation methods have any impact on how well the CPU or system runs or maybe only an effect in how quickly and efficiently the CPU identification and initilisation takes place during the BIOS boot.

So it sounds like it was only a co-incidence that the system seemed to run a lot quicker after the upgrade. Unless other things were done in the BIOS upgade that had an impact. But maybe more likely, maybe the hard disk installed was a very fast IDE/ATA disk and was well suited to that motherboard, CPU and other system components.

[NickS]

So if I understand correctly, the CPU microcode is only a part of the BIOS that allows it to correctly initialise new CPUs, and when you refer to loading the microcode update this is simply the process of correctly initialising the CPU?

Not quite. It is included in the BIOS, allows the BIOS to identify the CPU by means of the CPUID and therefore to initialise it correctly. It *may* also contain changes to the actual microcode of the CPU, depending on whether or not they got it right in the first place (remember the infamous Pentium floating point bug ?) as Denniss intimated.

And the CPU microcode has nothing to do with the code in the CPU itself,

It may.

or the mainboard circuitry or chips on it,

No.

and so should have no influence on how efficiently and quickly the system runs? What about the way in which the CPU is initialised, could different intialisation methods have any impact on how well the CPU or system runs or maybe only an effect in how quickly and efficiently the CPU identification and initilisation takes place during the BIOS boot.

Ajzchips's point was that your CPU is quite "mature", so unless you had a very early version of the microcode initally there is unlikely to have been much optimisation. Where Intel actually updated the microcode for a given CPU it was generally a bug fix anyway.

So it sounds like it was only a co-incidence that the system seemed to run a lot quicker after the upgrade. Unless other things were done in the BIOS upgade that had an impact.

Always possible, maybe someone screwed up cache handling or UDMA in an earlier BIOS.

But maybe more likely, maybe the hard disk installed was a very fast IDE/ATA disk and was well suited to that motherboard, CPU and other system components.

More than likely the HDD - especially if moving to a 7200 rpm UDMA drive with a 2 Mbyte buffer from a PIO Mode 4 5400 rpm drive with 32 Kbyte buffer?

[Ritchie]

Thanks for the detailed response Nick - that is quite clear now.