[SOLVED] AMD Microcode

3rensho · 03-11-2014, 10:37 AM

Just noticed that the AMD64 site was back up after a long down time so I grabbed the latest microcode and installed it. I've got an AMD FX-6200 CPU and am running Slackware64-current. After rebooting I see that following in dmesg -

dmesg|grep microcode
[ 6.096501] microcode: CPU0: patch_level=0x06000623
[ 6.256374] microcode: CPU0: new patch_level=0x0600063d
[ 6.241885] microcode: CPU1: patch_level=0x06000623
[ 6.256554] microcode: CPU2: patch_level=0x06000623
[ 6.271080] microcode: CPU2: new patch_level=0x0600063d
[ 6.271164] microcode: CPU3: patch_level=0x06000623
[ 6.271246] microcode: CPU4: patch_level=0x06000623
[ 6.285775] microcode: CPU4: new patch_level=0x0600063d
[ 6.285855] microcode: CPU5: patch_level=0x06000623
[ 6.241885] microcode: Microcode Update Driver: v2.00 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

It looks to me like CPU's 0, 2 and 4 are updating the new code but 1,3 and 5 are not. Am I misinterpreting this output or is something amiss? I'm not sure what it is supposed to look like. Thanks for any help.

Ztcoracat · 03-25-2014, 09:27 PM

No you are not misinterpreting the output.
I noticed that 1, 3, and 5 have not been patched/updated.

Sorry I am not good with microcode.
Maybe a member with that type of experience will chime in.

I (think) that maybe a developer has not made additional patches or updates for the new AMD FX-6200.

http://nlug.ml1.co.uk/2013/06/microc...r-devices/4116
https://wiki.archlinux.org/index.php/microcode
http://www.linuxquestions.org/questi...fs-4175433777/

3rensho · 03-26-2014, 12:31 AM

Many thanks for your reply and the links. Good info to have.

Ztcoracat · 03-26-2014, 12:35 AM

Quote:

Originally Posted by 3rensho

Many thanks for your reply and the links. Good info to have.

Your very Welcome.

Glad to be a help-

273 · 03-26-2014, 02:01 AM

My CPU only seems to have the first 4 "cores" patched so you're not the only one:

Code:

dmesg |grep microcode
[    2.144448] microcode: CPU0: patch_level=0x06000629
[    2.144508] microcode: CPU1: patch_level=0x06000629
[    2.144569] microcode: CPU2: patch_level=0x06000629
[    2.144632] microcode: CPU3: patch_level=0x06000629
[    2.144695] microcode: CPU4: patch_level=0x06000629
[    2.144762] microcode: CPU5: patch_level=0x06000629
[    2.144825] microcode: CPU6: patch_level=0x06000629
[    2.144887] microcode: CPU7: patch_level=0x06000629
[    2.146065] microcode: Microcode Update Driver: v2.00 <tigran@aivazian.fsnet.co.uk>, Peter Oruba
[    3.281820] platform microcode: firmware: direct-loading firmware amd-ucode/microcode_amd_fam15h.bin
[    3.291644] microcode: CPU0: new patch_level=0x0600063d
[    3.301433] microcode: CPU1: new patch_level=0x0600063d
[    3.311241] microcode: CPU3: new patch_level=0x0600063d
[    3.319662] microcode: CPU5: new patch_level=0x0600063d

Ztcoracat · 03-26-2014, 03:28 PM

Quote:

Originally Posted by 273

My CPU only seems to have the first 4 "cores" patched so you're not the only one:

Code:

dmesg |grep microcode
[    2.144448] microcode: CPU0: patch_level=0x06000629
[    2.144508] microcode: CPU1: patch_level=0x06000629
[    2.144569] microcode: CPU2: patch_level=0x06000629
[    2.144632] microcode: CPU3: patch_level=0x06000629
[    2.144695] microcode: CPU4: patch_level=0x06000629
[    2.144762] microcode: CPU5: patch_level=0x06000629
[    2.144825] microcode: CPU6: patch_level=0x06000629
[    2.144887] microcode: CPU7: patch_level=0x06000629
[    2.146065] microcode: Microcode Update Driver: v2.00 <tigran@aivazian.fsnet.co.uk>, Peter Oruba
[    3.281820] platform microcode: firmware: direct-loading firmware amd-ucode/microcode_amd_fam15h.bin
[    3.291644] microcode: CPU0: new patch_level=0x0600063d
[    3.301433] microcode: CPU1: new patch_level=0x0600063d
[    3.311241] microcode: CPU3: new patch_level=0x0600063d
[    3.319662] microcode: CPU5: new patch_level=0x0600063d

I see--
Thanks for chiming in 273!-

I'm not sure if patching for microcode is a similar practice as patching the kernel.
IF I have time I'll study it.

gradinaruvasile · 03-26-2014, 08:59 PM

I noticed this too. I assumed that its like this because the patching is done at module, not core level.

273 · 03-27-2014, 01:57 AM

Quote:

Originally Posted by gradinaruvasile

I noticed this too. I assumed that its like this because the patching is done at module, not core level.

That could well be it, however, the previous patches seem to be applied to all cores. Perhaps this one is per module.

Ztcoracat · 03-27-2014, 09:39 PM

Quote:

Originally Posted by 273

That could well be it, however, the previous patches seem to be applied to all cores. Perhaps this one is per module.

Neat!
How many moudules = the core? (or) equal one core?

273 · 03-28-2014, 02:02 AM

Quote:

Originally Posted by Ztcoracat

Neat!
How many moudules = the core? (or) equal one core?

Each module contains 2 "cores" so in the FX-6200 there are three modules meaning 6 "cores" and in the FX-8120 there are 4 modules meaning 8 "cores".

Ztcoracat · 03-28-2014, 06:14 PM

Quote:

Originally Posted by 273

Each module contains 2 "cores" so in the FX-6200 there are three modules meaning 6 "cores" and in the FX-8120 there are 4 modules meaning 8 "cores".

So the number of modules determines the number of cores?

Or to better understand perhaps I should go look up the Fx-6200 (to find out how you knew it had 3 cores) and compare it to the Fx-8120 and read the spec's- This is intresting--

273 · 03-28-2014, 06:20 PM

https://en.wikipedia.org/wiki/List_o...icroprocessors
https://en.wikipedia.org/wiki/Bulldo...Bulldozer_core

Ztcoracat · 03-28-2014, 06:46 PM

Quote:

Originally Posted by 273

https://en.wikipedia.org/wiki/List_o...icroprocessors
https://en.wikipedia.org/wiki/Bulldo...Bulldozer_core

Thanks!