Monday, January 25th 2016

ASRock Releases BIOS Update to Fix Intel "Skylake" Bug

Earlier, Intel confirmed that its Skylake processors were plagued by a bug which causes the system to freeze or act abnormally when executing complex workloads, such as the stress test benchmark Prime95. While the majority of computer users might not use Prime95 at all, or never had the urge to stress their system to the point of discovering this Intel Skylake bug, still ASRock has already released BIOS updates for fixing this flaw on all of their Intel 100 Series motherboards. So, on the bright side, after a quick ASRock BIOS update you will never notice or remember Intel's small hiccup.
Add your own comment

20 Comments on ASRock Releases BIOS Update to Fix Intel "Skylake" Bug

#1
RejZoR
What about H.264 encoded using AVX ? Are Skylakes prone to acting "abnormally" in those conditions as well? Because video encoding is nothing special compared to synthetic torture of Prime95?
Posted on Reply
#2
iO
RejZoR said:
What about H.264 encoded using AVX ? Are Skylakes prone to acting "abnormally" in those conditions as well? Because video encoding is nothing special compared to synthetic torture of Prime95?
Afaik the error is specific to prime numbers not heavy AVX workloads per se.
Posted on Reply
#3
newtekie1
Semi-Retired Folder
iO said:
Afaik the error is specific to prime numbers not heavy AVX workloads per se.
It is an AVX bug. To get it to happen you have to mod Prime95 to get it to use AVX, otherwise Prime95 works just fine. I believe the bug is caused by heavy AVX load on the CPU.

I just hope this patch doesn't hinder AVX performance, and kill BCLK overclocking...
Posted on Reply
#4
RejZoR
I'm asking that because I know many people got super stable overclocks in all torture programs, but same system got unstable when encoding longer H.264 videos using AVX extension.
Posted on Reply
#6
cracklez
AWESOME!!!!

Not that I've had any problems with my baby, but it's good to know ASRock did their homework and found a fix for it.
Posted on Reply
#7
hojnikb
cracklez said:
AWESOME!!!!

Not that I've had any problems with my baby, but it's good to know ASRock did their homework and found a fix for it.
Thats a microcode update, so asrock has nothing to do with it.
Posted on Reply
#8
Parn
I hope this microcode update is not going to partially disable AVX for certain workloads.

The famous TSX bug in Haswell was sorted out by disabling it through microcode update. So I hope it doesn't repeat with this (disabling a feature != fixing a bug in my book).
Posted on Reply
#9
R-T-B
hojnikb said:
Thats a microcode update, so asrock has nothing to do with it.
Indeed. And the microcode in question has been out since october last year. This is a case of ALL mobo vendors NOT doing their homework.
Posted on Reply
#10
RejZoR
Parn said:
I hope this microcode update is not going to partially disable AVX for certain workloads.

The famous TSX bug in Haswell was sorted out by disabling it through microcode update. So I hope it doesn't repeat with this (disabling a feature != fixing a bug in my book).
Wasn't TSX bug also present on Skylakes, making that extension useless and afaik they even disabled it?
Posted on Reply
#11
Parn
RejZoR said:
Wasn't TSX bug also present on Skylakes, making that extension useless and afaik they even disabled it?
I thought I heard it had been fixed in Haswell-EP/EX and Broadwell onwards. Maybe there has been another update to the matter which I'm not aware of.
Posted on Reply
#12
bug
Jizzler said:
Nice, flash done.
Does it retain your previous settings or do I have to take care of them myself?

TIA
Posted on Reply
#13
R-T-B
RejZoR said:
Wasn't TSX bug also present on Skylakes, making that extension useless and afaik they even disabled it?
No. It is an on-silicon bug in Haswell/Haswell-E only.
Posted on Reply
#14
RejZoR
I'm pretty certain Skylake was mentioned before as well. It's pretty lame they can't fix this stuff in 3 major CPU revisions...
Posted on Reply
#15
hojnikb
RejZoR said:
I'm pretty certain Skylake was mentioned before as well. It's pretty lame they can't fix this stuff in 3 major CPU revisions...
hardware bugs are not something you can fix in one day.
Posted on Reply
#16
RejZoR
hojnikb said:
hardware bugs are not something you can fix in one day.
Haswell-K (and neither the Haswell-E) wasn't exactly released "few days ago"...
Posted on Reply
#17
hojnikb
RejZoR said:
Haswell-K (and neither the Haswell-E) wasn't exactly released "few days ago"...
Well, i willing to bet that skylake design was well underway when haswell was released. CPU micro architectures are not exactly made overnight.
Posted on Reply
#18
looniam
anyone know if there is a change in BLCK adjusting for non-K processors?
Posted on Reply
#19
Jizzler
bug said:
Does it retain your previous settings or do I have to take care of them myself?

TIA
I don't remember off the top of my head if can do a complete Save All but on the OC Tweaker tab there is an option to save your CPU/Mem settings to disk. Would be nice if I was on Z170 with a K CPU but since my system is H170 with an i5-6600 I really only have 1 setting to change after a BIOS update - fan profile set to Silent from Standard - and I'm done.
Posted on Reply
#20
RejZoR
hojnikb said:
Well, i willing to bet that skylake design was well underway when haswell was released. CPU micro architectures are not exactly made overnight.
Still, I highly doubt that it was so far they couldn't fix it before 3rd iteration was released?
Posted on Reply
Add your own comment