Back in Mid December 2015, it was reported on Intel Communities that a group had potentially found a bug on Intel’s 6th generation CPU, codenamed Skylake. It was discovered by running certain Prime95 tests, which is a program that stress tests the CPU. When the bug is triggered, which happens sometime during the test (can be quick, can take a long time) the PC will freeze completely.
This sounds very worrying from the outset, as over the years Intel have been caught with a few different CPU bugs; back in 1994 was the first Pentium CPU’s FPU bug (I had one of these CPUs!) which caused a CPU recall, and for me personally, one racing game I couldn’t play. There was also the 1998 Pentium F00F bug, which was rather widespread.
Since the 90’s though, Major Intel CPU bugs have been very quiet. That was until 2014, when a TSX bug was discovered on Haswell and Broadwell CPUs. As it was a hard-coded bug, Intel just disabled the TSX functionality altogether – which was better than the alternative of leaving the bug in place.
There are hundreds of smaller bugs found in Intel CPUS (and AMD for that matter) but they’re usually minor, fixable or only under rare conditions. I couldn’t find a list of these bugs, just random references scattered over the internet – so if anyone can, please share!
Now, the Skylake Bug (no official name, so I’ll call it that) has been found, but doesn’t seem to be causing too much dramas. There’s no reports of problems in day to day usage, but even better is that Intel has worked out what causes the bug, and is liaising with motherboard manufacturers to push out BIOS updates to fix it. At this stage, I can’t find out if it’s disabling the feature, working around it or actually applying code and fixing it.
The first known motherboard manufacturuer to release a patch (in Beta currently) is MSI who have released the patch for two of their newer motherboards. We should see a lot more BIOS updates coming soon from all manufacturers, but the BIOS is probbaly the least patched component of a PC due to the risk and manual work required (i.e. it won’t happen by Windows Updates unless you have a Microsoft piece of hardware).
The last public comment I could find from Intel on providing more communication on this issue was that ‘by the end of this month’ (i.e. January 2016) a specification update will be communicated, ‘which will include information on this issue’.
As a side note, I thought I’d try to replicate this bug on my Lenovo ThinkPad Yoga 260 but could not reproduce after waiting a few hours. I’d already applied a BIOS update after receiving the hardware, but it’s unlikely that already contained the fix.