Read more.Debian advisory warns users to disable hyper-threading "immediately".
Read more.Debian advisory warns users to disable hyper-threading "immediately".
Ooops that's a pretty spectacular bug there...
Old puter - still good enuff till I save some pennies!
I'll stick with my Haswell-E for the time being.
If this bug affects Skylake and is so serious, why did it take so long to find? Hm.
Alternatively it's been crashing for ages but it's taken this long to rule out the OS, drivers, programmes running, data errors, viruses, malware etc and pinpoint the Unicode as a problem.
If you had a crash every time you tried to use a pivot table in Excel (as an example I know that isn't the fault described here,) on a Windows 10 machine how many guesses at the cause would it take before you got to the CPU microcode?
Going on the Debian.org mailing list the reports of malfunctions went back to around Q2 2016 and the issue started to be investigated around Jan 2017 so it does seem to have gone unnoticed for some times as IIRC Skylake came out something like a year before the reports of malfunctions.
The issue was being investigated by the OCaml community since 2017-01-06, with reports of malfunctions going at least as far back as Q2 2016. It was narrowed down to Skylake with hyper-threading, which is a strong indicative of a processor defect. Intel was contacted about it, but did not provide further feedback as far as we know.
With all their massive R&D budget boiling down to this...
However it is quite a hard one to find. I remember reading an article about a mathematician found an error in the actual ALU in an Intel Processor for certain mathematics that returned incorrectly. He went through a phenomenal amount of testing before finding it!
Disable hyperthreading - you now have the equivalent of a couple hundred quid cheaper Intel CPU which can't do it either.
I hope the upcoming Coffee Lake does not have this bug
So will any new reviews featuring affected CPUs disable hyperthreading while running the tests until this gets patched?
Well thats a bummer of a read just as I'm convincing myself to bite the bullet on a new Kaby Lake upgrade ....
A question for those with more technical knowledge than me: Can CPU microcode be patched? Or is it purely a case of working around it?
I seem to recall an issue with early Phenom chips that couldn't be patched, the solution IIRC was BIOS updates for motherboards which could then toggle off the affected functionality.
IDK what all the fuss is about TBH, it's not like this particular bug is going to effect the vast majority of people, you'd probably have more chance of being effected by one of the 99 (odd) other bugs.
Yes and no, it depends on the bug, some can be patched to fix a problem, some they just disable the problematic function, while others are avoided by developers, Intel lists all the errors and what the workarounds are in the errata section of their technical specifications, like this one for 7th gen (PDF).
Last edited by Corky34; 29-06-2017 at 11:31 AM.
spacein_vader (29-06-2017)
There are currently 1 users browsing this thread. (0 members and 1 guests)