informedperson

4 ms·

1. 1.▲
   by informedperson 10y ago

   True, my mistake. (While it doesn't apply for this specific example, it's still the reason why one generally will see microcode-fixed errata go away rather quickly, independently of BIOS updates)
2. 2.▲
   by informedperson 10y ago

   Desktop parts have no dedicated AVX clock (however, their base vs. all-core turbo pretty much does the same thing), while server SKUs (Socket R3 and later) basically have two distinct clocking ladders, one for non-AVX code and one for AVX c
3. 3.▲
   by informedperson 10y ago

   Windows 7 and later distribute and activate microcode updates through Windows Update.
4. 4.▲
   by informedperson 10y ago

   A power issue would also plausibly explain why disabling SMT avoids the problem, since disabling SMT powers off a bunch of stuff (a whole thread context does need a bit of power), and, in a multi-threaded scenario, tends to reduce core usag
5. 5.▲
   by informedperson 10y ago

   Well doing that tends to halve throughput for ALU-dense code, but obviously saves a bit of die space and more importantly reduces peak power draw, which has to be supported by the metallization on the chip. AVX2 might have also been just a
6. 6.▲
   by informedperson 10y ago

   You're not wrong - the Vcore boost when both AVX pipes are powered up causes increased dissipation, but a large part of that is also caused by ... well, AVX processing huge amounts of data. For similar reasons server processors (Xeon E