Message from @lidl papi
Discord ID: 644514987175313408
HLE is about machine code
it's an instruction
yeah but all thing i know have had problems/used it have had something to do with C so i was just quessing
HLE adds two instruction prefixes
which through some black magic fuckery are backwards compatible
so they can remove HLE without breaking anything
since it'll just fall back
the performance loss that tweet is talking about is unrelated to HLE instructions itself
the microcode update changed something relating to how jump instructions are decoded
the problem this tweet is talking about
compilers can know about this issue and generate different code
to avoid crossing 32 byte boundaries or ending up on a 32 byte boundary
for negligible side effects (basically just add a nop or reorder code)
but the problem is
everything has to be recompiled for that to happen
hence things compiled before now might try to do that and see reduced performance
and the way almost all modern software works
no one actually compiles for different cpus
unless you're some arch turboautist
Same difference
it's not, you don't build from source on arch unless outside package manageer
Performance in games seems to be in line with the 0-4% decrease Intel states
Mainly because games are GPU limited anyways
INTEL NO 1
>when desktop CPUs are more powerful than your 10 times more expensive server CPUs
LOL!
>actually taking passmark as a representative benchmark
LOL!
LOL!
you're laughing now until you realize that with the current going AMD users will get the same performance impact
yeah when that happens
it's already happening
intel submitted the patches to GCC today
Current Intel hardware mitigations do not cover TAA and current Cascade Lake CPUs remain vulnerable. TAA can allow leaking of data across processes, privilege boundaries and Hyper Threading. With Hyper Threading disabled, TAA can still leak data from protected domains.
yikes imagine leaking data even when HT is off
also the next time you cherrypick benchmarks
>In this case the assembler update didn't make any difference as Firefox wasn't rebuilt from source as part of the test profile or the Clear Linux revision.