Re:Spectre mitigations murder userspace performance
by p_l on Sunday June 20, 2021 @09:04AM
(#27569552)
The techniques used to make CPUs fast have possible vulnureable areas (not all of it is due to Out-Of-Order Execution, some was due to optimizations made on top).
The mitigations introduced stalls into those mechanism, clearing some of the state, resulting in slowdowns.
However, the mitigations are not as big slowdowns as not having the speculative execution mechanism at all.
Re:Spectre mitigations murder userspace performance
by dkarras on Sunday June 20, 2021 @07:08AM
(#27568948)
I'm not sure if I got you right: The performance boost with speculative execution is massive. It introduces a vuln. It is the fix for that vuln that kills the performance.
Re:Spectre mitigations murder userspace performance
by Popegaf on Sunday June 20, 2021 @10:03AM
(#27570070)
What I'm saying is if the fix for the vuln negates the performance boost from speculative execucution, then the net outcome is 0 performance boost and speculative execution might as well be removed entirely. It wouldn't surprise me if it just costs more energy to keep the vuln and mitigation active at the same time.
Re:Spectre mitigations murder userspace performance
by SAI_Peregrinus on Sunday June 20, 2021 @04:01PM
(#27573388)
The fix doesn't negate the performance boost. It decreases it. But it's still nonzero boost.
Are there CPUs out there / being made without these SoC predictions?