Re:Performance Debugging with LLVM-mca: Simulating the CPU
by fossa1 on Wednesday July 02, 2025 @08:51AM
(#44443140)
This is a textbook case of micro-architectural reality beats theoretical elegance. It's fascinating how replacing 5 loads with 2 loads + 3 vextq_f32 intrinsics, which should reduce memory pressure, ends up being slower due to execution port contention and dependency chains.
[ Parent
| on HN ]
There are no comments on this comment yet.