forked from OSchip/llvm-project
d073b19dbf
Noticed while trying to clean up the shift costs model for SSE4 targets using the script in D10369 - SLM double-pumps all the 128-bit vector conversion ops and only use FP0 pipe - numbers taken from Intel AOM + Agner. |
||
|---|---|---|
| .. | ||
| resources-aes.s | ||
| resources-cmov.s | ||
| resources-cmpxchg.s | ||
| resources-lea.s | ||
| resources-mmx.s | ||
| resources-movbe.s | ||
| resources-pclmul.s | ||
| resources-popcnt.s | ||
| resources-prefetchw.s | ||
| resources-rdrand.s | ||
| resources-rdseed.s | ||
| resources-sse1.s | ||
| resources-sse2.s | ||
| resources-sse3.s | ||
| resources-sse41.s | ||
| resources-sse42.s | ||
| resources-ssse3.s | ||
| resources-x86_32.s | ||
| resources-x86_64.s | ||
| resources-x87.s | ||