mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 04:32:44 +01:00
3655fd0fd3
There seems to be a small benefit to the legalized sequence for v2f16 round with packed instructions, so allow vectorizing it by reducing the cost. An unintended side effect is vectorization of f32 round also happens. The current FMA logic seems off to me, and isn't checking for packed instructions. |
||
---|---|---|
.. | ||
Analysis | ||
AsmParser | ||
BinaryFormat | ||
Bitcode | ||
Bitstream | ||
CodeGen | ||
DebugInfo | ||
Demangle | ||
DWARFLinker | ||
ExecutionEngine | ||
Frontend | ||
Fuzzer | ||
FuzzMutate | ||
IR | ||
IRReader | ||
LineEditor | ||
Linker | ||
LTO | ||
MC | ||
MCA | ||
Object | ||
ObjectYAML | ||
Option | ||
Passes | ||
ProfileData | ||
Remarks | ||
Support | ||
TableGen | ||
Target | ||
Testing | ||
TextAPI | ||
ToolDrivers | ||
Transforms | ||
WindowsManifest | ||
XRay | ||
CMakeLists.txt | ||
LLVMBuild.txt |