src/target · 50b5adaac2956712d65e14c163148c1e6279f5e2 · wenyuanbo / tic

[CodeGen][CUDA] Vectorization for intrinsics (#5101) · 05b0f7e0

- This allows to emit vectorized loads/stores
  for CUDA math intrinsics.

- A few intrinsics should be lowered as CUDAMath not CUDAFastMath ones.

- Fixed the code block identation.

committed Mar 22, 2020

05b0f7e0

Name	Last commit	Last update
..
datatype		Loading commit data...
llvm		Loading commit data...
opt		Loading commit data...
source		Loading commit data...
spirv		Loading commit data...
stackvm		Loading commit data...
build_common.h		Loading commit data...
codegen.cc		Loading commit data...
generic_func.cc		Loading commit data...
intrin_rule.cc		Loading commit data...
intrin_rule.h		Loading commit data...
target.cc		Loading commit data...
target_info.cc		Loading commit data...