src/target/source · 997a14eda9aec3b343e742e55c3018f9dc23d8c3 · 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
..
literal		Loading commit data...
codegen_aocl.cc		Loading commit data...
codegen_c.cc		Loading commit data...
codegen_c.h		Loading commit data...
codegen_c_host.cc		Loading commit data...
codegen_c_host.h		Loading commit data...
codegen_cuda.cc		Loading commit data...
codegen_cuda.h		Loading commit data...
codegen_metal.cc		Loading commit data...
codegen_metal.h		Loading commit data...
codegen_opencl.cc		Loading commit data...
codegen_opencl.h		Loading commit data...
codegen_opengl.cc		Loading commit data...
codegen_opengl.h		Loading commit data...
codegen_source_base.cc		Loading commit data...
codegen_source_base.h		Loading commit data...
codegen_vhls.cc		Loading commit data...
codegen_vhls.h		Loading commit data...
intrin_rule_aocl.cc		Loading commit data...
intrin_rule_cuda.cc		Loading commit data...
intrin_rule_metal.cc		Loading commit data...
intrin_rule_opencl.cc		Loading commit data...
intrin_rule_opengl.cc		Loading commit data...
intrin_rule_vhls.cc		Loading commit data...
source_module.cc		Loading commit data...