mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-04-19 22:39:03 +00:00
fcbb978828
* support dynamic tensor descriptor * use buffer load OOB feature for padding case * add navi support * add int8x4 inference kernel Co-authored-by: Chao Liu <chao@ixt-rack-81.local.lan> Co-authored-by: Jing Zhang <jizhan@amd.com>
26 lines
928 B
Bash
Executable File
26 lines
928 B
Bash
Executable File
rm *.ll *.s
|
|
|
|
BC_FILE=$1
|
|
|
|
/opt/rocm/llvm/bin/llvm-dis $BC_FILE -o original.ll
|
|
/opt/rocm/llvm/bin/opt -S -inline -inline-threshold=104857 original.ll > inline.ll
|
|
/opt/rocm/llvm/bin/opt -S -sroa inline.ll > sroa.ll
|
|
/opt/rocm/llvm/bin/opt -S -O3 sroa.ll > o3.ll
|
|
|
|
/opt/rocm/llvm/bin/llc -mcpu=gfx906 original.ll
|
|
/opt/rocm/llvm/bin/llc -mcpu=gfx906 inline.ll
|
|
/opt/rocm/llvm/bin/llc -mcpu=gfx906 sroa.ll
|
|
/opt/rocm/llvm/bin/llc -mcpu=gfx906 o3.ll
|
|
|
|
#/opt/rocm/llvm/bin/opt -S -O3 -sroa inline.ll > o3.ll
|
|
#/opt/rocm/llvm/bin/opt -S -O3 -sroa o3.ll > o3_2.ll
|
|
#/opt/rocm/llvm/bin/opt -S -O3 -sroa o3_2.ll > o3_3.ll
|
|
#/opt/rocm/llvm/bin/opt -S -O3 -sroa o3_3.ll > o3_4.ll
|
|
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 opt.ll
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 inline.ll
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 o3.ll
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 o3_2.ll
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 o3_3.ll
|
|
#/opt/rocm/llvm/bin/llc -mcpu=gfx908 o3_4.ll
|