build.sh

# -DCMAKE_INSTALL_PREFIX=/home/mlevental/dev_projects/iree/iree-install
#-DIREE_BUILD_PYTHON_BINDINGS=ON
#-DIREE_ENABLE_ASSERTIONS=ON
#-DCMAKE_C_COMPILER=clang-18
#-DCMAKE_CXX_COMPILER=clang++-18
#-DIREE_ENABLE_LLD=ON
#-DCMAKE_C_COMPILER_LAUNCHER=ccache
#-DCMAKE_CXX_COMPILER_LAUNCHER=ccache
#-DMLIR_ENABLE_BINDINGS_PYTHON=ON
#-DIREE_ENABLE_RUNTIME_TRACING=ON
#-DIREE_TRACING_PROVIDER=console
#-DIREE_TRACING_CONSOLE_FLUSH=1
#-DIREE_TRACING_MODE=4
#-DIREE_BUILD_TESTS=ON
#-DBENCHMARK_ENABLE_TESTING:BOOL=ON
#-DIREE_TARGET_BACKEND_ROCM=ON
#-DIREE_HIP_TEST_TARGET_CHIP=gfx1100
#-DCMAKE_C_FLAGS="-DIREE_VM_EXECUTION_TRACING_ENABLE=1 -DIREE_VM_EXECUTION_TRACING_FORCE_ENABLE=1-DIREE_VM_EXECUTION_TRACING_SRC_LOC_ENABLE=1"

rm -rf dump
mkdir -p dump

/home/mlevental/miniconda3/envs/IREE/bin/python \
  /home/mlevental/dev_projects/iree/tests/e2e/matmul/generate_e2e_matmul_tests.py \
  --output_matmul_mlir=$PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.mlir \
   --output_calls_mlir=$PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.mlir \
   --lhs_rhs_type=f16 --acc_type=f32 --compilation_info=LLVMGPUVectorDistributeWMMA --shapes=easy_large_static

/home/mlevental/dev_projects/iree/cmake-build-debug/tools/iree-compile \
  --output-format=vm-bytecode \
  --mlir-print-op-on-diagnostic=false \
  --iree-hal-target-backends=rocm \
  --iree-hip-target=gfx1100 \
  --iree-hal-dump-executable-files-to=$PWD/dump \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.mlir -o \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.vmfb \
  --iree-hal-executable-object-search-path="/home/mlevental/dev_projects/iree/cmake-build-debug"

/home/mlevental/dev_projects/iree/cmake-build-debug/tools/iree-compile \
  --output-format=vm-bytecode \
  --mlir-print-op-on-diagnostic=false \
  --iree-hal-target-backends=rocm \
  --iree-hip-target=gfx1100 \
  --iree-hal-dump-executable-files-to=$PWD/dump \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.mlir -o \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.vmfb \
  --iree-hal-executable-object-search-path="/home/mlevental/dev_projects/iree/cmake-build-debug"

/home/mlevental/dev_projects/iree/cmake-build-debug/tools/iree-compile \
  --output-format=vm-bytecode \
  --iree-vm-bytecode-module-output-format=flatbuffer-text \
  --mlir-print-op-on-diagnostic=false \
  --iree-hal-target-backends=rocm \
  --iree-hip-target=gfx1100 \
  --iree-hal-dump-executable-files-to=$PWD/dump \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.mlir -o \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.vmfb.txt \
  --iree-hal-executable-object-search-path="/home/mlevental/dev_projects/iree/cmake-build-debug"

/home/mlevental/dev_projects/iree/cmake-build-debug/tools/iree-compile \
  --output-format=vm-bytecode \
  --iree-vm-bytecode-module-output-format=flatbuffer-text \
  --mlir-print-op-on-diagnostic=false \
  --iree-hal-target-backends=rocm \
  --iree-hip-target=gfx1100 \
  --iree-hal-dump-executable-files-to=$PWD/dump \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.mlir -o \
  $PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.vmfb.txt \
  --iree-hal-executable-object-search-path="/home/mlevental/dev_projects/iree/cmake-build-debug"

/home/mlevental/dev_projects/iree/cmake-build-debug/tools/testing/e2e/iree-e2e-matmul-test \
  --module=$PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_matmul.vmfb \
  --module=$PWD/dump/e2e_matmul_rocm_f16_large_rdna3_wmma_rocm_hip_calls.vmfb \
  --require_exact_results=false \
  --acceptable_fp_delta=1e-04 --device=hip