This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from b3b6024027 [Relax] Add FRelaxInferLayout for gather_elements operator
(#18642)
add fa905d2b69 [Compile] accelerate compilation speed using NVRTC (#18519)
No new revisions were added by this update.
Summary of changes:
cmake/modules/CUDA.cmake | 2 -
cmake/utils/FindCUDA.cmake | 9 -
docker/Dockerfile.ci_gpu | 3 +
...ll_gluoncv.sh => ubuntu_install_cuda_python.sh} | 2 +-
python/tvm/contrib/nvcc.py | 329 +++++++++++++++++++--
.../tvm/script/ir_builder/tir/external_kernel.py | 24 +-
src/runtime/contrib/nvshmem/init.cc | 18 +-
src/target/opt/build_cuda_on.cc | 122 ++------
src/target/source/codegen_cuda.cc | 8 +-
src/target/source/literal/cuda_half_t.h | 8 +-
tests/python/codegen/test_target_codegen_cuda.py | 32 +-
tests/python/disco/test_nvshmem.py | 30 +-
.../test_tir_transform_inject_ptx_async_copy.py | 7 +
13 files changed, 443 insertions(+), 151 deletions(-)
copy docker/install/{ubuntu_install_gluoncv.sh =>
ubuntu_install_cuda_python.sh} (97%)
mode change 100755 => 100644
