This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch last-successful
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 832856d109 [CRT runtime] Added functions TVMPlatformBeforeMeasurement
and TVMPlatformAfterMeasurement (#11244)
add 0b46efa33e [BUG] Disable second PlanDevices pass (#11662)
add 53d163c968 [TIR, CUDA] Add pass to replace global to shared memory
copy with cp.async (#11658)
No new revisions were added by this update.
Summary of changes:
include/tvm/tir/stmt.h | 5 +
include/tvm/tir/transform.h | 6 +
python/tvm/testing/utils.py | 7 +
python/tvm/tir/transform/transform.py | 11 ++
src/driver/driver_api.cc | 8 +
src/relay/backend/vm/compiler.cc | 6 +-
src/target/source/ptx.cc | 3 +-
src/tir/transforms/inject_ptx_async_copy.cc | 145 ++++++++++++++++
tests/python/unittest/test_tir_ptx_cp_async.py | 4 +-
.../test_tir_schedule_tensorize_ldmatrix_mma.py | 8 +-
.../test_tir_transform_inject_ptx_async_copy.py | 183 +++++++++++++++++++++
11 files changed, 373 insertions(+), 13 deletions(-)
create mode 100644 src/tir/transforms/inject_ptx_async_copy.cc
create mode 100644
tests/python/unittest/test_tir_transform_inject_ptx_async_copy.py
