This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch unity
in repository https://gitbox.apache.org/repos/asf/tvm.git
omit 7c88732056 [Unity][Pass] Wellformed Analysis (#14032)
omit ae8f9e373b [Unity][BlockBuilder] CallTE convert PrimValue args
(#14028)
omit bb5c8df4c7 [Unity][Pass] Normalize Pass (#14031)
omit 705753df25 [Unity] Relay -> Relax translator (#14026)
omit 9e36bb15c4 [Unity][Pass][TuningAPI] Introduce TuningAPI and
MetaSchedule pass (#14014)
omit ba475010db [Unity][Pass] BindParams pass, FoldConstant pass (#14016)
omit 9af6afccbb [Unity][VM] Supporting "compiled" exec mode. (#14015)
omit ab43ba8ae4 [Unity][Pass] LambdaLift pass (#14012)
omit 85b8a419ab [Unity][Pass] Operator Fusion Passes (#14001)
omit e842690f8a [Unity] NestedMsg Support utility (#13995)
omit a62fcd4470 [Unity] Relax op: manipulation (#13989)
omit dbc81a0d67 [Unity] Relax op: search (#13992)
omit 58bcc122cd [Unity] Relax op: linear algebra (#13988)
omit d79b0682fa [Unity] Relax op: creation (#13984)
omit 9925df67e0 [Unity] Relax op: neural networks (#13993)
omit ac9af7ebc7 [Unity] Relax op: statistical (#13991)
omit 7df7b048bf [Unity] Relax op: arithmetic, comparison (#13983)
omit 3454010052 [Unity] Relax op: image (#13994)
omit c35386efdc [Unity] Relax op: set (#13990)
omit 7ae3d773ed [Unity] Relax op: datatype (#13986)
omit 547d3ac357 [Unity] Relax op: index (#13987)
omit 895fb6cf21 [Unity][TVMScript] Use explicit `R.shape` in TVMScript
(#13979)
omit 202caefb28 [Unity] e2e Relax minimum build flow (#13961)
omit 0e24b127c3 [Unity] Relax VM shape lowering pass (#13956)
omit 3561b631aa [Unity] Relax VM codegen (#13954)
omit aae3011397 [Unity] Relax TVMScript Printer (#13944)
omit 34f2a0df4e [Unity] Relax TVMScript Parser. (#13932)
omit abe2677391 [Unity] Relax BlockBuilder and ExprMutator (#13926)
omit 15f7d55839 [Unity] Basic StructInfo Analysis and Expr construction
(#13916)
omit cd399ed507 [Unity][CI] Unity specific jenkins setup (do not upstream
to main) (#13910)
omit f93d56ccbe [Unity][IR] First-class StructInfo (#13907)
omit 3806a2e348 [Unity] Relax expressions and types (#13901)
omit 50b9504a64 [Unity] Relax VM (#13878)
add f6cebb5d60 [microTVM] additional refactoring for enabling USE_MICRO in
more builds (#13909)
add e516eaaea1 [testing][py_converter] Enhance py_converter to better
support entire modules (#13769)
add affcc11c2e [Keras] Fix importing conv2d_transpose for NHWC layout
(#13998)
add 1e5a830784 [microTVM]Refactor crt_config.h header file generation
(#13955)
add a6f4694bf9 [CI][Flaky] Skip zephyr_qemu-x86 tests that are part of
task_python_microtvm (#14005)
add 967d2bb807 [Build] Fix find_include_path when using TVM python package
(#14007)
add aa927df6b3 [Metaschedule] EvolutionarySearchNode::State constructor
typo fix (#14002)
add e7ad4bc071 [CLML] Changes corresponding to OpenCL workspace
refactorization (#13972)
add 0dd3d4aab6 [ACL][TESTING] Use pytest.mark.parametrize in ACL conv2d
tests (#14011)
add 0c965f442a [ONNX] Support ScatterElements with reduction (#13894)
add d7253fbc7e [TIR] Add cp.async support for tir.if_then_else (#13966)
add d12a636897 Refactor test to make it easier for user to understand how
tensor_intrin works (#14017)
add b1397367fc [FIX][RUNTIME] Convert container with function value type
(#14024)
add a543f6bebf [Arith] Implemented PMatchesOneOf and matches_one_of
(#13933)
add 14bc5e4585 [FIX][TOPI] Clip with IntImm/FloatImm (#14027)
add f4a0323c22 [Unity] Relax VM (#13878)
add 13af1f26fa [Unity] Relax expressions and types (#13901)
add 80a32cb0bc [Unity][IR] First-class StructInfo (#13907)
add ad41f97f00 [Unity][CI] Unity specific jenkins setup (do not upstream
to main) (#13910)
add d56a12aa49 [Unity] Basic StructInfo Analysis and Expr construction
(#13916)
add abdd35aac2 [Unity] Relax BlockBuilder and ExprMutator (#13926)
add 48a3423248 [Unity] Relax TVMScript Parser. (#13932)
add 81ebb3f938 [Unity] Relax TVMScript Printer (#13944)
add 998069c519 [Unity] Relax VM codegen (#13954)
add afe63b9a72 [Unity] Relax VM shape lowering pass (#13956)
add 7c66a535eb [Unity] e2e Relax minimum build flow (#13961)
add 568ceebaf2 [Unity][TVMScript] Use explicit `R.shape` in TVMScript
(#13979)
add 342b670cde [Unity] Relax op: index (#13987)
add d9d79c0be9 [Unity] Relax op: datatype (#13986)
add d4db3c092b [Unity] Relax op: set (#13990)
add 08cbcdf11f [Unity] Relax op: image (#13994)
add 716f3aca7c [Unity] Relax op: arithmetic, comparison (#13983)
add 7d8095e8be [Unity] Relax op: statistical (#13991)
add 9a7a5fbe10 [Unity] Relax op: neural networks (#13993)
add b3de34f1f2 [Unity] Relax op: creation (#13984)
add 15800607f5 [Unity] Relax op: linear algebra (#13988)
add e59b4e78fd [Unity] Relax op: search (#13992)
add 1991cecbbe [Unity] Relax op: manipulation (#13989)
add 5bf538fa5b [Unity] NestedMsg Support utility (#13995)
add dc5612fa0a [Unity][Pass] Operator Fusion Passes (#14001)
add 13402fc028 [Unity][Pass] LambdaLift pass (#14012)
add dd5cfaa0bc [Unity][VM] Supporting "compiled" exec mode. (#14015)
add 47f8c8df58 [Unity][Pass] BindParams pass, FoldConstant pass (#14016)
add 70c645d4ef [Unity][Pass][TuningAPI] Introduce TuningAPI and
MetaSchedule pass (#14014)
add 5dd5411d9b [Unity] Relay -> Relax translator (#14026)
add c6c48deb94 [Unity][Pass] Normalize Pass (#14031)
add cc3eb0c8e6 [Unity][BlockBuilder] CallTE convert PrimValue args
(#14028)
add d80451bc3f [Unity][Pass] Wellformed Analysis (#14032)
This update added new revisions after undoing existing revisions.
That is to say, some revisions that were in the old version of the
branch are not in the new version. This situation occurs
when a user --force pushes a change and generates a repository
containing something like this:
* -- * -- B -- O -- O -- O (7c88732056)
\
N -- N -- N refs/heads/unity (d80451bc3f)
You should already have received notification emails for all of the O
revisions, and so the following emails describe only the N revisions
from the common base, B.
Any revisions marked "omit" are not gone; other references still
refer to them. Any revisions marked "discard" are gone forever.
No new revisions were added by this update.
Summary of changes:
CMakeLists.txt | 13 +-
apps/bundle_deploy/Makefile | 14 +-
apps/bundle_deploy/build_model.py | 8 +
apps/bundle_deploy/crt_config/crt_config.h | 53 --
.../template_project/crt_config/crt_config.h | 57 ---
apps/microtvm/cmsisnn/Makefile | 1 +
apps/microtvm/cmsisnn/convert_image.py | 8 +
apps/microtvm/cmsisnn/include/crt_config.h | 26 -
apps/microtvm/ethosu/Makefile | 2 +
apps/microtvm/ethosu/convert_labels.py | 7 +
apps/microtvm/ethosu/include/crt_config.h | 26 -
.../template_project/crt_config/crt_config.h | 67 ---
apps/microtvm/zephyr_cmsisnn/CMakeLists.txt | 1 +
apps/microtvm/zephyr_cmsisnn/include/crt_config.h | 54 --
.../zephyr_cmsisnn/model/convert_labels.py | 14 +
cmake/modules/Arduino.cmake | 81 ---
cmake/modules/CRT.cmake | 79 ---
cmake/modules/Micro.cmake | 102 ++++
cmake/modules/StandaloneCrt.cmake | 97 ++--
.../{modules/Metal.cmake => utils/CRTConfig.cmake} | 34 +-
include/tvm/relay/attrs/transform.h | 12 +
include/tvm/runtime/c_runtime_api.h | 7 +
include/tvm/runtime/crt/microtvm_rpc_server.h | 2 +
python/tvm/_ffi/libinfo.py | 8 +-
python/tvm/micro/__init__.py | 1 +
python/tvm/micro/build.py | 17 +
python/tvm/relay/frontend/keras.py | 2 +
python/tvm/relay/frontend/onnx.py | 52 +-
python/tvm/relay/op/_transform.py | 10 +
python/tvm/relay/op/op_attrs.py | 5 +
python/tvm/relay/op/strategy/cuda.py | 14 +
python/tvm/relay/op/strategy/generic.py | 22 +
python/tvm/relay/op/transform.py | 35 ++
python/tvm/relay/testing/py_converter.py | 83 +++-
python/tvm/runtime/object_generic.py | 15 +-
python/tvm/testing/aot.py | 7 +-
python/tvm/topi/__init__.py | 1 +
python/tvm/topi/cuda/__init__.py | 1 +
python/tvm/topi/cuda/scatter_elements.py | 167 +++++++
python/tvm/topi/scatter_elements.py | 150 ++++++
src/arith/pattern_match.h | 106 +++-
src/arith/rewrite_simplify.cc | 546 ++++++++++-----------
.../search_strategy/evolutionary_search.cc | 2 +-
src/relay/op/tensor/transform.cc | 56 +++
src/runtime/container.cc | 1 +
src/runtime/contrib/clml/clml_runtime.cc | 127 ++---
src/runtime/crt/common/crt_runtime_api.c | 8 +-
...crt_config-template.h => crt_config.h.template} | 38 +-
src/runtime/crt/host/microtvm_api_server.py | 2 +-
src/runtime/micro/crt_config.h | 58 ---
src/runtime/opencl/opencl_common.h | 2 +
src/target/source/codegen_cuda.cc | 8 +-
src/target/source/ptx.cc | 31 ++
src/target/source/ptx.h | 16 +
src/tir/ir/expr.cc | 4 +-
src/tir/transforms/inject_ptx_async_copy.cc | 156 +++---
tests/lint/check_file_type.py | 1 +
tests/micro/common/test_tvmc.py | 1 +
tests/micro/stm32/test_code_emitter.py | 5 +-
.../test_runtime_packed_func.py | 13 +
.../contrib/test_arm_compute_lib/test_conv2d.py | 395 +++++++--------
tests/python/frontend/keras/test_forward.py | 11 +
tests/python/frontend/onnx/test_forward.py | 80 ++-
tests/python/relay/test_py_converter.py | 65 ++-
.../unittest/test_cp_async_in_if_then_else.py | 238 +++++++++
tests/python/unittest/test_te_schedule.py | 23 +-
66 files changed, 2090 insertions(+), 1258 deletions(-)
delete mode 100644 apps/bundle_deploy/crt_config/crt_config.h
delete mode 100644
apps/microtvm/arduino/template_project/crt_config/crt_config.h
delete mode 100644 apps/microtvm/cmsisnn/include/crt_config.h
delete mode 100644 apps/microtvm/ethosu/include/crt_config.h
delete mode 100644
apps/microtvm/zephyr/template_project/crt_config/crt_config.h
delete mode 100644 apps/microtvm/zephyr_cmsisnn/include/crt_config.h
delete mode 100644 cmake/modules/Arduino.cmake
delete mode 100644 cmake/modules/CRT.cmake
copy cmake/{modules/Metal.cmake => utils/CRTConfig.cmake} (50%)
create mode 100644 python/tvm/topi/cuda/scatter_elements.py
create mode 100644 python/tvm/topi/scatter_elements.py
rename src/runtime/crt/{crt_config-template.h => crt_config.h.template} (57%)
delete mode 100644 src/runtime/micro/crt_config.h
create mode 100644 tests/python/unittest/test_cp_async_in_if_then_else.py
