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 84d05e9a24 [LLVM][Codegen] Cast NaN to bool gives true (#18772)
add 82b01c9486 [REFACTOR][S-TIR] More migrations to s-tir (#18776)
add 2a448ce4bb [CI] Update Pillow, pytest-rerunfailures, junitparser,
xgboost, onnx and pytorch (#18777)
add f8dbbdaf5c [REFACTOR] Phase out dmlc dep (#18779)
No new revisions were added by this update.
Summary of changes:
3rdparty/nvbench/l2_cache_flush.h | 2 +-
CMakeLists.txt | 16 +-
Makefile | 5 +-
apps/android_rpc/app/src/main/jni/Android.mk | 1 -
apps/android_rpc/app/src/main/jni/tvm_runtime.h | 1 -
apps/cpp_rpc/CMakeLists.txt | 1 -
apps/cpp_rpc/main.cc | 3 -
apps/cpp_rpc/rpc_env.cc | 2 +-
apps/cpp_rpc/rpc_server.cc | 2 -
apps/hexagon_launcher/cmake/HexagonLauncher.cmake | 1 -
apps/hexagon_launcher/cmake/android/CMakeLists.txt | 1 -
apps/hexagon_launcher/cmake/hexagon/CMakeLists.txt | 2 -
apps/hexagon_launcher/launcher_core.cc | 88 ++--
apps/hexagon_launcher/launcher_core.h | 10 +-
apps/ios_rpc/tvmrpc.xcodeproj/project.pbxproj | 4 -
cmake/modules/CUDA.cmake | 1 -
cmake/modules/LLVM.cmake | 4 +-
cmake/modules/LibInfo.cmake | 1 -
cmake/modules/contrib/CUTLASS.cmake | 2 -
cmake/modules/contrib/MSCCLPP.cmake | 1 -
docker/install/ubuntu_install_onnx.sh | 65 +--
docker/install/ubuntu_install_python_package.sh | 6 +-
docker/install/ubuntu_install_redis.sh | 2 +-
docs/Doxyfile | 2 +-
include/tvm/ir/attrs.h | 1 -
include/tvm/ir/op.h | 2 +-
include/tvm/runtime/disco/session.h | 2 +
include/tvm/runtime/logging.h | 20 +-
include/tvm/runtime/module.h | 2 +-
include/tvm/runtime/serializer.h | 49 +-
include/tvm/runtime/tensor.h | 29 +-
include/tvm/runtime/vm/executable.h | 39 +-
include/tvm/s_tir/analysis.h | 121 +++++
.../tvm/{tir => s_tir}/backend/adreno/transform.h | 23 +-
include/tvm/s_tir/schedule/instruction.h | 2 +-
include/tvm/s_tir/transform.h | 41 ++
include/tvm/support/io.h | 129 +++++
include/tvm/support/serializer.h | 232 +++++++++
include/tvm/support/with.h | 4 +-
include/tvm/target/tag.h | 2 +-
include/tvm/target/target_info.h | 77 ---
include/tvm/target/target_kind.h | 2 +-
include/tvm/tir/analysis.h | 145 ------
include/tvm/tir/builtin.h | 13 -
include/tvm/tir/stmt_functor.h | 9 -
include/tvm/tir/transform.h | 163 -------
python/tvm/contrib/hexagon/meta_schedule.py | 2 +-
python/tvm/contrib/hexagon/tools.py | 2 -
python/tvm/contrib/msc/core/runtime/runner.py | 2 +-
.../contrib/msc/framework/tvm/runtime/runner.py | 4 +-
python/tvm/libinfo.py | 5 -
python/tvm/relax/frontend/nn/extern.py | 7 -
python/tvm/s_tir/__init__.py | 19 +
python/tvm/s_tir/analysis/__init__.py | 96 +++-
python/tvm/s_tir/backend/adreno/__init__.py | 1 +
python/tvm/s_tir/backend/adreno/pipeline.py | 4 +-
.../backend/adreno/transform/__init__.py | 0
.../backend/adreno/transform/_ffi_api.py} | 5 +-
.../backend/adreno/transform/transform.py | 0
.../s_tir/meta_schedule/builder/local_builder.py | 2 +-
python/tvm/s_tir/pipeline.py | 2 -
python/tvm/s_tir/transform/transform.py | 60 +++
python/tvm/tir/analysis/analysis.py | 100 +---
python/tvm/tir/backend/adreno/__init__.py | 2 -
.../tvm/tir/backend/adreno/transform/_ffi_api.py | 21 -
python/tvm/tir/pipeline.py | 3 -
python/tvm/tir/stmt_functor.py | 19 -
python/tvm/tir/transform/transform.py | 349 +------------
src/arith/solve_linear_equation.cc | 2 +
src/contrib/msc/core/codegen/base_codegen.h | 7 +-
src/contrib/msc/core/codegen/codegen_json.cc | 2 +-
src/contrib/msc/core/codegen/codegen_json.h | 25 +-
src/contrib/msc/core/codegen/codegen_utils.h | 60 ++-
src/contrib/msc/core/codegen/cpp_codegen.h | 1 -
src/contrib/msc/core/codegen/py_codegen.h | 1 -
src/contrib/msc/core/ir/graph.cc | 48 +-
src/contrib/msc/core/ir/graph.h | 537 ++++++++++++++-------
src/contrib/msc/core/ir/graph_builder.h | 88 ++--
src/contrib/msc/core/ir/plugin.cc | 22 +-
src/contrib/msc/core/ir/plugin.h | 326 ++++++++-----
src/contrib/msc/core/printer/msc_base_printer.h | 34 +-
src/contrib/msc/core/transform/set_expr_name.cc | 2 +
.../msc/framework/tensorflow/codegen_utils.h | 8 +-
src/contrib/msc/framework/tensorflow/config.h | 13 +-
src/contrib/msc/framework/tensorrt/codegen_utils.h | 66 +--
.../msc/framework/tensorrt/transform_tensorrt.cc | 28 +-
src/contrib/msc/framework/torch/codegen_utils.h | 8 +-
src/contrib/msc/framework/torch/config.h | 13 +-
src/contrib/msc/framework/tvm/codegen_utils.h | 18 +-
src/contrib/msc/framework/tvm/config.h | 18 +-
src/contrib/msc/plugin/base_codegen.h | 7 +-
src/contrib/msc/plugin/codegen_utils.h | 81 +++-
src/contrib/msc/plugin/tensorrt_codegen.h | 13 +-
src/contrib/msc/plugin/torch_codegen.h | 18 +-
src/contrib/msc/plugin/tvm_codegen.cc | 1 -
src/contrib/msc/plugin/tvm_codegen.h | 18 +-
src/ir/instrument.cc | 17 +-
src/ir/transform.cc | 12 +-
src/node/structural_hash.cc | 11 +-
src/relax/analysis/var2value.cc | 2 +
src/relax/backend/contrib/clml/codegen.cc | 10 +-
.../backend/contrib/codegen_json/codegen_json.h | 52 +-
src/relax/backend/contrib/cublas/codegen.cc | 2 +-
src/relax/backend/contrib/cudnn/codegen.cc | 4 +-
src/relax/backend/contrib/nnapi/codegen.cc | 20 +-
src/relax/backend/contrib/tensorrt/codegen.cc | 2 +-
src/relax/backend/vm/exec_builder.cc | 1 +
.../transform/lower_global_view_to_local_view.cc | 3 +-
src/relax/ir/transform.cc | 1 -
.../transform/attach_attr_layout_free_buffers.cc | 3 +-
src/relax/transform/compute_prim_value.cc | 3 +-
src/relax/transform/fuse_tir.cc | 5 +-
src/relax/transform/split_call_tir_by_pattern.cc | 5 +-
.../transform/split_layout_rewrite_preproc.cc | 5 +-
src/runtime/const_loader_module.cc | 38 +-
src/runtime/contrib/clml/clml_runtime.cc | 50 +-
src/runtime/contrib/coreml/coreml_runtime.mm | 23 +-
src/runtime/contrib/cublas/cublas_utils.cc | 11 +-
src/runtime/contrib/cudnn/cudnn_json_runtime.cc | 1 +
src/runtime/contrib/cudnn/cudnn_utils.cc | 6 +-
src/runtime/contrib/hipblas/hipblas_utils.cc | 11 +-
src/runtime/contrib/json/json_node.h | 292 +++++------
src/runtime/contrib/json/json_runtime.h | 72 ++-
src/runtime/contrib/miopen/miopen_utils.cc | 6 +-
src/runtime/contrib/mps/mps_utils.h | 1 -
src/runtime/contrib/mps/mps_utils.mm | 7 +-
src/runtime/contrib/mrvl/mrvl_hw_runtime.cc | 35 +-
src/runtime/contrib/mrvl/mrvl_runtime.cc | 23 +-
src/runtime/contrib/msc/tensorrt_runtime.cc | 1 -
src/runtime/contrib/random/random.cc | 6 +-
src/runtime/contrib/rocblas/rocblas.cc | 12 +-
src/runtime/contrib/tensorrt/tensorrt_runtime.cc | 76 +--
src/runtime/cpu_device_api.cc | 10 +-
src/runtime/cuda/cuda_device_api.cc | 8 +-
src/runtime/cuda/cuda_module.cc | 25 +-
src/runtime/cuda/l2_cache_flush.cc | 8 +-
src/runtime/device_api.cc | 1 +
src/runtime/disco/loader.cc | 1 +
src/runtime/disco/message_queue.h | 16 +-
src/runtime/disco/protocol.h | 11 +-
src/runtime/disco/threaded_session.cc | 12 +-
src/runtime/file_utils.cc | 138 +++---
src/runtime/file_utils.h | 6 +-
src/runtime/hexagon/hexagon_device_api.cc | 10 +-
src/runtime/hexagon/hexagon_module.cc | 16 +-
src/runtime/logging.cc | 4 +-
src/runtime/meta_data.h | 27 +-
src/runtime/metal/metal_device_api.mm | 8 +-
src/runtime/metal/metal_module.mm | 42 +-
src/runtime/opencl/opencl_device_api.cc | 12 +-
src/runtime/opencl/opencl_module.cc | 48 +-
src/runtime/opencl/opencl_module_spirv.cc | 14 +-
.../opencl/opencl_wrapper/opencl_wrapper.cc | 1 -
src/runtime/pack_args.h | 1 +
src/runtime/profiling.cc | 90 ++--
src/runtime/rocm/rocm_device_api.cc | 8 +-
src/runtime/rocm/rocm_module.cc | 25 +-
src/runtime/rpc/rpc_endpoint.cc | 30 +-
src/runtime/rpc/rpc_pipe_impl.cc | 6 +-
src/runtime/rpc/rpc_socket_impl.cc | 14 +-
src/runtime/spirv/spirv_shader.h | 24 +-
src/runtime/static_library.cc | 22 +-
src/runtime/thread_pool.cc | 11 +-
src/runtime/thread_storage_scope.h | 1 +
src/runtime/vm/executable.cc | 37 +-
src/runtime/vm/tensor_cache_support.cc | 2 +-
src/runtime/vulkan/vulkan_device.cc | 4 +-
src/runtime/vulkan/vulkan_module.cc | 19 +-
src/runtime/vulkan/vulkan_wrapped_func.cc | 27 +-
src/s_tir/analysis/calculate_allocated_memory.cc | 12 +-
src/{tir => s_tir}/analysis/estimate_flops.cc | 7 +-
src/s_tir/analysis/find_anchor_sblock.cc | 3 +-
src/{tir => s_tir}/analysis/identify_memcpy.cc | 7 +-
src/{tir => s_tir}/analysis/is_pure_function.cc | 9 +-
src/{tir => s_tir}/analysis/oob_checker.cc | 15 +-
.../analysis/sblock_access_region_detector.cc | 1 +
src/{tir => s_tir}/analysis/verify_gpu_code.cc | 15 +-
.../backend/adreno/inject_texture_alloc.cc | 14 +-
.../backend/adreno/texture_flatten.cc | 12 +-
src/s_tir/meta_schedule/arg_info.cc | 4 +-
.../feature_extractor/per_store_feature.cc | 2 +-
src/s_tir/meta_schedule/mutator/mutate_parallel.cc | 1 +
.../postproc/disallow_async_strided_mem_copy.cc | 2 +-
.../meta_schedule/postproc/verify_gpu_code.cc | 6 +-
.../meta_schedule/postproc/verify_vtcm_limit.cc | 6 +-
.../search_strategy/evolutionary_search.cc | 2 +
.../meta_schedule/task_scheduler/task_scheduler.cc | 3 +-
src/s_tir/meta_schedule/utils.h | 16 +-
src/s_tir/schedule/primitive/hide_buffer_access.cc | 2 +
.../schedule/primitive/reorder_block_iter_var.cc | 2 +
.../transform/decorate_device_scope.cc | 15 +-
.../transform/default_gpu_schedule.cc | 21 +-
src/s_tir/transform/inject_software_pipeline.cc | 1 +
src/s_tir/transform/lower_async_dma.cc | 6 +-
.../transform/merge_shared_memory_allocations.cc | 2 +
src/{tir => s_tir}/transform/remove_store_undef.cc | 22 +-
.../remove_weight_layout_rewrite_block.cc | 13 +-
src/{tir => s_tir}/transform/renew_defs.cc | 11 +-
src/s_tir/transform/storage_access.cc | 1 -
.../transform/using_assume_to_reduce_branches.cc | 13 +-
.../printer/doc_printer/python_doc_printer.cc | 1 +
src/support/base64.h | 21 +-
src/support/bytes_io.h | 103 ++++
src/support/env.h | 63 +++
src/support/libinfo.cc | 1 -
src/support/pipe.h | 4 +-
src/support/socket.h | 4 +-
src/target/codegen.cc | 31 +-
src/target/intrin_rule.cc | 20 +
src/target/llvm/codegen_hexagon.cc | 2 +-
src/target/llvm/llvm_instance.cc | 3 +-
src/target/llvm/llvm_module.cc | 2 +-
src/target/source/codegen_webgpu.cc | 25 +-
src/target/source/source_module.cc | 42 +-
src/target/spirv/spirv_utils.cc | 4 +-
src/target/target.cc | 12 +-
src/target/target_info.cc | 52 --
src/tir/ir/buffer.cc | 1 +
src/tir/ir/function.cc | 4 +-
src/tir/op/builtin.cc | 4 -
src/tir/transform/bind_target.cc | 3 +-
src/tir/transform/combine_context_call.cc | 123 -----
src/tir/transform/convert_for_loops_serial.cc | 78 ---
.../transform/lower_device_storage_access_info.cc | 141 ------
src/tir/transform/lower_intrin.cc | 1 +
src/tir/transform/make_unpacked_api.cc | 210 --------
.../reduce_branching_through_overcompute.cc | 187 -------
src/tir/transform/storage_rewrite.cc | 26 +-
src/topi/einsum.cc | 3 +
tests/lint/cpplint.sh | 11 +-
tests/lint/jnilint.sh | 3 +-
tests/python/contrib/test_msc/test_plugin.py | 2 +-
.../contrib/test_msc/test_translate_tensorrt.py | 2 +-
tests/python/relax/test_codegen_cublas.py | 2 +-
tests/python/relax/test_codegen_cutlass.py | 6 +-
tests/python/relax/test_contrib_vllm.py | 2 +-
...eliminate_pad_branch_using_buffer_assumption.py | 6 +-
.../python/relax/test_frontend_nn_extern_module.py | 2 -
tests/python/relax/test_frontend_nn_op.py | 8 +-
tests/python/relax/test_transform_codegen_pass.py | 4 +-
.../relax/test_transform_meta_schedule_tuning.py | 2 +-
tests/python/relax/test_vm_cuda_graph.py | 2 +-
tests/python/relax/test_vm_multi_device.py | 2 +-
...t_s_tir_analysis_calculate_allocated_memory.py} | 6 +-
.../test_s_tir_analysis_estimate_tir_flops.py} | 2 +-
.../test_s_tir_analysis_identify_memcpy.py} | 2 +-
.../test_s_tir_analysis_is_pure_function.py} | 2 +-
.../analysis/test_s_tir_analysis_oob.py} | 10 +-
.../test_s_tir_renew_defs.py} | 12 +-
.../test_s_tir_transform_decorate_device_scope.py} | 2 +-
.../test_s_tir_transform_default_gpu_schedule.py} | 2 +-
.../test_s_tir_transform_inject_double_buffer.py | 2 +-
.../test_s_tir_transform_remove_undef.py} | 12 +-
...ransform_remove_weight_layout_rewrite_block.py} | 2 +-
.../test_tir_transform_combine_context_call.py | 113 -----
.../test_tir_transform_convert_for_loops_serial.py | 62 ---
...r_transform_lower_device_storage_access_info.py | 155 ------
.../test_tir_transform_make_unpacked_api.py | 306 ------------
...ansform_reduce_branching_through_overcompute.py | 280 -----------
.../test_tir_transform_storage_rewrite.py | 19 -
web/Makefile | 2 +-
web/emcc/tvmjs_support.cc | 3 +-
web/emcc/wasm_runtime.cc | 1 -
web/emcc/webgpu_runtime.cc | 32 +-
264 files changed, 3085 insertions(+), 4481 deletions(-)
rename include/tvm/{tir => s_tir}/backend/adreno/transform.h (73%)
create mode 100644 include/tvm/support/io.h
create mode 100644 include/tvm/support/serializer.h
delete mode 100644 include/tvm/target/target_info.h
rename python/tvm/{tir => s_tir}/backend/adreno/transform/__init__.py (100%)
copy python/tvm/{ir/_ffi_analysis_api.py =>
s_tir/backend/adreno/transform/_ffi_api.py} (87%)
rename python/tvm/{tir => s_tir}/backend/adreno/transform/transform.py (100%)
delete mode 100644 python/tvm/tir/backend/adreno/transform/_ffi_api.py
rename src/{tir => s_tir}/analysis/estimate_flops.cc (98%)
rename src/{tir => s_tir}/analysis/identify_memcpy.cc (98%)
rename src/{tir => s_tir}/analysis/is_pure_function.cc (94%)
rename src/{tir => s_tir}/analysis/oob_checker.cc (92%)
rename src/{tir => s_tir}/analysis/verify_gpu_code.cc (96%)
rename src/{tir => s_tir}/backend/adreno/inject_texture_alloc.cc (90%)
rename src/{tir => s_tir}/backend/adreno/texture_flatten.cc (94%)
rename src/{tir => s_tir}/transform/decorate_device_scope.cc (79%)
rename src/{tir => s_tir}/transform/default_gpu_schedule.cc (91%)
rename src/{tir => s_tir}/transform/remove_store_undef.cc (88%)
rename src/{tir => s_tir}/transform/remove_weight_layout_rewrite_block.cc (93%)
rename src/{tir => s_tir}/transform/renew_defs.cc (98%)
rename src/{tir => s_tir}/transform/using_assume_to_reduce_branches.cc (98%)
create mode 100644 src/support/bytes_io.h
create mode 100644 src/support/env.h
delete mode 100644 src/target/target_info.cc
delete mode 100644 src/tir/transform/combine_context_call.cc
delete mode 100644 src/tir/transform/convert_for_loops_serial.cc
delete mode 100644 src/tir/transform/lower_device_storage_access_info.cc
delete mode 100644 src/tir/transform/make_unpacked_api.cc
delete mode 100644 src/tir/transform/reduce_branching_through_overcompute.cc
rename
tests/python/{tir-analysis/test_tir_analysis_calculate_allocated_memory.py =>
s_tir/analysis/test_s_tir_analysis_calculate_allocated_memory.py} (96%)
rename tests/python/{tir-analysis/test_tir_analysis_estimate_tir_flops.py =>
s_tir/analysis/test_s_tir_analysis_estimate_tir_flops.py} (98%)
rename tests/python/{tir-analysis/test_tir_analysis_identify_memcpy.py =>
s_tir/analysis/test_s_tir_analysis_identify_memcpy.py} (99%)
rename tests/python/{tir-analysis/test_tir_analysis_is_pure_function.py =>
s_tir/analysis/test_s_tir_analysis_is_pure_function.py} (97%)
rename tests/python/{tir-analysis/test_tir_analysis_oob.py =>
s_tir/analysis/test_s_tir_analysis_oob.py} (84%)
rename tests/python/{tir-base/test_tir_renew_defs.py =>
s_tir/test_s_tir_renew_defs.py} (96%)
rename tests/python/{tir-transform/test_tir_transform_decorate_device_scope.py
=> s_tir/transform/test_s_tir_transform_decorate_device_scope.py} (93%)
rename tests/python/{tir-transform/test_transform_default_gpu_schedule.py =>
s_tir/transform/test_s_tir_transform_default_gpu_schedule.py} (99%)
rename tests/python/{tir-transform/test_tir_transform_remove_undef.py =>
s_tir/transform/test_s_tir_transform_remove_undef.py} (90%)
rename
tests/python/{tir-transform/test_tir_transform_remove_weight_layout_rewrite_block.py
=> s_tir/transform/test_s_tir_transform_remove_weight_layout_rewrite_block.py}
(97%)
delete mode 100644
tests/python/tir-transform/test_tir_transform_combine_context_call.py
delete mode 100644
tests/python/tir-transform/test_tir_transform_convert_for_loops_serial.py
delete mode 100644
tests/python/tir-transform/test_tir_transform_lower_device_storage_access_info.py
delete mode 100644
tests/python/tir-transform/test_tir_transform_make_unpacked_api.py
delete mode 100644
tests/python/tir-transform/test_tir_transform_reduce_branching_through_overcompute.py
