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 14e41c681a [Relax][ONNX] Support dynamic repeats for Tile (#18878)
add 689d2b51b2 [REFACTOR][TIR] Remove body from AllocBuffer and DeclBuffer
(#18876)
add 17d1a28900 [FIX][Adreno] Replace AllocBuffer with Bind in texture
alloc injection (#18881)
add c0a305dfeb [TARGET] Fix round-trip reconstruction of targets with
canonicalizer-generated `feature.*` attrs (#18883)
No new revisions were added by this update.
Summary of changes:
include/tvm/script/ir_builder/tir/frame.h | 107 ----
include/tvm/script/ir_builder/tir/ir.h | 32 +-
include/tvm/tir/stmt.h | 19 +-
python/tvm/script/ir_builder/tir/frame.py | 23 +-
python/tvm/script/ir_builder/tir/ir.py | 59 +--
python/tvm/script/ir_builder/tir/utils.py | 42 +-
python/tvm/tir/stmt.py | 22 +-
python/tvm/topi/gpu/scan.py | 18 +-
python/tvm/topi/gpu/sort.py | 440 +++++++---------
python/tvm/topi/searchsorted.py | 46 +-
python/tvm/topi/sparse_reshape.py | 217 +++-----
python/tvm/topi/vision/nms.py | 47 +-
python/tvm/topi/vision/nms_util.py | 32 +-
src/relax/op/tensor/inspect.cc | 5 +-
src/s_tir/analysis/calculate_allocated_memory.cc | 18 +-
src/s_tir/analysis/estimate_flops.cc | 4 +-
src/s_tir/backend/adreno/inject_texture_alloc.cc | 5 +-
src/s_tir/schedule/primitive/cache_read_write.cc | 41 +-
src/s_tir/schedule/transform.cc | 22 +-
src/s_tir/transform/compact_buffer_region.cc | 43 +-
src/s_tir/transform/inject_double_buffer.cc | 33 +-
src/s_tir/transform/inject_ptx_ldg32.cc | 6 +-
src/s_tir/transform/inject_virtual_thread.cc | 8 +-
src/s_tir/transform/lower_match_buffer.cc | 28 +-
src/s_tir/transform/lower_opaque_block.cc | 2 +-
src/s_tir/transform/lower_thread_allreduce.cc | 136 ++++-
src/s_tir/transform/lower_vtcm_alloc.cc | 6 +-
.../transform/merge_shared_memory_allocations.cc | 15 +-
src/s_tir/transform/profile_instrumentation.cc | 3 +-
src/s_tir/transform/thread_storage_sync.cc | 15 +-
src/script/ir_builder/tir/frame.cc | 21 -
src/script/ir_builder/tir/ir.cc | 52 +-
src/script/printer/tir/stmt.cc | 19 +-
src/target/llvm/codegen_amdgpu.cc | 4 +-
src/target/llvm/codegen_llvm.cc | 18 +-
src/target/llvm/codegen_nvptx.cc | 4 +-
src/target/source/codegen_c.cc | 12 +-
src/target/source/codegen_c.h | 2 +
src/target/source/codegen_cuda.cc | 4 +-
src/target/source/codegen_metal.cc | 4 +-
src/target/source/codegen_webgpu.cc | 1 -
src/target/spirv/codegen_spirv.cc | 12 +-
src/target/target.cc | 35 +-
src/tir/ir/script/script_complete.cc | 18 +-
src/tir/ir/stmt.cc | 18 +-
src/tir/ir/stmt_functor.cc | 10 +-
src/tir/ir/tir_visitor_with_path.cc | 14 +-
src/tir/ir/tir_visitor_with_path.h | 4 +-
src/tir/transform/common_subexpr_elim.cc | 27 +-
src/tir/transform/flatten_buffer.cc | 2 +-
src/tir/transform/ir_utils.cc | 56 +-
src/tir/transform/lower_tvm_builtin.cc | 53 +-
src/tir/transform/lower_warp_memory.cc | 50 +-
src/tir/transform/remove_no_op.cc | 20 +-
src/tir/transform/split_host_device.cc | 2 +-
src/tir/transform/storage_rewrite.cc | 59 ++-
src/tir/transform/tvm_ffi_binder.cc | 2 +-
src/tir/transform/unsupported_dtype_legalize.cc | 11 +-
tests/cpp/ir_functor_test.cc | 79 ++-
tests/cpp/target_test.cc | 57 ++
tests/python/codegen/test_target_codegen_llvm.py | 14 +-
tests/python/codegen/test_target_codegen_vulkan.py | 4 +-
tests/python/s_tir/test_s_tir_renew_defs.py | 10 +-
.../test_s_tir_transform_compact_buffer_region.py | 10 +-
.../test_s_tir_transform_inject_double_buffer.py | 10 +-
.../test_s_tir_transform_inject_virtual_thread.py | 26 +-
...test_s_tir_transform_lower_thread_all_reduce.py | 578 +++------------------
...form_merge_dynamic_shared_memory_allocations.py | 165 ++----
.../test_s_tir_transform_rewrite_unsafe_select.py | 15 +-
.../transform/test_s_tir_transform_thread_sync.py | 50 +-
.../test_tir_analysis_undefined_vars.py | 15 +-
.../test_tir_analysis_verify_well_formed.py | 35 +-
tests/python/tir-base/test_tir_constructor.py | 3 +-
tests/python/tir-base/test_tir_specialize.py | 3 +-
.../tir-base/test_tir_stmt_functor_substitute.py | 20 +-
.../test_tir_transform_convert_ssa.py | 6 +-
.../test_tir_transform_flatten_buffer.py | 8 +-
.../test_tir_transform_lower_tvm_builtin.py | 55 +-
.../test_tir_transform_narrow_datatype.py | 7 +-
...est_tir_transform_pointer_value_type_rewrite.py | 48 +-
.../test_tir_transform_remove_no_op.py | 6 +-
.../tir-transform/test_tir_transform_simplify.py | 7 +-
.../test_tir_transform_storage_rewrite.py | 108 ++--
.../test_tir_transform_unroll_loop.py | 18 +-
.../tvmscript/test_tvmscript_error_report.py | 10 +-
.../tvmscript/test_tvmscript_ir_builder_tir.py | 39 +-
.../test_tvmscript_printer_structural_equal.py | 6 +-
.../python/tvmscript/test_tvmscript_printer_tir.py | 80 +--
tests/python/tvmscript/test_tvmscript_roundtrip.py | 90 ++--
89 files changed, 1413 insertions(+), 2197 deletions(-)
