This is an automated email from the ASF dual-hosted git repository.
mbrookhart pushed a change to branch ci-docker-staging
in repository https://gitbox.apache.org/repos/asf/tvm.git.
discard c6064b0 support explicit padding for NCHW TF padding test
discard 38b3165 manage TF memory use in TF1 tests
discard 7f5a959 next try at docker images
discard 275cc43 skip a test until update complete
discard 01f058a try updating docker images again
discard ca7497e disable test until CI update complete
discard 1d06784 Don't force output shape for conv transpose tests, add 1D and
3D cases
discard 77e6077 support convtranspose opset 11 autopadding
discard b7a65f3 point jenkins at new docker
discard fa513d5 add failing onnx tets
add 15bdf28 Fix address and port reported by android_rpc to tracker
(#8405)
add c81d533 [Bugfix] Fix broadcast type func with incomplete type (#8438)
add cd5a20e [COMMUNITY] @junrushao1994 -> PMC (#8450)
add 62adc77 [MyPy] Extend type checking and annotation for TIR (#8429)
add d043cb9 [BugFix][TOPI] Fix the integer overflow problem of the
scatter_nd op. (#8415)
add 807373c Add qnn batch_matmul operator (#8401)
add 80f48c7 [microTVM] Fix Stack Size Issue for Zephyr AOT Demo on
Physical Hardware (#8453)
add 957cc12 [Relay] Modify create_executor to pass params (#8418)
add d67514b [PROFILING] Use PAPI to collect hardware performance counters
on CPU and CUDA (#7983)
add 136f218 [Relay][ONNX] Batch_matmul to dense optimization (#8440)
add f62917e [TOPI] Add support for arbitrary dtypes to CSRMV and CSRMM
(#8437)
add 1a26733 [Refactor] Enforce attaching storage scope to PointerType
(#8366)
add 73b38e8 [Fix] Explicitly retain `__hash__` of `StringImm` (#8449)
add 5c1a1cf [CUDA] Improve injective schedule to enable half2 (#8457)
add c16d61b [Fix] Remove unused variable in GraphExecutorCodegen (#8465)
add a425d265 [Docs] Corrected typo in googletest build instructions.
(#8459)
add f15be8b [RPC] Fix cpp_rpc connection to rpc_tracker (#8388)
add e1b3ff4 [Relay][Frontend][ONNX] Add ConvInteger support. (#8456)
add b3e8d61 add failing onnx tets
add 3aa5c64 point jenkins at new docker
add 8f2c1aa support convtranspose opset 11 autopadding
add 6352a12 Don't force output shape for conv transpose tests, add 1D and
3D cases
add a1e278e disable test until CI update complete
add c0b6bf4 try updating docker images again
add 68cef66 skip a test until update complete
add 2bf210d next try at docker images
add 7ae767d manage TF memory use in TF1 tests
add 56da2ab support explicit padding for NCHW TF padding test
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 (c6064b0)
\
N -- N -- N refs/heads/ci-docker-staging (56da2ab)
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 | 2 +
CONTRIBUTORS.md | 2 +-
apps/cpp_rpc/rpc_env.cc | 9 +-
apps/cpp_rpc/rpc_server.cc | 2 +-
apps/cpp_rpc/rpc_tracker_client.h | 15 +-
.../modules/contrib/PAPI.cmake | 17 +-
docs/dev/inferbound.rst | 2 -
docs/index.rst | 1 +
docs/install/from_source.rst | 2 +-
.../{api/python/driver.rst => profiling/index.rst} | 10 +-
docs/profiling/papi.rst | 114 ++++++++
.../tvm/runtime/contrib/papi.h | 39 ++-
include/tvm/runtime/profiling.h | 102 ++++++-
include/tvm/runtime/threading_backend.h | 8 +
include/tvm/te/operation.h | 15 +-
include/tvm/tir/buffer.h | 3 +-
include/tvm/tir/stmt.h | 9 +-
.../tvm/rpc/ConnectTrackerServerProcessor.java | 4 +-
python/tvm/contrib/debugger/debug_executor.py | 9 +-
python/tvm/relay/build_module.py | 9 +-
python/tvm/relay/frontend/onnx.py | 106 +++++++-
python/tvm/relay/op/strategy/x86.py | 6 +-
python/tvm/relay/qnn/op/qnn.py | 38 +++
python/tvm/runtime/profiler_vm.py | 8 +-
python/tvm/runtime/profiling.py | 48 ----
python/tvm/runtime/profiling/__init__.py | 85 ++++++
.../tvm/{arith => runtime/profiling}/_ffi_api.py | 7 +-
python/tvm/script/scope_handler.py | 2 +-
python/tvm/script/special_stmt.py | 16 ++
python/tvm/te/hybrid/parser.py | 3 +-
python/tvm/tir/analysis/analysis.py | 34 ++-
python/tvm/tir/buffer.py | 12 +-
python/tvm/tir/data_layout.py | 28 +-
python/tvm/tir/expr.py | 158 ++++++-----
python/tvm/tir/function.py | 4 +-
python/tvm/tir/generic.py | 12 +-
python/tvm/tir/ir_builder.py | 13 +-
python/tvm/tir/op.py | 70 ++---
python/tvm/tir/stmt.py | 67 +++--
python/tvm/tir/stmt_functor.py | 6 +-
python/tvm/tir/transform/function_pass.py | 17 +-
python/tvm/tir/transform/transform.py | 113 ++++----
python/tvm/topi/cuda/injective.py | 36 ++-
python/tvm/topi/cuda/scatter.py | 6 +-
python/tvm/topi/nn/batch_matmul.py | 28 +-
python/tvm/topi/nn/sparse.py | 12 +-
python/tvm/topi/sparse/csrmm.py | 15 +-
python/tvm/topi/sparse/csrmv.py | 15 +-
python/tvm/topi/x86/batch_matmul.py | 25 +-
src/contrib/hybrid/codegen_hybrid.cc | 9 +-
src/contrib/hybrid/codegen_hybrid.h | 2 -
src/printer/tvmscript_printer.cc | 14 +-
src/relay/backend/aot_executor_codegen.cc | 3 +-
src/relay/backend/graph_executor_codegen.cc | 4 +-
src/relay/op/nn/nn.cc | 53 +---
src/relay/op/nn/nn.h | 54 ++++
src/relay/op/tensor/transform.cc | 10 +-
src/relay/qnn/op/batch_matmul.cc | 216 +++++++++++++++
src/runtime/contrib/papi/papi.cc | 299 +++++++++++++++++++++
.../graph_executor/debug/graph_executor_debug.cc | 15 +-
src/runtime/profiling.cc | 98 ++++---
src/runtime/thread_pool.cc | 38 ++-
src/runtime/thread_storage_scope.h | 4 +-
src/runtime/vm/executable.cc | 13 +
src/runtime/vm/profiler/vm.cc | 53 ++--
src/runtime/vm/profiler/vm.h | 5 +-
src/target/llvm/codegen_amdgpu.cc | 5 +-
src/target/llvm/codegen_llvm.cc | 8 +-
src/target/llvm/codegen_llvm.h | 2 -
src/target/llvm/codegen_nvptx.cc | 6 +-
src/target/source/codegen_c.h | 1 +
src/target/source/codegen_cuda.cc | 8 +-
src/target/spirv/codegen_spirv.cc | 14 +-
src/target/spirv/codegen_spirv.h | 2 -
src/te/operation/compute_op.cc | 4 +-
src/te/operation/create_primfunc.cc | 5 +-
src/te/operation/cross_thread_reduction.cc | 5 +-
src/te/operation/extern_op.cc | 4 +-
src/te/operation/hybrid_op.cc | 4 +-
src/te/operation/placeholder_op.cc | 2 +-
src/te/operation/scan_op.cc | 4 +-
src/te/schedule/schedule_ops.cc | 11 +-
src/te/schedule/schedule_postproc_to_primfunc.cc | 19 +-
src/tir/ir/buffer.cc | 5 +-
src/tir/ir/stmt.cc | 17 +-
src/tir/transforms/ir_utils.cc | 6 +
src/tir/transforms/ir_utils.h | 6 +
src/tir/transforms/lower_thread_allreduce.cc | 40 ++-
src/tir/transforms/lower_warp_memory.cc | 12 +-
src/tir/transforms/storage_access.cc | 23 +-
src/tir/transforms/storage_access.h | 4 +-
src/tir/transforms/storage_flatten.cc | 20 +-
src/tir/transforms/storage_rewrite.cc | 34 +--
src/tir/transforms/thread_storage_sync.cc | 27 +-
src/tir/transforms/update_pointer_storage_scope.cc | 91 +++++++
.../transforms/update_pointer_storage_scope.h} | 48 ++--
tests/micro/zephyr/test_zephyr_aot.py | 4 +-
tests/python/frontend/onnx/test_forward.py | 165 +++++++++++-
tests/python/relay/test_op_qnn_batch_matmul.py | 247 +++++++++++++++++
tests/python/topi/python/test_topi_sparse.py | 35 ++-
tests/python/unittest/test_runtime_profiling.py | 66 ++++-
tests/python/unittest/test_te_hybrid_script.py | 4 +-
.../python/unittest/test_te_schedule_tensorize.py | 4 +-
tests/python/unittest/test_te_tensor.py | 2 +-
.../unittest/test_tir_transform_flatten_buffer.py | 8 +-
.../python/unittest/test_tir_transform_hoist_if.py | 2 +-
.../unittest/test_tir_transform_loop_partition.py | 4 +-
.../test_tir_transform_lower_warp_memory.py | 4 +-
tests/python/unittest/test_tvmscript_roundtrip.py | 4 +-
tests/scripts/task_mypy.sh | 6 +
110 files changed, 2483 insertions(+), 763 deletions(-)
copy apps/microtvm/zephyr/aot_demo/boards/mps2_an521.conf =>
cmake/modules/contrib/PAPI.cmake (72%)
copy docs/{api/python/driver.rst => profiling/index.rst} (88%)
create mode 100644 docs/profiling/papi.rst
copy tests/cpp/profiling_test.cc => include/tvm/runtime/contrib/papi.h (54%)
delete mode 100644 python/tvm/runtime/profiling.py
create mode 100644 python/tvm/runtime/profiling/__init__.py
copy python/tvm/{arith => runtime/profiling}/_ffi_api.py (89%)
create mode 100644 src/relay/qnn/op/batch_matmul.cc
create mode 100644 src/runtime/contrib/papi/papi.cc
create mode 100644 src/tir/transforms/update_pointer_storage_scope.cc
copy src/{te/schedule/verify_compact_buffer.cc =>
tir/transforms/update_pointer_storage_scope.h} (54%)
create mode 100644 tests/python/relay/test_op_qnn_batch_matmul.py
