This is an automated email from the ASF dual-hosted git repository.
leandron pushed a change to branch ci-docker-staging
in repository https://gitbox.apache.org/repos/asf/tvm.git.
discard 2b46855 [Hexagon] Enable running CI tests via simulator
add ede5601 [ci] Use available CPUs in builds (#10359)
add 64bf0ef [ci] Fix slow test script permissions (#10457)
add 83f8e54 [runtime][Hexagon] AOTExecutor implementation for C Codegen
(#10311)
add b7922f3 [microTVM] Zephyr: add B-U585I-IOT02A board support (#10416)
add 1271e6c [MetaSchedule] Fix Cyclic Dependency in PyClass Family
(#10368)
add 33b9b5c [Hotfix] Black format (#10482)
add 9679e68 [MetaSchedule] Keep Task / Trial / Iter / Postproc Number
Consistent in Log (#10478)
add 646d8cd [Torch] fix torch version check (#10481)
add bf8e7e0 [microNPU] Remove unused code from testing infra (#10462)
add 865f34e [MetaSchedule] Enable AutoTVM-style template-based search
space (#10461)
add 6ad7638 [MetaSchedule] update misc parts (#10444)
add 0f2392c [Arith] Handle mod/floormod in modular set analysis (#10453)
add d1046f5 Correctly enable architecture extensions in CMSIS-NN Zephyr
Demo (#10458)
add 4ae142f [Minor][Build] Fix compiler warnings hidden overloaded
virtual function (#10485)
add 86ba517 [Refactor] Expose meta-schedule related packed func in a
header and call it directly (#10470)
add 8fe75fd [ci] Skip CI based on globs (#10456)
add 3f96f3d [skip ci][ci] hotfix Jenkinsfile (#10492)
add db8cf2f don't rely on cudnn for compilation (#10495)
add fd92947 [ci] Disable flaky tuning test (#10490)
add 837d5d4 [ci] Disable flaky ethosu tests (#10488)
add 70f223b [CodeGen][CUDA] use hrint for cuda half rounding (#10460)
add 9415bea [ETHOSN] Implement tanh operator (#10486)
add 0c0245a [Hexagon] Fix scripts to enable automated testing on hardware
(#10491)
add ff3a48e [TIR] Fix Tensorization IR-Comparator for Annotations (#10498)
add 7cfaa88 [Hexagon] Enable running CI tests via simulator (#10473)
add 085d36c [MetaSchedule] Refactor testing workloads (#10497)
add 96069c1 [CI] Remove `llvm -device=arm_cpu` and `cuda -libs=cudnn`
from the default test target list (#10500)
add 8729f6b [MetaSchedule] Update scripts for subgraph tuning (#10501)
add d7ccb91 [AOT] Rerun FVP test incase of first attempt failure (#10408)
add 7dd6335 [CMSIS-NN] Only run ScalarToTensorConstants pass on CMSISNN
external functions (#10375)
add 1970223 Fixed additional deprecation warning in file (#10318)
add 174d09e [TE][TIR] Implement layout transformations, non-flat memory
buffers (#9727)
add 12f213a [BYOC-DNNL] Support DNNL optimal layout (#10421)
add 22abfc4 support returned function in relay.build (#10502)
add 57556fe [Hexagon] Add default constructor to struct Optional in
session.cc (#10517)
add 71d6f46 fix: select narrow dtype (#10519)
add 5c0ea30 [vulkan] Add integer dot product (4xint8, 4xuint8)
tensorization for the vulkan SPIR-V target. (#10391)
add 6ac684b Automatically close open files to prevent ResourceWarning
(#10526)
add 704ac8b [FIX] Only allow autoscheduler layout rewritting in
conv2d_nhwc (#10522)
add f4b74ba Add Check about negative uint constant (#10484)
add acf8be7 [microNPU] modify the demo to use USMP (#10511)
add 5e81389 [microNPU] Introduce a pass to remove redundant identity
operations (#10254)
add 5804d20 [microNPU] Fix stride bug in strided slice legalization
(#10286)
add 7688db7 [PTX] Support mma.sp to use Sparse Tensor Cores and refactor
mma codegen (#10339)
add 35d2e8b [TE COMPILER] Propagate structural hash from relay function
to TIR function (#10475)
add a8a2574 [skip ci][Bugfix] Allow constant folding of 0U - 0U (#10535)
add ffd5f70 Revert "[TE][Fix] Comparison of the output tensor (#9829)"
(#10540)
add fe33ed6 RFC: initial stab at TorchScript fallback (#7401)
add f9d3918 [LLVM][TIR] Propagate variable names to parameters. (#10514)
add 060d9d2 [AOT] Introduce checks for return values from operators
(#10424)
add 0fa3540 [QNN] Add nn.adaptive_avg_pool1d to FQ2I (#10541)
add 1f60529 [Hexagon] Resolve breakage in
test_hexagon/test_cache_read_write (#10520)
add 48793f3 Add ONNX LinearRegressor operator support (#10477)
add 39bf923 [AOT] Use python temporary directory for AOT tests (#10518)
add a1fb44d [microNPU] Add support for TFLite FULLY_CONNECTED (#10345)
add 15dc4aa [Runtime][ThreadPool] Remove a cout log output. (#10560)
add 201a584 [MetaSchedule] Bug Fix for Relay Integration (#10534)
add c02a8b7 [FIX,HEXAGON] Gitignore generated Hexagon files (#10552)
add ac66072 Interpreter call in FoldConstant now always uses graph
executor with link-params=0 (#10465)
add de945e0 [CI] Enable TOPI tests in ci_arm (#10564)
add a57397e [Bugfix] Simultaneous layout transform and axis separators.
(#10553)
add 0825e93 [TVMScript] Add intrinsic to look up llvm intrinsic id
(#10551)
add 45ef533 [PyTorch][BugFix] PyTorch-TVM Bridge Build Scripts (#10527)
add 3894991 [ci] Remove commit check on ci skipping logic (#10537)
add e2211a2 [CI] Upgrade Python dependencies as part of Docker image build
add 7e49f53 [AUTO_SCHEDULER] Add feature extraction directly from
PrimFunc (#10455)
add 5b76768 Unit test for DFPatternRewriter on deeply nested sub-graph
with attributes on call. (#10533)
add 3a9e77b Fix TorchScript fallback build (#10556)
add 0b37bd2 Remove CODEOWNERS (#10192)
add 51ae845 [Minor][MetaSchedule] Remove Unused Imports (#10577)
add 076fa33 [TECompiler] Decouple TE compute and schedule lowering in
ScheduleBuilder (#10561)
add 05cda49 [CMSIS-NN] Include clip in the qnn binary op patterns (#10548)
add 4e4f607 [BYOC][TENSOORT] Add support for FP16 on TensorRT BYOC flow
(#10388)
add 3f9cdee [TVMSCRIPT] Add type definition for preflattened_buffer
(#10550)
add a4a481f [Refactor] Reduced repetition in CodeGenLLVM's buffer access
(#10567)
add e34985b [Hexagon] Add doc on TVM - Hexagon RPC flow (#10507)
add 6f3158b [CMAKE] Add option to enable custom logging (#10531)
add 678e76b [TIR] Restrict Buffer indices, only last index can be
multi-lane (#10513)
add 39487d8 [ci] Build GPU libraries on CPU nodes (#10539)
add 409ddef [ci] Delay pytest errors until all invocations have run
(#10521)
add 5dc4015 [CMAKE,HEXAGON] Only enable Hexagon custom logging when
building for Hexagon (#10587)
add 4cdbf5c [TE] Promote substituted variable to iter_var's dtype (#10571)
add 975086e [Arith] Support dtype promotion in TIR comparison expr
creation (#10584)
add aa47018 [QNN] unary op for quantized resize2d and test (#10589)
add ce2f81a Upgrade Windows build to use windows-2019 runner (#10585)
add 5775f64 [Fix] Refactor the roundtrip test. (#10592)
add 3187753 [Minor] fix redundant compute (#10580)
add 5eb93df [CMSIS-NN] Scalar to tensor constant pass to support only
qnn.add and qnn.multiply (#10563)
add 8bddaab [TFLite] Quantized unary elemwise ops (#10566)
add 4d88a45 [microNPU] Improve cycles estimates for memory transfers
(#10508)
add 7d5ef84 [CUDA] Various int8 fix (cublas, cutlass, etc) (#10596)
add 8418026 [FQ2I] Add leaky relu to FQ21 (#10378)
add 47cd410 Deploy docs to tvm-site/asf-site on main (#10494)
add c3168d1 [microTVM][RVM] Improve base-box-tool 'build' command (#8738)
add d7af2e3 [TIR] Updated python docstring and parameter names for
AllocateConst (#10602)
add f9f9f1d [Runtime][PipelineExecutor] Add the pipeline internal
forwarding logic. (#10543)
add 50c632e [BYOC][TENSORRT] Fix bug of Segmentation Fault when loading
engine file. (#10597)
add 8bbb206 [TVMScript] fix print target's host (#10598)
add 2b7013e [Arith] Improve floordiv / floormod rewrite simplifing rules
(#10591)
add ff54011 [Bugfix][MetaSchedule] Fix over-simplification of Select
(#10605)
add 2f7bb58 [Hexagon] Generalize builtin for Nd memory alloc with storage
scope and add lowering for VTCM / Hexagon (#10558)
add 0b61256 Fix bug check trt (#10600)
add 0ea07a7 [CI] Pin numpy version in image build (#10611)
add 4797850 [microTVM] Zephyr: add mps3_an547 board support (#10479)
add 8a636a9 [Hexagon] Codegen for 2d Load/Store (#10586)
add 232a84e [FQ2I] Add mean op to FQ2I (#10607)
add 1584da4 [Hexagon] Deprecate SDK 3.x, rewrite HexagonSDK.cmake (#10612)
add 4806b80 [Hexagon] Refactor tvm.contrib.hexagon, NFC (#10616)
add 0d068f4 Add remaining targets to ci.py (#10425)
add ab4289d [skip ci][ci] Add missing guard to skip CI check (#10625)
add ce335c3 [Metaschedule] New relay backend for meta schedule task
extraction (#10578)
add 1cd9814 [Hexagon] Remove double ".hexagon.hexagon." from registered
names, NFC (#10624)
add 6991396 [Arith] Fix floormod rewrite simplify rule (#10626)
add ab074d2 [CI] Remove mps3_an547 from the CI (#10621)
add 3e30a5f [Bugfix] Handled TransformNode in
PassUpBitMaskOr/PassDownBitMaskOr (#10620)
add 794e1e3 [Testing] Add model loader for int8 BERT (#10622)
add 690581c [Pytorch] Add `aten::fmod` and `aten::remainder` (#10613)
add 69bb71c [Hexagon] Add support for Hexagon v69, deprecate v60 and v62
(#10623)
add 6d5d3e2 [Docker] Move psutil installation to
ubuntu_install_python_package (#10615)
add 1fd1b79 [ci] Move pip dependencies to docker images, add ninja /
shellcheck (#10257)
add 08038fa [microNPU] re-enable network tests (#10565)
new c573163 Upgrade tensorflow to version to 2.6.x
new e033731 Fix VitisAI.cmake to use Python 3.7.
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 (2b46855)
\
N -- N -- N refs/heads/ci-docker-staging (e033731)
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.
The 2 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "add" were already present in the repository and have only
been added to this reference.
Summary of changes:
.github/{CODEOWNERS => CODEOWNERSHIP} | 11 +
.github/workflows/main.yml | 3 +-
.gitignore | 8 +
CMakeLists.txt | 20 +-
Jenkinsfile | 126 +-
apps/cpp_rpc/CMakeLists.txt | 10 +-
apps/hexagon_launcher/README.md | 4 +-
apps/hexagon_launcher/cmake/HexagonLauncher.cmake | 24 +-
apps/hexagon_launcher/cmake/android/CMakeLists.txt | 26 +-
apps/hexagon_launcher/cmake/hexagon/CMakeLists.txt | 24 +-
apps/microtvm/ethosu/corstone300.ld | 16 +-
apps/microtvm/ethosu/include/tvm_runtime.h | 9 +-
apps/microtvm/ethosu/run_demo.sh | 3 +
apps/microtvm/ethosu/src/demo_bare_metal.c | 3 -
apps/microtvm/ethosu/src/demo_freertos.c | 2 -
apps/microtvm/ethosu/src/tvm_ethosu_runtime.c | 8 +-
apps/microtvm/reference-vm/base-box-tool.py | 25 +-
apps/microtvm/zephyr/template_project/boards.json | 17 +
.../zephyr/template_project/microtvm_api_server.py | 13 +-
.../template_project/qemu-hack/qemu-system-i386 | 7 +-
apps/microtvm/zephyr_cmsisnn/CMakeLists.txt | 6 +-
apps/microtvm/zephyr_cmsisnn/prj.conf | 4 +-
apps/microtvm/zephyr_cmsisnn/src/main.c | 3 +-
cmake/config.cmake | 18 +-
cmake/libs/hexagon_rpc_skel/CMakeLists.txt | 120 -
cmake/modules/Hexagon.cmake | 45 +-
cmake/modules/HexagonSDK.cmake | 199 +-
cmake/modules/Logging.cmake | 9 +
cmake/modules/Vulkan.cmake | 6 +-
.../modules/contrib/LibTorch.cmake | 21 +-
cmake/modules/contrib/PT_TVMDSOOP.cmake | 9 +-
cmake/modules/contrib/VitisAI.cmake | 4 +-
cmake/utils/FindVulkan.cmake | 26 +-
conda/recipe/bld.bat | 1 +
docker/Dockerfile.ci_lint | 2 +-
docker/bash.sh | 12 +-
docker/dev_common.sh | 2 +-
docker/install/ubuntu_install_core.sh | 2 +-
docker/install/ubuntu_install_python_package.sh | 7 +-
docker/install/ubuntu_install_redis.sh | 1 -
docker/install/ubuntu_install_tensorflow.sh | 6 +-
.../install/ubuntu_install_vitis_ai_packages_ci.sh | 2 +-
docs/contribute/ci.rst | 5 +-
.../how_to/deploy_models/deploy_prequantized.py | 3 +-
include/tvm/auto_scheduler/feature.h | 10 +-
include/tvm/ir/attrs.h | 41 +
include/tvm/meta_schedule/integration.h | 62 +-
include/tvm/meta_schedule/runner.h | 10 +-
include/tvm/meta_schedule/space_generator.h | 2 +-
.../libtorch_runtime.h} | 21 +-
include/tvm/te/operation.h | 9 +-
include/tvm/te/schedule.h | 123 +-
include/tvm/tir/buffer.h | 44 +-
include/tvm/tir/builtin.h | 42 +-
include/tvm/tir/expr.h | 16 +
include/tvm/tir/function.h | 44 +-
include/tvm/tir/index_map.h | 140 +
include/tvm/tir/stmt.h | 17 +
include/tvm/tir/transform.h | 7 +
include/tvm/tir/var.h | 4 +
include/tvm/topi/broadcast.h | 38 +
include/tvm/topi/transform.h | 8 +-
python/tvm/auto_scheduler/feature.py | 78 +-
python/tvm/auto_scheduler/relay_integration.py | 12 +-
python/tvm/auto_scheduler/workload_registry.py | 10 +-
python/tvm/autotvm/record.py | 13 +-
python/tvm/contrib/cudnn.py | 28 +-
python/tvm/contrib/cutlass/gen_conv2d.py | 5 +
python/tvm/contrib/cutlass/gen_gemm.py | 9 +
python/tvm/contrib/ethosu/cascader/block_config.py | 14 +-
.../tvm/contrib/ethosu/cascader/device_config.py | 7 +-
python/tvm/contrib/ethosu/cascader/graph.py | 4 +
.../tvm/contrib/ethosu/cascader/tensor_config.py | 20 +-
python/tvm/contrib/hexagon/__init__.py | 4 +-
python/tvm/contrib/hexagon/build.py | 37 +-
python/tvm/contrib/hexagon/session.py | 1 +
python/tvm/contrib/hexagon/tools.py | 177 +
.../contrib/hexagon/{hexagon.py => transform.py} | 104 +-
python/tvm/contrib/nvcc.py | 9 +-
python/tvm/meta_schedule/builder/builder.py | 43 +-
python/tvm/meta_schedule/builder/local_builder.py | 66 +-
python/tvm/meta_schedule/cost_model/cost_model.py | 124 +-
.../tvm/meta_schedule/cost_model/random_model.py | 19 +-
python/tvm/meta_schedule/cost_model/xgb_model.py | 3 +-
python/tvm/meta_schedule/database/database.py | 127 +-
.../feature_extractor/feature_extractor.py | 62 +-
python/tvm/meta_schedule/integration.py | 103 +-
.../measure_callback/measure_callback.py | 75 +-
python/tvm/meta_schedule/mutator/mutator.py | 81 +-
python/tvm/meta_schedule/postproc/postproc.py | 81 +-
python/tvm/meta_schedule/runner/__init__.py | 2 +-
python/tvm/meta_schedule/runner/config.py | 2 +-
python/tvm/meta_schedule/runner/local_runner.py | 153 +-
python/tvm/meta_schedule/runner/rpc_runner.py | 223 +-
python/tvm/meta_schedule/runner/runner.py | 90 +-
.../meta_schedule/schedule_rule/schedule_rule.py | 87 +-
.../search_strategy/search_strategy.py | 118 +-
.../meta_schedule/space_generator/schedule_fn.py | 2 +
.../space_generator/space_generator.py | 67 +-
.../meta_schedule/task_scheduler/task_scheduler.py | 203 +-
python/tvm/meta_schedule/testing/__init__.py | 3 -
python/tvm/meta_schedule/testing/byoc_trt.py | 53 -
.../meta_schedule/testing/conv2d_winograd_cpu.py | 172 +
.../meta_schedule/testing/conv2d_winograd_cuda.py | 173 +
.../meta_schedule/testing/custom_builder_runner.py | 140 +
python/tvm/meta_schedule/testing/relay_workload.py | 435 +-
.../testing/run_subgraph_auto_scheduler.py | 137 +
.../testing/run_subgraph_meta_schedule.py | 120 +
python/tvm/meta_schedule/testing/tlcbench.py | 124 +
python/tvm/meta_schedule/tune.py | 6 +-
python/tvm/meta_schedule/utils.py | 144 +-
python/tvm/relay/backend/contrib/ethosu/codegen.py | 14 +
.../tvm/relay/backend/contrib/ethosu/legalize.py | 96 +-
.../contrib/ethosu/tir/binary_elementwise.py | 6 +-
.../backend/contrib/ethosu/tir/convolution.py | 12 +-
.../relay/backend/contrib/ethosu/tir/depthwise.py | 12 +-
python/tvm/relay/backend/contrib/ethosu/tir/dma.py | 58 +-
.../relay/backend/contrib/ethosu/tir/identity.py | 18 +-
.../tvm/relay/backend/contrib/ethosu/tir/passes.py | 425 +-
.../relay/backend/contrib/ethosu/tir/pooling.py | 4 +-
.../tvm/relay/backend/contrib/ethosu/tir/spec.py | 15 +-
.../relay/backend/contrib/ethosu/tir/transform.py | 11 +-
.../contrib/ethosu/tir/unary_elementwise.py | 6 +-
.../tvm/relay/backend/contrib/ethosu/tir/utils.py | 28 +-
.../backend/contrib/ethosu/tir_to_cs_translator.py | 31 +-
python/tvm/relay/backend/contrib/ethosu/util.py | 14 +
python/tvm/relay/build_module.py | 31 +-
python/tvm/relay/frontend/onnx.py | 31 +
python/tvm/relay/frontend/pytorch.py | 2 +
python/tvm/relay/frontend/pytorch_utils.py | 6 +-
python/tvm/relay/frontend/tflite.py | 18 -
python/tvm/relay/op/_tensor.py | 4 +
python/tvm/relay/op/_transform.py | 6 +-
python/tvm/relay/op/contrib/__init__.py | 1 +
python/tvm/relay/op/contrib/cmsisnn.py | 13 +-
python/tvm/relay/op/contrib/dnnl.py | 246 +-
python/tvm/relay/op/contrib/ethosn.py | 14 +
python/tvm/relay/op/contrib/ethosu.py | 109 +
.../__init__.py => relay/op/contrib/libtorch.py} | 28 +-
python/tvm/relay/op/contrib/tensorrt.py | 140 +-
python/tvm/relay/op/dyn/_transform.py | 2 +-
python/tvm/relay/op/dyn/nn/_nn.py | 10 +-
python/tvm/relay/op/strategy/cuda.py | 64 +-
python/tvm/relay/op/tensor.py | 36 +
python/tvm/relay/qnn/op/legalizations.py | 37 +-
python/tvm/relay/qnn/op/qnn.py | 27 +
python/tvm/relay/testing/tf.py | 53 +-
.../transform/fake_quantization_to_integer.py | 34 +
python/tvm/script/context_maintainer.py | 3 +
python/tvm/script/parser.py | 65 +-
python/tvm/script/tir/__init__.pyi | 15 +-
python/tvm/script/tir/intrin.py | 7 +
python/tvm/script/tir/node.py | 6 +-
python/tvm/script/tir/scope_handler.py | 49 +-
python/tvm/script/tir/special_stmt.py | 54 +
python/tvm/target/target.py | 10 +-
python/tvm/te/__init__.py | 8 +-
python/tvm/te/hybrid/parser.py | 2 +-
python/tvm/te/schedule.py | 152 +-
python/tvm/te/tensor.py | 5 +-
python/tvm/tir/buffer.py | 38 +
python/tvm/tir/expr.py | 4 +
python/tvm/tir/function.py | 37 +-
python/tvm/tir/ir_builder.py | 112 +-
python/tvm/tir/stmt.py | 14 +-
python/tvm/tir/transform/transform.py | 15 +-
python/tvm/topi/cuda/batch_matmul.py | 13 +-
python/tvm/topi/cuda/batch_matmul_tensorcore.py | 2 +
python/tvm/topi/cuda/conv2d_alter_op.py | 17 +-
python/tvm/topi/cuda/conv2d_int8.py | 17 +-
python/tvm/topi/cuda/conv2d_nhwc_winograd.py | 27 +-
python/tvm/topi/cuda/conv2d_winograd.py | 20 +-
python/tvm/topi/cuda/dense.py | 12 +-
python/tvm/topi/cuda/group_conv2d_nchw.py | 11 +-
python/tvm/topi/cuda/sparse.py | 12 +-
python/tvm/topi/cuda/tensor_intrin.py | 26 +-
python/tvm/topi/nn/conv2d.py | 38 +-
python/tvm/topi/nn/pad.py | 8 +-
python/tvm/topi/utils.py | 33 +-
python/tvm/topi/x86/batch_matmul.py | 1 +
python/tvm/topi/x86/dense.py | 1 +
src/arith/canonical_simplify.cc | 1 +
src/arith/const_fold.h | 4 +
src/arith/int_constraints.cc | 6 +-
src/arith/modular_set.cc | 39 +-
src/arith/rewrite_simplify.cc | 18 +-
src/auto_scheduler/feature.cc | 155 +-
src/autotvm/feature_visitor.cc | 10 +-
src/autotvm/feature_visitor.h | 4 +-
src/contrib/ethosu/cascader/block_config.cc | 15 +-
src/contrib/ethosu/cascader/block_config.h | 11 +-
src/contrib/ethosu/cascader/graph.cc | 1 +
src/contrib/ethosu/cascader/graph.h | 7 +-
src/contrib/ethosu/cascader/parts/ethosu.cc | 5 +-
src/contrib/ethosu/cascader/parts/inline.cc | 4 +-
src/contrib/ethosu/cascader/plan_generator.cc | 53 +-
src/contrib/ethosu/cascader/tensor_config.cc | 9 +-
src/contrib/ethosu/cascader/tensor_config.h | 12 +-
src/contrib/hybrid/codegen_hybrid.cc | 8 +
src/contrib/hybrid/codegen_hybrid.h | 2 +
src/driver/driver_api.cc | 1 +
.../feature_extractor/per_store_feature.cc | 20 +-
src/meta_schedule/integration.cc | 76 +-
.../measure_callback/echo_statistics.cc | 4 +-
.../measure_callback/update_cost_model.cc | 15 +-
src/meta_schedule/postproc/verify_gpu_code.cc | 3 +-
.../schedule_rule/multi_level_tiling.cc | 13 +-
src/meta_schedule/schedule_rule/winograd.cc | 96 +
.../space_generator/post_order_apply.cc | 54 +-
src/meta_schedule/task_scheduler/task_scheduler.cc | 6 +-
src/printer/tir_text_printer.cc | 3 +
src/printer/tvmscript_printer.cc | 135 +-
src/relay/backend/aot_executor_codegen.cc | 49 +-
src/relay/backend/build_module.cc | 9 +-
src/relay/backend/contrib/cmsisnn/relay_to_tir.cc | 55 +-
.../contrib/cmsisnn/scalar_to_tensor_constant.cc | 98 +-
src/relay/backend/contrib/dnnl/codegen.cc | 42 +-
src/relay/backend/contrib/dnnl/query_layout.cc | 378 ++
src/relay/backend/contrib/ethosn/codegen.cc | 28 +
src/relay/backend/contrib/ethosn/codegen_ethosn.h | 1 +
src/relay/backend/contrib/ethosn/ethosn_api.cc | 28 +
src/relay/backend/contrib/ethosn/ethosn_api.h | 6 +
src/relay/backend/contrib/ethosu/codegen.cc | 76 +
.../contrib/example_target_hooks/relay_to_tir.cc | 8 +-
.../backend/contrib/libtorch/libtorch_codegen.cc | 138 +
src/relay/backend/contrib/tensorrt/codegen.cc | 13 +-
src/relay/backend/task_extraction.cc | 86 +
src/relay/backend/te_compiler.cc | 19 +-
src/relay/backend/te_compiler_cache.cc | 284 +-
src/relay/backend/te_compiler_cache.h | 11 +
src/relay/backend/utils.cc | 53 +
src/relay/backend/utils.h | 41 +-
src/relay/backend/vm/compiler.cc | 9 +-
src/relay/backend/vm/manifest_lifetimes.cc | 2 +
src/relay/ir/dataflow_matcher.cc | 1 +
src/relay/op/contrib/ethosu/binary_elementwise.cc | 98 +-
src/relay/op/contrib/ethosu/convolution.cc | 85 +-
src/relay/op/contrib/ethosu/depthwise.cc | 89 +-
src/relay/op/contrib/ethosu/identity.cc | 29 +-
src/relay/op/contrib/ethosu/op_attrs.h | 480 ++
src/relay/op/contrib/ethosu/pooling.cc | 81 +-
src/relay/op/contrib/ethosu/unary_elementwise.cc | 62 +-
src/relay/op/nn/convolution.cc | 48 +-
src/relay/op/tensor/binary.cc | 10 +
src/relay/op/tensor/transform.cc | 6 +-
src/relay/qnn/op/leaky_relu.cc | 130 +
src/relay/transforms/fold_constant.cc | 11 +-
src/relay/transforms/type_infer.cc | 48 +-
src/runtime/aot_executor/aot_executor.cc | 7 +-
src/runtime/contrib/cublas/cublas.cc | 2 +-
src/runtime/contrib/dnnl/dnnl_json_runtime.cc | 291 +-
.../contrib/ethosu/bare_metal/tvm_ethosu_runtime.c | 8 +-
src/runtime/contrib/libtorch/libtorch_runtime.cc | 169 +
src/runtime/contrib/tensorrt/tensorrt_builder.cc | 29 +-
src/runtime/contrib/tensorrt/tensorrt_builder.h | 2 +-
src/runtime/contrib/tensorrt/tensorrt_ops.cc | 38 +-
src/runtime/contrib/tensorrt/tensorrt_ops.h | 2 +-
src/runtime/contrib/tensorrt/tensorrt_runtime.cc | 14 +-
src/runtime/crt/microtvm_rpc_server/rpc_server.cc | 2 +-
.../hexagon/android/target/fastrpc/CMakeLists.txt | 46 +-
src/runtime/hexagon/hexagon/hexagon_buffer.cc | 3 +
src/runtime/hexagon/hexagon/hexagon_common.cc | 5 +-
.../hexagon/hexagon/hexagon_device_api_v2.cc | 85 +-
.../hexagon/hexagon/hexagon_device_api_v2.h | 17 +
src/runtime/hexagon/rpc/hexagon/rpc_server.cc | 3 +
src/runtime/hexagon/rpc/simulator/session.cc | 1 +
src/runtime/minrpc/minrpc_server.h | 2 +-
src/runtime/opencl/opencl_device_api.cc | 30 +-
src/runtime/pipeline/pipeline_struct.h | 297 +-
src/runtime/pipeline/spsc_queue.h | 83 +
src/runtime/thread_pool.cc | 1 -
src/runtime/threading_backend.cc | 2 +-
src/runtime/vm/vm.cc | 2 +-
src/runtime/vulkan/vulkan_device.cc | 25 +-
src/runtime/vulkan/vulkan_device.h | 1 +
src/runtime/vulkan/vulkan_device_api.cc | 5 +
src/target/llvm/codegen_cpu.cc | 14 +-
src/target/llvm/codegen_hexagon.cc | 39 +-
src/target/llvm/codegen_llvm.cc | 326 +-
src/target/llvm/codegen_llvm.h | 39 +-
src/target/source/codegen_c.cc | 211 +-
src/target/source/codegen_c.h | 11 +-
src/target/source/codegen_c_host.cc | 4 +
src/target/source/codegen_cuda.cc | 59 +-
src/target/source/codegen_cuda.h | 3 +-
src/target/source/codegen_metal.cc | 5 +
src/target/source/codegen_opencl.cc | 32 +-
src/target/source/codegen_opencl.h | 7 +-
src/target/source/codegen_source_base.cc | 4 +
src/target/source/intrin_rule_cuda.cc | 2 +
src/target/source/ptx_mma.cc | 1806 ++-----
src/target/source/ptx_mma.h | 30 +-
src/target/spirv/codegen_spirv.cc | 64 +-
src/target/spirv/codegen_spirv.h | 4 +-
src/target/spirv/ir_builder.cc | 38 +
src/target/spirv/ir_builder.h | 11 +
src/target/spirv/spirv_support.cc | 13 +
src/target/spirv/spirv_support.h | 14 +
src/target/stackvm/codegen_stackvm.cc | 46 +-
src/target/stackvm/codegen_stackvm.h | 2 +
src/target/tag.cc | 29 +-
src/target/target_kind.cc | 8 +-
src/te/operation/create_primfunc.cc | 7 +-
.../operation/create_primfunc.h} | 29 +-
src/te/operation/cross_thread_reduction.cc | 46 +-
src/te/operation/extern_op.cc | 1 -
src/te/operation/hybrid_op.cc | 13 +-
src/te/operation/op_utils.cc | 41 +-
src/te/operation/scan_op.cc | 2 -
src/te/operation/tensor_compute_op.cc | 1 -
src/te/schedule/bound.cc | 8 +-
src/te/schedule/message_passing.cc | 159 +-
src/te/schedule/schedule_dataflow_rewrite.cc | 18 +-
src/te/schedule/schedule_lang.cc | 96 +
src/te/schedule/schedule_ops.cc | 64 +-
src/te/schedule/schedule_postproc_to_primfunc.cc | 287 +-
src/te/tensor.cc | 30 +-
src/tir/analysis/block_access_region_detector.cc | 10 +-
src/tir/analysis/buffer_access_lca_detector.cc | 21 +-
src/tir/analysis/device_constraint_utils.cc | 27 +-
src/tir/analysis/var_touch.cc | 14 +-
src/tir/analysis/verify_gpu_code.cc | 12 +-
src/tir/analysis/verify_memory.cc | 14 +-
src/tir/ir/buffer.cc | 241 +-
src/tir/ir/data_layout.cc | 1 -
src/tir/ir/expr.cc | 26 +-
src/tir/ir/expr_functor.cc | 12 +-
src/tir/ir/function.cc | 10 +-
src/tir/ir/index_map.cc | 154 +
src/tir/ir/stmt.cc | 30 +-
src/tir/ir/stmt_functor.cc | 82 +-
src/tir/op/builtin.cc | 9 +-
src/tir/schedule/ir_comparator.cc | 1 -
src/tir/schedule/primitive/blockize_tensorize.cc | 5 +-
src/tir/schedule/primitive/cache_read_write.cc | 30 +-
src/tir/schedule/primitive/compute_inline.cc | 42 +-
src/tir/transforms/arg_binder.cc | 77 +-
src/tir/transforms/bf16_legalize.cc | 108 +-
src/tir/transforms/bind_params.cc | 9 +-
src/tir/transforms/bound_checker.cc | 166 +-
src/tir/transforms/compact_buffer_region.cc | 9 +-
src/tir/transforms/coproc_sync.cc | 19 +-
src/tir/transforms/flatten_buffer.cc | 105 +-
src/tir/transforms/inject_copy_intrin.cc | 42 +-
src/tir/transforms/inject_double_buffer.cc | 93 +-
src/tir/transforms/inject_virtual_thread.cc | 160 +-
src/tir/transforms/ir_utils.cc | 128 +-
src/tir/transforms/ir_utils.h | 16 +-
src/tir/transforms/lower_cross_thread_reduction.cc | 2 +-
src/tir/transforms/lower_custom_datatypes.cc | 78 +-
src/tir/transforms/lower_match_buffer.cc | 19 +-
src/tir/transforms/lower_thread_allreduce.cc | 262 +-
src/tir/transforms/lower_tvm_builtin.cc | 208 +-
src/tir/transforms/lower_vtcm_alloc.cc | 80 +
src/tir/transforms/lower_warp_memory.cc | 107 +-
src/tir/transforms/make_packed_api.cc | 87 +-
.../merge_dynamic_shared_memory_allocations.cc | 87 +-
src/tir/transforms/narrow_datatype.cc | 74 +-
src/tir/transforms/renormalize_split_pattern.cc | 2 +
src/tir/transforms/rewrite_unsafe_select.cc | 15 +-
src/tir/transforms/simplify.cc | 34 +-
src/tir/transforms/split_host_device.cc | 26 +-
src/tir/transforms/storage_access.cc | 49 +-
src/tir/transforms/storage_access.h | 9 +-
src/tir/transforms/storage_flatten.cc | 618 ++-
src/tir/transforms/storage_rewrite.cc | 353 +-
src/tir/transforms/texture_flatten.cc | 9 +-
src/tir/transforms/thread_storage_sync.cc | 71 +-
src/tir/transforms/unify_thread_binding.cc | 6 +-
src/tir/transforms/unroll_loop.cc | 5 +
src/tir/transforms/update_pointer_storage_scope.cc | 56 +-
src/tir/transforms/update_pointer_storage_scope.h | 8 +
src/tir/transforms/vectorize_loop.cc | 191 +-
src/tir/usmp/analysis/extract_buffer_info.cc | 12 +-
src/tir/usmp/transform/assign_pool_info.cc | 4 +-
.../convert_pool_allocations_to_offsets.cc | 109 +-
tests/cpp/relay/df_pattern_rewrite_test.cc | 100 +
tests/cpp/tir_analysis_side_effect.cc | 5 +-
tests/lint/check_file_type.py | 2 +-
tests/micro/zephyr/test_zephyr_aot.py | 2 +-
tests/micro/zephyr/test_zephyr_armv7m.py | 3 +-
.../python/contrib/test_cmsisnn/test_binary_ops.py | 11 +-
.../test_cmsisnn/test_scalar_to_tensor_constant.py | 198 +-
tests/python/contrib/test_cublas.py | 12 +-
tests/python/contrib/test_cutlass.py | 20 +
tests/python/contrib/test_dnnl.py | 158 +-
tests/python/contrib/test_ethosn/test_tanh.py | 80 +
.../contrib/test_ethosu/cascader/conftest.py | 30 +-
.../cascader/test_ethosu_block_config.py | 9 +
.../test_ethosu/cascader/test_ethosu_part.py | 2 +-
.../cascader/test_ethosu_part_performance.py | 6 +-
tests/python/contrib/test_ethosu/infra.py | 105 +-
tests/python/contrib/test_ethosu/test_codegen.py | 51 +-
.../contrib/test_ethosu/test_encode_constants.py | 136 +-
.../contrib/test_ethosu/test_identity_optimizer.py | 321 ++
tests/python/contrib/test_ethosu/test_legalize.py | 180 +-
.../contrib/test_ethosu/test_lookup_table.py | 1 -
.../test_ethosu/test_remove_concatenates.py | 29 +-
.../contrib/test_ethosu/test_replace_conv2d.py | 188 +-
.../contrib/test_ethosu/test_replace_copy.py | 44 +-
.../test_ethosu/test_replace_unary_elementwise.py | 4 +-
tests/python/contrib/test_ethosu/test_scheduler.py | 32 +-
.../test_ethosu/test_tir_to_cs_translator.py | 234 +-
tests/python/contrib/test_ethosu/test_vela_api.py | 26 +-
tests/python/contrib/test_hexagon/README_RPC.md | 364 ++
tests/python/contrib/test_hexagon/conftest.py | 27 +-
.../test_hexagon/test_2d_physical_buffers.py | 263 +
.../contrib/test_hexagon/test_cache_read_write.py | 25 +-
tests/python/contrib/test_hexagon/test_launcher.md | 4 +-
tests/python/contrib/test_hexagon/test_launcher.py | 358 +-
tests/python/contrib/test_libtorch_ops.py | 91 +
tests/python/contrib/test_tensorrt.py | 480 +-
tests/python/frontend/onnx/test_forward.py | 44 +
tests/python/frontend/pytorch/test_forward.py | 13 +
tests/python/frontend/tflite/test_forward.py | 512 +-
tests/python/integration/test_tuning.py | 1 +
tests/python/relay/aot/aot_test_utils.py | 236 +-
tests/python/relay/aot/test_c_device_api.py | 40 +-
tests/python/relay/aot/test_crt_aot.py | 50 +-
tests/python/relay/test_op_level1.py | 4 +
tests/python/relay/test_op_qnn_conv2_transpose.py | 2 +-
tests/python/relay/test_op_qnn_leaky_relu.py | 65 +
.../test_pass_fake_quantization_to_integer.py | 50 +
tests/python/relay/test_pass_fold_constant.py | 19 +
tests/python/relay/test_pipeline_executor.py | 1 +
tests/python/relay/test_relay_te_compiler.py | 24 +
tests/python/topi/python/test_topi_conv2d_int8.py | 245 +-
.../python/test_topi_group_conv2d_transpose.py | 45 +-
tests/python/unittest/test_arith_modular_set.py | 27 +
.../python/unittest/test_arith_rewrite_simplify.py | 9 +
.../python/unittest/test_auto_scheduler_feature.py | 28 +
tests/python/unittest/test_ci.py | 47 +-
tests/python/unittest/test_lower_build.py | 36 +-
.../unittest/test_meta_schedule_byoc_tensorrt.py | 187 +-
.../unittest/test_meta_schedule_cost_model.py | 9 +-
.../test_meta_schedule_custom_rule_winograd_cpu.py | 209 +
...test_meta_schedule_custom_rule_winograd_cuda.py | 243 +
.../test_meta_schedule_feature_extractor.py | 8 +-
...schedule_feature_extractor_per_store_feature.py | 50 +-
.../unittest/test_meta_schedule_integration.py | 141 +-
.../test_meta_schedule_measure_callback.py | 89 +-
.../test_meta_schedule_post_order_apply.py | 51 +-
.../test_meta_schedule_postproc_verify_gpu_code.py | 93 +-
tests/python/unittest/test_meta_schedule_runner.py | 19 +-
.../unittest/test_meta_schedule_search_strategy.py | 51 +-
.../unittest/test_meta_schedule_space_generator.py | 16 +-
.../unittest/test_meta_schedule_task_scheduler.py | 99 +-
.../unittest/test_meta_schedule_tune_relay.py | 360 +-
.../test_runtime_module_based_interface.py | 2 +-
tests/python/unittest/test_runtime_module_load.py | 2 +-
tests/python/unittest/test_target_codegen_cuda.py | 48 +-
.../python/unittest/test_target_codegen_hexagon.py | 4 +-
tests/python/unittest/test_target_codegen_llvm.py | 11 +-
.../python/unittest/test_target_codegen_vulkan.py | 5 +-
tests/python/unittest/test_target_target.py | 18 +-
tests/python/unittest/test_te_tensor.py | 4 -
.../test_tir_analysis_calculate_workspace.py | 60 +-
.../test_tir_analysis_detect_buffer_access_lca.py | 6 +-
tests/python/unittest/test_tir_buffer.py | 30 +-
tests/python/unittest/test_tir_constructor.py | 25 +-
tests/python/unittest/test_tir_intrin.py | 5 +-
tests/python/unittest/test_tir_ir_builder.py | 4 +-
.../python/unittest/test_tir_lower_match_buffer.py | 4 +-
tests/python/unittest/test_tir_nodes.py | 17 +-
tests/python/unittest/test_tir_ptx_mma.py | 151 +-
tests/python/unittest/test_tir_ptx_mma_sp.py | 346 ++
.../unittest/test_tir_schedule_cache_read_write.py | 6 +-
.../unittest/test_tir_schedule_compute_inline.py | 20 +-
tests/python/unittest/test_tir_schedule_reorder.py | 4 +-
tests/python/unittest/test_tir_schedule_rfactor.py | 94 +
.../unittest/test_tir_schedule_split_fuse.py | 6 +-
.../python/unittest/test_tir_schedule_tensorize.py | 104 +
.../test_tir_transform_combine_context_call.py | 6 +-
.../test_tir_transform_common_subexpr_elim.py | 30 +-
.../test_tir_transform_compact_buffer_region.py | 19 +-
.../test_tir_transform_convert_for_loops_serial.py | 20 +-
.../test_tir_transform_extract_constants.py | 6 +-
.../unittest/test_tir_transform_flatten_buffer.py | 56 +-
.../test_tir_transform_inject_double_buffer.py | 4 +-
.../test_tir_transform_inject_rolling_buffer.py | 62 +-
.../test_tir_transform_inject_virtual_thread.py | 41 +-
...test_tir_transform_instrument_bound_checkers.py | 6 +-
.../python/unittest/test_tir_transform_ir_utils.py | 9 +-
.../unittest/test_tir_transform_loop_partition.py | 80 +-
...t_tir_transform_lower_cross_thread_reduction.py | 24 +-
.../test_tir_transform_lower_tvm_builtin.py | 4 +-
.../unittest/test_tir_transform_make_packed_api.py | 16 +-
.../unittest/test_tir_transform_narrow_datatype.py | 14 +-
.../unittest/test_tir_transform_remove_no_op.py | 2 +-
...test_tir_transform_renormalize_split_pattern.py | 49 +-
.../python/unittest/test_tir_transform_simplify.py | 4 +-
.../unittest/test_tir_transform_storage_flatten.py | 47 +-
.../unittest/test_tir_transform_unroll_loop.py | 27 +-
.../unittest/test_tir_transform_vectorize.py | 9 +-
tests/python/unittest/test_tir_usmp_algo.py | 124 +-
.../test_tir_usmp_analysis_extract_bufferinfo.py | 540 +--
...ransform_convert_pool_allocations_to_offsets.py | 381 +-
tests/python/unittest/test_tir_usmp_utils.py | 40 +-
tests/python/unittest/test_transform_layout.py | 543 +++
tests/python/unittest/test_tvmscript_complete.py | 6 -
.../python/unittest/test_tvmscript_error_report.py | 5 +-
tests/python/unittest/test_tvmscript_roundtrip.py | 5012 ++++++++++----------
tests/scripts/ci.py | 329 +-
tests/scripts/cmd_utils.py | 61 +
tests/scripts/git_change_docs.sh | 14 +-
tests/scripts/git_skip_ci.py | 10 +-
tests/scripts/git_skip_ci_globs.py | 76 +
tests/scripts/setup-pytest-env.sh | 29 +-
tests/scripts/should_run_slow_tests.py | 0
tests/scripts/task_build.py | 83 +
tests/scripts/task_build.sh | 19 -
tests/scripts/task_ci_setup.sh | 1 +
tests/scripts/task_config_build_arm.sh | 6 +-
tests/scripts/task_config_build_cpu.sh | 6 +-
tests/scripts/task_config_build_gpu.sh | 6 +-
tests/scripts/task_config_build_gpu_other.sh | 5 +-
tests/scripts/task_config_build_hexagon.sh | 5 +-
tests/scripts/task_config_build_i386.sh | 6 +-
tests/scripts/task_config_build_qemu.sh | 6 +-
tests/scripts/task_config_build_wasm.sh | 6 +-
tests/scripts/task_cpp_unittest.sh | 5 +-
tests/scripts/task_python_docs.sh | 1 +
tests/scripts/task_python_hexagon.sh | 5 +-
tests/scripts/task_python_hexagon_simulator.sh | 5 +-
tests/scripts/task_python_microtvm.sh | 4 -
tests/scripts/task_python_topi.sh | 2 +-
tests/scripts/task_show_node_info.sh | 10 +-
vta/python/vta/transform.py | 97 +-
528 files changed, 24328 insertions(+), 11582 deletions(-)
rename .github/{CODEOWNERS => CODEOWNERSHIP} (91%)
delete mode 100644 cmake/libs/hexagon_rpc_skel/CMakeLists.txt
copy python/tvm/contrib/debugger/debug_runtime.py =>
cmake/modules/contrib/LibTorch.cmake (64%)
mode change 100755 => 100644 include/tvm/auto_scheduler/feature.h
copy include/tvm/runtime/{executor_info.h => contrib/libtorch_runtime.h} (67%)
create mode 100644 include/tvm/tir/index_map.h
create mode 100644 python/tvm/contrib/hexagon/tools.py
rename python/tvm/contrib/hexagon/{hexagon.py => transform.py} (61%)
delete mode 100644 python/tvm/meta_schedule/testing/byoc_trt.py
create mode 100644 python/tvm/meta_schedule/testing/conv2d_winograd_cpu.py
create mode 100644 python/tvm/meta_schedule/testing/conv2d_winograd_cuda.py
create mode 100644 python/tvm/meta_schedule/testing/custom_builder_runner.py
create mode 100644
python/tvm/meta_schedule/testing/run_subgraph_auto_scheduler.py
create mode 100644
python/tvm/meta_schedule/testing/run_subgraph_meta_schedule.py
create mode 100644 python/tvm/meta_schedule/testing/tlcbench.py
copy python/tvm/{meta_schedule/task_scheduler/__init__.py =>
relay/op/contrib/libtorch.py} (57%)
mode change 100755 => 100644 src/auto_scheduler/feature.cc
create mode 100644 src/meta_schedule/schedule_rule/winograd.cc
create mode 100755 src/relay/backend/contrib/dnnl/query_layout.cc
create mode 100644 src/relay/backend/contrib/libtorch/libtorch_codegen.cc
create mode 100644 src/relay/backend/task_extraction.cc
create mode 100644 src/relay/op/contrib/ethosu/op_attrs.h
create mode 100644 src/relay/qnn/op/leaky_relu.cc
create mode 100644 src/runtime/contrib/libtorch/libtorch_runtime.cc
create mode 100644 src/runtime/pipeline/spsc_queue.h
copy src/{target/func_registry_generator.h => te/operation/create_primfunc.h}
(63%)
create mode 100644 src/tir/ir/index_map.cc
create mode 100644 src/tir/transforms/lower_vtcm_alloc.cc
create mode 100644 tests/cpp/relay/df_pattern_rewrite_test.cc
create mode 100644 tests/python/contrib/test_ethosn/test_tanh.py
create mode 100644 tests/python/contrib/test_ethosu/test_identity_optimizer.py
create mode 100644 tests/python/contrib/test_hexagon/README_RPC.md
create mode 100755
tests/python/contrib/test_hexagon/test_2d_physical_buffers.py
create mode 100644 tests/python/contrib/test_libtorch_ops.py
create mode 100644 tests/python/relay/test_op_qnn_leaky_relu.py
create mode 100644
tests/python/unittest/test_meta_schedule_custom_rule_winograd_cpu.py
create mode 100644
tests/python/unittest/test_meta_schedule_custom_rule_winograd_cuda.py
create mode 100644 tests/python/unittest/test_tir_ptx_mma_sp.py
create mode 100755 tests/python/unittest/test_transform_layout.py
create mode 100644 tests/scripts/cmd_utils.py
create mode 100755 tests/scripts/git_skip_ci_globs.py
mode change 100644 => 100755 tests/scripts/should_run_slow_tests.py
create mode 100755 tests/scripts/task_build.py
delete mode 100755 tests/scripts/task_build.sh
