This is an automated email from the ASF dual-hosted git repository.
mehrdadh pushed a change to branch ci-docker-staging
in repository https://gitbox.apache.org/repos/asf/tvm.git
discard 619466c4ae Validating new gpu image
add 3414b3fae0 [MetaSchedule] Allow optional params to be None (#11188)
add e23a1b8d57 Make microtvm_template_projects available in tutorials.
(#11164)
add 9ea4fa2450 [fix] vec * mat in matmul in onnx converter (#11174)
add f14070e8f9 [AOT] Return module list from AotExecutorFactory (#11191)
add b772d27273 Enable USE_CUSTOM_LOGGING for Hexagon Launcher (#11185)
add 552f06ed45 support round-trip for T.Ptr in tvmscript (#11179)
add 26cefab5df [TIR] Bind iter domain in analyzer in CreatePrimFunc
(#11187)
add 17b687e400 [ONNX] Reshape op (#11047)
add 5ad70b7a8e [CI] Update GPU image to use newer CMake (#11194)
add b6b0bafdef [Relay] Create header file for realize.cc (#11093)
add 8eae317d28 Remove support for run-time linked-params from codegen
(#11144)
add be90c656e8 [FIX] Avoid stack overflow in TargetHookVisitor with large
modules (#11135)
add 9284d32e3a [TRT] Add check to support split op with TRT 5.1.5+ (#11154)
add 169f824d69 [TIR] Reduced duplication in op.h (#11129)
add 6e23e22b17 [TRT] Add check to use setBindingDimensions in TRT 6.0.1+
(#11178)
add b878dcf316 Run cpplint in quiet mode. (#10292)
add f00c8db5ce Fix make format (#11197)
add 91d351c1f7 [CI] Update GoogleTest (#11162)
add 5733d9754a [microNPU] Match requantize in min/max with activation
pattern (#11010)
add 58a85b80ed [docker][RVM][microtvm] Refactor CMSIS installation to add
to RVM (#11148)
add 0fb155c3c9 [Hexagon] Add support for on-device unit testing using
gtest (#11145)
add 633fb54614 [Runtime][PipelineExecutor] Refactor PipelineExecutor.py
and Add cross compile support for pipeline executor. (#11133)
add 530091acce [Hexagon][Runtime] Add QuRT thread pool backend (#11018)
add 5c204c6246 Remove device type dependency (#11198)
add eb3ce911d0 [Hexagon] Add schedule and test for conv2d_transpose_nchw
(#11175)
add 90084ab8db [PROFILER] Theoretical roofline models (#11066)
add 184b3a9474 [FIX,AUTO_SCHEDULER] Handle manually unrolled loops in auto
scheduler features (#11166)
add e1917f4943 [TIR] Fix an index out of bound issue of compact buffer
region (#11201)
add 01e160616f [MetaSchedule] Logging Interface Unification (#11157)
add a0c61038b6 [COMMUNITY] Altan Haan -> Reviewers (#11205)
add f0854ae456 [COMMUNITY] Xiyou Zhou -> Committer (#11206)
add 3d0fc36ae4 [CI] Update GoogleTest in ci_wasm (#11207)
add 8de64cfba1 [Hexagon] Update launcher cmake flags for Android (#11213)
add 521b80af21 [Relay] Support 'external codegen targets'. (#11173)
add e1acd4b872 [Hexagon] Removes directory after stopping the server
(#11212)
add bafa3e9c22 [Hexagon] Add AoT capability to Hexagon launcher (#11214)
add aa3bcd9d33 Implemented rpc logging (#10967)
add 5007033086 [TIR] Fix printing enum in TransformLayout::AsPython
(#11211)
add eae836cdf6 Fix mixed precision output type to original type (#11142)
add ff7efe767a Revert "Implemented rpc logging (#10967)" (#11227)
add f9d9221739 [COMMUNITY] Nicola Lancellotti -> Reviewers (#11226)
add fb32997369 [CI] Update all Docker Images to 20220505-060045-500703308
(#11219)
add 98aa41e329 [Relay] Flexible shape dispatch transformation (#11199)
add 4622dfefd6 [CI] fix docker group exists with different GID (#11184)
add 31be30062b [TIR] Add schedule primitive SetAxisSeparator (#11225)
add 5393f6dc2d Update debugger.rst (#11231)
add bc7f45e765 [AOT] Enable A-Normal Form in the AOT executor (#11091)
add a3d75ae535 [Frontend][PyTorch] Add: Relay stft operator (#11190)
add 62d3a674d8 [LLVM] Fix a possible tbaa issue (#11181)
add c75eea775e fix two check typo in codegen (#11240)
add d04907919b [TIR] Fix reverse_compute_at for trivial region with
trivial block var (#11234)
add e854c0a4d6 [CI] Identify non-inclusive language in commits (#11230)
add 731af42d1b [CMSIS-NN] Increase partitioning accuracy for pooling
(#11229)
add 5ecb8c384a [ETHOSN] Minor corner case fixes (#11218)
add 588679ee83 [microNPU] Add support for conv2d running on two cores on
U65 (#10251)
add 66cc14f69b [Hexagon] Add mobilenet test with AOT (#11204)
add 0d6403baa2 [Hexagon]Disable hexagon gtest (#11236)
add 29c6d93290 [tir] remove unused member variable (#11248)
add fb6eb38f31 [microNPU] Remove spurious prints and improve
documentation (#11247)
add 3911b85ed7 [USMP] Change internal workspace section (#11246)
add 9e404f0957 [CMSIS-NN] Fix memory alignment bug in CMSIS-NN demo
(#11221)
add d2a7f93beb [ROOFLINE] Calculate roofline from existing TIR PrimFunc
(#11238)
add 24d6c3f3f0 [LLVM] Make sure all functions have target-related
attributes set (#11222)
add 8d4f4dd73f [microTVM] Add support for host-driven AoT Executor (#11044)
add be2ae9433e [TENSORRT] Improvements and fixes for TensorRT (#11203)
add 0f6abea1ca [OpenCL] Change of OpenCL profiling logic (#11180)
add 0828076732 [ci][docker] Add Jinja2 to image (#11265)
add 025c3a3f69 [ci] Use r5.large nodes for builds and lint (#11258)
add 391b4d99ab [USMP] Fix assert condition for TVMBackendAllocWorkspace
(#11270)
add cfb5674c25 [ETHOSN] Remove remaining support for the N77 variant
(#11262)
add 3be5622c7e [ETHOSN] Adding support for Leaky ReLU (#11261)
add 2049185165 Fix running gtest on Hexagon hardware (#11257)
add f5d2d667f9 [ci] Add --docker-image option to ci.py (#11118)
add 02047e60f1 [ci] Bump i386 shards (#11271)
add bae8216c30 [COMMUNITY] mikepapadim -> Reviewer (#11276)
add 63b6b49f03 fix expand onnx conversion (#11278)
add 4eb6497adb [ARM][Strategy] Fix is_int8_hw_support check function
(#11193)
add 775457ceff [ci][docker] Use sccache everywhere by default (#11267)
add 1a8c64bb91 [Hexagon] capture gtest output and return over FFI (#11239)
add 116ccef024 [Frontend] [Paddle] fix testing problem (#11259)
add 497f5f6223 Fix a case of linking to wrong OpenCL library (#11215)
add 366a566977 [microNPU] Adding a option to enable striping (#11263)
add 1b3990808e [microNPU] Update existing microNPU tutorial for CMSIS-NN
(#11285)
add 53fe596682 [rpc] Implemented rpc logging (#11232)
add ec129cf48c [ci] Run docker prune directly in Jenkins (#11275)
add ca1030cf15 [ci][docs] Seed autotvm tutorial (#11147)
add 384d7791bf [ci][build] Use ninja instead of Makefiles (#10934)
add 3d05362226 [ci] Disable dependabot PRs (#11072)
add bd029cb52c [skip ci][wasm][ci] Fix WASM build and JS doc build (#11299)
add e7f1224a04 Fix json serialization for NDArray (#11303)
add fa834f67eb Prevent simplifing unit IterVar in CreatePrimFunc (#11292)
add 7c75b77df8 [microNPU] Add various options to the cascader (#10509)
add c2d1905779 Add Adreno GPU target and topi supporting textures with
dynamically allocated textures (#11161)
add 2ba4588393 [Hexagon] Update Readme (#11283)
add 80d827018a [logging] LOG(FATAL) calls [[noreturn]] functions (#11310)
add 6c339eaa8f [Hexagon] Remove sim_options from tvm.target.hexagon()
(#11293)
add 0e8107bfc7 [TIR][Arith] Implemented padded inverses in IndexMap
(#11235)
add 2eb742d6da [profiler] Skip i386 skip condition (#11280)
add d871bbd96f [QNN] Enable constant folding for QNN operations. (#11228)
add 78142ad2e4 [Relay] Fix a corner case of fused identity (#11217)
add 636463d16c Consider pad value and input zero point in
FoldExplicitPading (#11127)
add 19ce068149 Avoid use of MemoryInfo when undefined in StorageRewrite
(#11254)
add 17c07013e5 [Hexagon][Docker]Add HEXAGON_SDK_ROOT ENV variable (#11291)
add be65732b2a [ROOFLINE] Roofline analysis over RPC (#11252)
add 2023a20339 [build][hexagon] Respect x86 C/C++ compiler choice (#11312)
add aa67a6a01c [Hexagon] Add USMP tests (#11279)
add 67a72d27d7 [ci][docker] Update images to include sccache changes
(#11314)
add f59c70226b [PYTORCH] [FRONTEND] torch.bool support for data type
conversion (#11290)
add 87366b56ed Oneflow fronted support more model and fix bug (#11321)
add 325f6fed56 [MetaSchedule] Allow Easy Logging Level Setting (#11305)
add 9ab3a1121a [TVMScript] Represent ramp as index slice (#11308)
add eed21eeff2 [ONNX] Fix cast op to/from bfloat16 (#11171)
add 02d57bbc06 [BugFix][Topi] Fix 'duplicated iterator names in the
compute definition' bug of roi_align (#11322)
add df51d979d3 adding ramana to reviewers list (#11311)
add 0e27bf5ee9 [frontend][ONNX]support ConvTranspose explicitly specified
output_shape (#11076)
add 0ed0f1d651 Missed out_layout field of conv1d attrs (#11325)
add 672ce33657 [TIR] Propagate storage scope of undefined vars in
SplitHostDevice. (#11255)
add b4c4215d7f Add vlogging for type-table registration. (#11041)
add ebf3607067 [build][hexagon] fix several compiler warnings (#11245)
add 55b6be598a [TIR] Support affine expressions as indices in reverse
compute inline (#11317)
add 4eff08863b update docker version
add a3a4155943 [TIR] Simplify indices in layout transform (#11330)
add de21c8f2ef [CMSIS-NN] Align CMSIS-NN in TVM to TFLu SHA (#11273)
add 0e2e61912d [CI] Update Docker images for new CMSIS-NN (#11336)
add a4aeb1b462 Merge branch 'main' into hexagon/update_jenkins
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 (619466c4ae)
\
N -- N -- N refs/heads/ci-docker-staging (a4aeb1b462)
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:
.github/dependabot.yml | 15 +
3rdparty/dmlc-core | 2 +-
CMakeLists.txt | 23 +-
CONTRIBUTORS.md | 5 +
Jenkinsfile | 309 ++++--
Makefile | 2 +-
apps/android_camera/app/src/main/jni/tvm_runtime.h | 2 +-
apps/android_rpc/app/src/main/jni/tvm_runtime.h | 2 +-
apps/hexagon_api/CMakeLists.txt | 15 +
apps/hexagon_launcher/CMakeLists.txt | 1 +
apps/hexagon_launcher/README.md | 19 +
apps/hexagon_launcher/cmake/HexagonLauncher.cmake | 3 +
apps/hexagon_launcher/cmake/android/CMakeLists.txt | 4 +-
apps/hexagon_launcher/cmake/hexagon/CMakeLists.txt | 1 +
apps/hexagon_launcher/launcher_core.cc | 37 +-
apps/hexagon_launcher/launcher_core.h | 4 +-
apps/hexagon_launcher/launcher_hexagon.cc | 45 +-
apps/hexagon_launcher/launcher_main.cc | 9 +-
apps/ios_rpc/tvmrpc/TVMRuntime.mm | 2 +-
apps/microtvm/cmsisnn/Makefile | 17 +-
apps/microtvm/cmsisnn/convert_image.py | 2 +-
apps/microtvm/cmsisnn/corstone300.ld | 6 +-
apps/microtvm/ethosu/Makefile | 19 +-
apps/microtvm/ethosu/corstone300.ld | 6 +-
apps/microtvm/reference-vm/base-box-tool.py | 1 +
.../zephyr/base-box/base_box_provision.sh | 6 +-
.../reference-vm/zephyr/provision_setup.sh | 1 +
cmake/modules/CUDA.cmake | 4 +
cmake/modules/Hexagon.cmake | 34 +-
cmake/modules/LibInfo.cmake | 1 +
cmake/modules/StandaloneCrt.cmake | 3 +
cmake/utils/FindOpenCL.cmake | 4 +-
docker/Dockerfile.ci_arm | 4 +
docker/Dockerfile.ci_cpu | 9 +
docker/Dockerfile.ci_gpu | 4 +
docker/Dockerfile.ci_hexagon | 10 +-
docker/Dockerfile.ci_i386 | 4 +
docker/Dockerfile.ci_lint | 2 +-
docker/Dockerfile.ci_qemu | 9 +
docker/Dockerfile.ci_wasm | 4 +
docker/install/ubuntu_install_cmsis.sh | 52 +
docker/install/ubuntu_install_core.sh | 24 +-
.../install/ubuntu_install_ethosu_driver_stack.sh | 11 +-
docker/install/ubuntu_install_googletest.sh | 57 +
docker/install/ubuntu_install_sccache.sh | 2 +
docker/with_the_same_user | 7 +-
docs/Doxyfile | 2 +-
docs/arch/debugger.rst | 15 +-
docs/conf.py | 6 +-
docs/how_to/deploy/tensorrt.rst | 4 +
docs/topic/vta/dev/hardware.rst | 2 +-
.../deploy_object_detection_pytorch.py | 2 +-
.../tune_conv2d_layer_cuda.py | 2 +-
.../tune_with_autoscheduler/tune_network_cuda.py | 2 +-
gallery/how_to/work_with_microtvm/micro_ethosu.py | 41 +-
gallery/tutorial/autotvm_relay_x86.py | 7 +-
include/tvm/arith/iter_affine_map.h | 73 +-
include/tvm/ir/module.h | 35 -
include/tvm/meta_schedule/apply_history_best.h | 5 +-
include/tvm/meta_schedule/task_scheduler.h | 23 +-
include/tvm/meta_schedule/tune_context.h | 5 +
include/tvm/relay/attrs/nn.h | 6 +
include/tvm/relay/attrs/transform.h | 25 +
include/tvm/relay/transform.h | 10 +-
include/tvm/runtime/c_runtime_api.h | 16 +-
include/tvm/runtime/crt/aot_executor.h | 107 ++
...aph_executor_module.h => aot_executor_module.h} | 12 +-
include/tvm/runtime/crt/error_codes.h | 12 +-
include/tvm/runtime/crt/graph_executor_module.h | 2 +-
include/tvm/runtime/logging.h | 5 +-
include/tvm/runtime/metadata.h | 69 +-
include/tvm/runtime/metadata_types.h | 89 ++
include/tvm/runtime/profiling.h | 15 +
include/tvm/target/compilation_config.h | 133 +--
include/tvm/target/target.h | 38 +-
include/tvm/target/target_kind.h | 20 +
include/tvm/tir/function.h | 10 -
include/tvm/tir/index_map.h | 20 +-
include/tvm/tir/op.h | 41 +-
include/tvm/tir/schedule/schedule.h | 12 +
include/tvm/topi/transform.h | 3 +-
jenkins/Jenkinsfile.j2 | 201 ++--
jenkins/macros.j2 | 21 +-
python/tvm/_ffi/runtime_ctypes.py | 11 +
python/tvm/auto_scheduler/dispatcher.py | 29 +-
python/tvm/autotvm/task/dispatcher.py | 35 +-
python/tvm/autotvm/task/relay_integration.py | 2 +-
python/tvm/autotvm/tophub.py | 5 +-
python/tvm/contrib/ethosu/cascader/__init__.py | 1 +
python/tvm/contrib/ethosu/cascader/block_config.py | 12 +
.../contrib/ethosu/cascader/cascader_options.py | 28 +
.../tvm/contrib/ethosu/cascader/device_config.py | 85 +-
python/tvm/contrib/ethosu/cascader/logging.py | 70 ++
python/tvm/contrib/ethosu/cascader/pareto.py | 6 +-
.../tvm/contrib/ethosu/cascader/plan_generator.py | 10 +-
python/tvm/contrib/ethosu/cascader/scheduler.py | 45 +-
python/tvm/contrib/hexagon/build.py | 7 +
python/tvm/contrib/nvcc.py | 14 +-
python/tvm/contrib/pipeline_executor.py | 608 +----------
...line_executor.py => pipeline_executor_build.py} | 321 +-----
python/tvm/driver/tvmc/composite_target.py | 5 -
python/tvm/meta_schedule/apply_history_best.py | 13 +-
python/tvm/meta_schedule/builder/local_builder.py | 6 +-
python/tvm/meta_schedule/relay_integration.py | 2 +
python/tvm/meta_schedule/runner/local_runner.py | 5 +-
python/tvm/meta_schedule/runner/rpc_runner.py | 2 +-
.../meta_schedule/task_scheduler/gradient_based.py | 5 +
.../meta_schedule/task_scheduler/round_robin.py | 5 +
.../meta_schedule/task_scheduler/task_scheduler.py | 6 +
.../testing/tune_relay_meta_schedule.py | 6 +-
.../meta_schedule/testing/tune_te_meta_schedule.py | 6 +-
python/tvm/meta_schedule/tune.py | 153 ++-
python/tvm/meta_schedule/tune_context.py | 12 +-
python/tvm/meta_schedule/utils.py | 94 +-
python/tvm/micro/session.py | 1 +
python/tvm/relay/backend/contrib/ethosu/codegen.py | 7 +-
.../tvm/relay/backend/contrib/ethosu/legalize.py | 4 +-
.../backend/contrib/ethosu/tir/convolution.py | 71 +-
.../tvm/relay/backend/contrib/ethosu/tir/passes.py | 168 ++-
.../tvm/relay/backend/contrib/ethosu/tir/spec.py | 4 +
.../tvm/relay/backend/contrib/ethosu/tir/utils.py | 18 -
.../backend/contrib/ethosu/tir_to_cs_translator.py | 22 +-
python/tvm/relay/backend/contrib/ethosu/util.py | 6 +
.../tvm/relay/backend/contrib/ethosu/vela_api.py | 27 +-
python/tvm/relay/backend/graph_executor_codegen.py | 12 +-
python/tvm/relay/backend/vm.py | 91 +-
python/tvm/relay/build_module.py | 20 +-
python/tvm/relay/frontend/oneflow.py | 418 +++++---
python/tvm/relay/frontend/onnx.py | 124 ++-
python/tvm/relay/frontend/pytorch.py | 17 +-
python/tvm/relay/op/_transform.py | 58 +-
python/tvm/relay/op/contrib/cmsisnn.py | 38 +-
python/tvm/relay/op/contrib/ethosn.py | 54 +-
python/tvm/relay/op/contrib/ethosu.py | 47 +-
python/tvm/relay/op/contrib/tensorrt.py | 1123 +++++++++-----------
python/tvm/relay/op/dyn/_transform.py | 10 +-
python/tvm/relay/op/strategy/__init__.py | 1 +
python/tvm/relay/op/strategy/adreno.py | 162 +++
python/tvm/relay/op/strategy/arm_cpu.py | 5 +
python/tvm/relay/op/strategy/cuda.py | 11 +
python/tvm/relay/op/strategy/generic.py | 33 +
python/tvm/relay/op/strategy/hexagon.py | 20 +
python/tvm/relay/op/transform.py | 72 +-
python/tvm/relay/testing/inception_v3.py | 4 +-
python/tvm/relay/transform/__init__.py | 1 +
python/tvm/relay/transform/flexible_shape.py | 369 +++++++
python/tvm/relay/transform/transform.py | 21 +-
python/tvm/rpc/client.py | 13 +-
python/tvm/runtime/ndarray.py | 2 +
python/tvm/runtime/profiling/__init__.py | 47 +
python/tvm/script/parser.py | 30 +-
python/tvm/script/tir/__init__.pyi | 3 +
python/tvm/script/tir/node.py | 63 +-
python/tvm/script/tir/scope_handler.py | 7 +-
python/tvm/script/tir/special_stmt.py | 15 +-
python/tvm/script/tir/ty.py | 27 +-
python/tvm/script/tir/utils.py | 55 -
python/tvm/target/compilation_config.py | 14 +-
python/tvm/target/target.py | 152 +--
.../tvm/testing/usmp.py | 43 +-
python/tvm/tir/function.py | 110 +-
python/tvm/tir/schedule/schedule.py | 81 +-
python/tvm/topi/__init__.py | 2 +
.../tvm/topi/adreno/__init__.py | 24 +-
python/tvm/topi/adreno/conv2d_alter_op.py | 211 ++++
python/tvm/topi/adreno/conv2d_nchw.py | 344 ++++++
python/tvm/topi/adreno/conv2d_nhwc.py | 339 ++++++
python/tvm/topi/adreno/depthwise_conv2d_nchw.py | 316 ++++++
python/tvm/topi/adreno/depthwise_conv2d_nhwc.py | 311 ++++++
python/tvm/topi/adreno/pooling.py | 89 ++
python/tvm/topi/adreno/utils.py | 549 ++++++++++
python/tvm/topi/arm_cpu/conv2d_int8.py | 2 +-
python/tvm/topi/cuda/__init__.py | 1 +
python/tvm/topi/cuda/stft.py | 135 +++
python/tvm/topi/hexagon/conv2d.py | 26 +
python/tvm/topi/stft.py | 125 +++
python/tvm/topi/vision/rcnn/roi_align.py | 4 +-
.../tvm/utils/__init__.py | 18 +-
python/tvm/utils/roofline.py | 488 +++++++++
src/arith/analyzer.cc | 3 +
src/arith/interval_set.h | 8 +-
src/arith/iter_affine_map.cc | 627 ++++++++---
src/arith/rewrite_simplify.cc | 4 +
src/auto_scheduler/feature.cc | 34 +-
src/contrib/ethosu/cascader/cascader_options.cc | 31 +-
src/contrib/ethosu/cascader/cascader_options.h | 19 +-
src/contrib/ethosu/cascader/pareto.cc | 20 +-
src/contrib/ethosu/cascader/pareto.h | 7 +-
src/contrib/ethosu/cascader/plan_generator.cc | 49 +-
src/contrib/ethosu/cascader/plan_generator.h | 5 +-
src/contrib/ethosu/cascader/proposal_generator.cc | 3 +-
src/ir/module.cc | 7 -
src/meta_schedule/apply_history_best.cc | 10 +-
.../measure_callback/echo_statistics.cc | 27 +-
.../schedule_rule/cross_thread_reduction.cc | 10 +-
.../schedule_rule/multi_level_tiling.cc | 2 +-
.../search_strategy/evolutionary_search.cc | 23 +-
.../space_generator/post_order_apply.cc | 8 +-
src/meta_schedule/task_scheduler/gradient_based.cc | 28 +-
src/meta_schedule/task_scheduler/round_robin.cc | 16 +-
src/meta_schedule/task_scheduler/task_scheduler.cc | 37 +-
src/meta_schedule/tune_context.cc | 5 +-
src/meta_schedule/utils.h | 48 +
src/node/structural_hash.cc | 34 +-
src/printer/tvmscript_printer.cc | 40 +-
src/relay/backend/aot_executor_codegen.cc | 152 ++-
src/relay/backend/build_module.cc | 42 +-
src/relay/backend/contrib/cmsisnn/relay_to_tir.cc | 78 +-
.../relay/backend/contrib/cublas/target.cc | 32 +-
.../relay/backend/contrib/cudnn/target.cc | 30 +-
.../relay/backend/contrib/cutlass/target.cc | 31 +-
src/relay/backend/contrib/ethosn/codegen.cc | 46 +-
src/relay/backend/contrib/ethosn/codegen_ethosn.h | 1 +
src/relay/backend/contrib/ethosn/ethosn_api.cc | 38 +-
src/relay/backend/contrib/ethosn/ethosn_api.h | 7 +
src/relay/backend/contrib/ethosu/codegen.cc | 79 +-
src/relay/backend/contrib/ethosu/compiler_attrs.cc | 44 +
.../contrib/example_target_hooks/relay_to_tir.cc | 60 +-
src/relay/backend/contrib/tensorrt/codegen.cc | 178 +++-
.../relay/backend/contrib/tensorrt/target.cc | 30 +-
src/relay/backend/graph_executor_codegen.cc | 33 +-
src/relay/backend/interpreter.cc | 15 +-
src/relay/backend/te_compiler.cc | 60 +-
src/relay/backend/te_compiler.h | 20 +-
src/relay/backend/te_compiler_cache.cc | 28 +-
src/relay/backend/utils.cc | 10 +-
src/relay/backend/utils.h | 5 +-
src/relay/backend/vm/compiler.cc | 76 +-
src/relay/backend/vm/compiler.h | 42 +-
src/relay/op/dyn/tensor/transform.cc | 3 +-
src/relay/op/make_op.h | 2 +-
src/relay/op/tensor/transform.cc | 78 +-
src/relay/quantize/realize.cc | 41 +-
src/relay/quantize/realize.h | 77 ++
src/relay/transforms/fold_constant.cc | 27 +-
src/relay/transforms/fold_explicit_padding.cc | 15 +-
src/relay/transforms/inline_composites.cc | 94 --
src/relay/transforms/target_hooks.cc | 43 +-
src/relay/transforms/to_mixed_precision.cc | 60 +-
src/runtime/aot_executor/aot_executor_factory.cc | 5 +
src/runtime/contrib/json/json_node.h | 6 +
src/runtime/contrib/tensorrt/tensorrt_builder.cc | 53 +-
src/runtime/contrib/tensorrt/tensorrt_calibrator.h | 2 +-
src/runtime/contrib/tensorrt/tensorrt_ops.cc | 306 ++++--
src/runtime/contrib/tensorrt/tensorrt_ops.h | 35 +-
src/runtime/contrib/tensorrt/tensorrt_runtime.cc | 10 +-
src/runtime/crt/Makefile | 2 +
src/runtime/crt/aot_executor/aot_executor.c | 244 +++++
.../crt/aot_executor_module/aot_executor_module.c | 196 ++++
.../graph_executor_module/graph_executor_module.c | 38 +-
src/runtime/crt/host/Makefile | 12 +-
src/runtime/crt/host/main.cc | 5 +
src/runtime/crt/microtvm_rpc_server/rpc_server.cc | 2 -
src/runtime/hexagon/hexagon_buffer.cc | 25 +-
src/runtime/hexagon/hexagon_common.cc | 2 +-
src/runtime/hexagon/hexagon_device_api.cc | 28 +-
src/runtime/hexagon/hexagon_user_dma.cc | 13 +-
src/runtime/hexagon/rpc/simulator/session.cc | 21 +
src/runtime/micro/micro_session.cc | 8 +
src/runtime/minrpc/minrpc_interfaces.h | 93 ++
src/runtime/minrpc/minrpc_logger.cc | 291 +++++
src/runtime/minrpc/minrpc_logger.h | 296 ++++++
src/runtime/minrpc/minrpc_server.h | 651 +++++++-----
src/runtime/minrpc/minrpc_server_logging.h | 166 +++
src/runtime/object.cc | 6 +
src/runtime/opencl/opencl_common.h | 68 +-
src/runtime/opencl/opencl_device_api.cc | 11 +-
src/runtime/opencl/opencl_module.cc | 19 +-
src/runtime/profiling.cc | 133 ++-
src/runtime/rpc/rpc_channel_logger.h | 185 ++++
src/runtime/rpc/rpc_endpoint.h | 2 +
src/runtime/rpc/rpc_socket_impl.cc | 21 +-
src/runtime/texture.h | 6 +
src/runtime/thread_storage_scope.h | 7 +
src/runtime/threading_backend.cc | 89 +-
src/support/libinfo.cc | 5 +
src/target/compilation_config.cc | 281 ++---
src/target/llvm/codegen_cpu.cc | 10 +
src/target/llvm/codegen_hexagon.cc | 19 -
src/target/llvm/codegen_llvm.cc | 163 +--
src/target/llvm/codegen_llvm.h | 26 +-
src/target/llvm/llvm_module.cc | 19 +-
src/target/metadata_module.cc | 19 +-
src/target/metadata_module.h | 3 +-
src/target/source/codegen_c_host.cc | 1 -
src/target/source/codegen_opencl.cc | 2 +-
src/target/source/codegen_source_base.cc | 5 +
src/target/source/source_module.cc | 75 +-
src/target/source/source_module.h | 4 +-
src/target/stackvm/codegen_stackvm.cc | 2 +-
src/target/target.cc | 31 +-
src/target/target_info.cc | 1 +
src/target/target_kind.cc | 6 +-
src/tir/ir/index_map.cc | 72 ++
src/tir/schedule/concrete_schedule.cc | 10 +
src/tir/schedule/concrete_schedule.h | 3 +
src/tir/schedule/primitive.h | 11 +
src/tir/schedule/primitive/block_annotate.cc | 107 +-
src/tir/schedule/primitive/compute_at.cc | 6 +
src/tir/schedule/primitive/compute_inline.cc | 174 ++-
.../schedule/primitive/layout_transformation.cc | 123 ++-
src/tir/schedule/schedule.cc | 7 +-
src/tir/schedule/state.cc | 1 +
src/tir/schedule/traced_schedule.cc | 13 +
src/tir/schedule/traced_schedule.h | 3 +
src/tir/schedule/transform.cc | 81 ++
src/tir/schedule/transform.h | 55 +
src/tir/transforms/compact_buffer_region.cc | 5 +-
src/tir/transforms/legalize_packed_calls.cc | 3 +-
src/tir/transforms/split_host_device.cc | 7 +-
src/tir/transforms/storage_rewrite.cc | 16 +-
src/tir/usmp/transform/assign_pool_info.cc | 1 +
.../hexagon/hexagon_buffer_tests.cc} | 3 +-
tests/cpp-runtime/hexagon/run_all_tests.cc | 68 ++
tests/cpp-runtime/hexagon/run_unit_tests.cc | 122 +++
tests/cpp/aot_metadata_test.cc | 16 +
tests/cpp/relay/transforms/device_domains_test.cc | 5 +-
tests/cpp/relay_build_module_test.cc | 5 +-
tests/cpp/runtime_test.cc | 5 +-
tests/cpp/target/compilation_config_test.cc | 264 +++--
tests/cpp/target_test.cc | 28 +
.../task_web_wasm.sh => lint/blocklint.sh} | 23 +-
tests/lint/cpplint.sh | 5 +-
tests/micro/arduino/test_arduino_workflow.py | 2 +-
.../contrib/test_arm_compute_lib/test_reshape.py | 1 +
tests/python/contrib/test_cmsisnn/test_pooling.py | 73 +-
.../python/contrib/test_ethosn/test_leaky_relu.py | 86 ++
.../contrib/test_ethosn/test_partition_params.py | 50 -
tests/python/contrib/test_ethosu/cascader/infra.py | 10 +
.../cascader/test_ethosu_block_config.py | 143 ++-
.../test_ethosu/cascader/test_memory_reduction.py | 57 +-
.../contrib/test_ethosu/cascader/test_pareto.py | 2 +-
.../test_ethosu/cascader/test_plan_generator.py | 133 ++-
.../cascader/test_proposal_generator.py | 21 +
.../contrib/test_ethosu/cascader/test_scheduler.py | 5 +-
tests/python/contrib/test_ethosu/test_codegen.py | 4 +-
.../contrib/test_ethosu/test_encode_constants.py | 329 ++++--
tests/python/contrib/test_ethosu/test_legalize.py | 13 +
.../test_ethosu/test_remove_concatenates.py | 8 +-
.../contrib/test_ethosu/test_replace_conv2d.py | 102 +-
.../contrib/test_ethosu/test_replace_copy.py | 6 +-
tests/python/contrib/test_ethosu/test_scheduler.py | 4 +-
.../test_ethosu/test_tir_to_cs_translator.py | 34 +-
tests/python/contrib/test_ethosu/test_vela_api.py | 13 +
tests/python/contrib/test_hexagon/README.md | 141 ++-
tests/python/contrib/test_hexagon/conftest.py | 25 +-
.../contrib/test_hexagon/{ => conv2d}/README.md | 0
tests/python/contrib/test_hexagon/test_launcher.md | 138 ---
tests/python/contrib/test_hexagon/test_launcher.py | 13 +-
tests/python/contrib/test_hexagon/test_models.py | 69 +-
.../contrib/test_hexagon/test_run_unit_tests.py | 47 +
.../contrib/test_hexagon/test_thread_pool.py | 101 ++
tests/python/contrib/test_hexagon/test_usmp.py | 112 ++
.../contrib/test_hexagon/topi/test_batch_matmul.py | 5 +-
.../test_hexagon/topi/test_cache_read_write.py | 7 +-
.../contrib/test_hexagon/topi/test_conv2d_nchw.py | 3 +-
.../contrib/test_hexagon/topi/test_conv2d_nhwc.py | 3 +-
.../test_hexagon/topi/test_conv2d_transpose.py | 158 +++
.../python/contrib/test_hexagon/topi/test_dense.py | 10 +-
.../test_hexagon/topi/test_depthwise_conv2d.py | 3 +-
.../contrib/test_hexagon/topi/test_pooling.py | 3 +-
.../contrib/test_hexagon/topi/test_reduce.py | 5 +-
.../contrib/test_hexagon/topi/test_softmax.py | 3 +-
tests/python/contrib/test_tensorrt.py | 42 +-
tests/python/driver/tvmc/test_compiler.py | 24 +-
tests/python/driver/tvmc/test_composite_target.py | 1 -
tests/python/driver/tvmc/test_target.py | 22 +-
tests/python/frontend/caffe/test_forward.py | 8 +-
.../frontend/mxnet/model_zoo/inception_v3.py | 4 +-
tests/python/frontend/oneflow/test_forward.py | 199 ++++
.../python/frontend/oneflow/test_vision_models.py | 150 +++
tests/python/frontend/onnx/test_forward.py | 124 ++-
tests/python/frontend/paddlepaddle/test_forward.py | 5 +
tests/python/frontend/pytorch/test_forward.py | 45 +-
.../frontend/pytorch/test_object_detection.py | 3 +-
tests/python/relay/aot/corstone300.ld | 6 +-
tests/python/relay/aot/corstone300.mk | 7 +-
tests/python/relay/aot/test_cpp_aot.py | 18 +
tests/python/relay/aot/test_crt_aot.py | 47 +
tests/python/relay/aot/test_crt_aot_usmp.py | 47 +-
tests/python/relay/test_auto_scheduler_tuning.py | 6 +
tests/python/relay/test_backend_graph_executor.py | 24 +
tests/python/relay/test_build_module.py | 9 +-
tests/python/relay/test_conv2d_nchw_texture.py | 394 +++++++
tests/python/relay/test_conv2d_nhwc_texture.py | 556 ++++++++++
.../relay/test_depthwise_conv2d_nchw_texture.py | 194 ++++
.../relay/test_depthwise_conv2d_nhwc_texture.py | 233 ++++
tests/python/relay/test_op_level3.py | 158 ++-
.../relay/test_pass_annotate_spans_defuse.py | 9 +-
.../relay/test_pass_flexible_shape_dispatch.py | 119 +++
tests/python/relay/test_pass_fold_constant.py | 183 ++++
.../relay/test_pass_fold_explicit_padding.py | 29 +
tests/python/relay/test_pass_inline_composites.py | 165 ---
tests/python/relay/test_pass_partition_graph.py | 22 +-
tests/python/relay/test_pass_plan_devices.py | 5 +-
tests/python/relay/test_pipeline_executor.py | 25 +-
tests/python/relay/test_to_mixed_precision.py | 39 +-
tests/python/relay/test_vm.py | 2 +-
tests/python/relay/utils/adreno_utils.py | 118 ++
tests/python/target/test_arm_target.py | 56 +
.../python/unittest/test_auto_scheduler_feature.py | 23 +-
tests/python/unittest/test_autotvm_record.py | 5 +
tests/python/unittest/test_crt.py | 78 +-
tests/python/unittest/test_index_map.py | 189 ++++
.../unittest/test_meta_schedule_tune_relay.py | 4 +-
tests/python/unittest/test_node_reflection.py | 45 +-
tests/python/unittest/test_runtime_profiling.py | 100 ++
tests/python/unittest/test_runtime_rpc.py | 23 +-
.../python/unittest/test_target_codegen_hexagon.py | 61 --
tests/python/unittest/test_target_codegen_llvm.py | 44 +
tests/python/unittest/test_target_target.py | 43 +-
.../python/unittest/test_tir_schedule_analysis.py | 12 +-
.../unittest/test_tir_schedule_compute_at.py | 43 +
.../unittest/test_tir_schedule_compute_inline.py | 140 +++
.../test_tir_schedule_set_axis_separator.py | 139 +++
.../unittest/test_tir_schedule_transform_layout.py | 40 +
tests/python/unittest/test_tir_texture_scope.py | 62 ++
.../test_tir_transform_compact_buffer_region.py | 58 +-
.../python/unittest/test_tvmscript_error_report.py | 51 +-
tests/python/unittest/test_tvmscript_roundtrip.py | 50 +
tests/scripts/ci.py | 35 +-
tests/scripts/task_build.py | 9 +-
tests/scripts/task_build_hexagon_api.sh | 6 +-
tests/scripts/task_ci_setup.sh | 11 +-
tests/scripts/task_cpp_unittest.sh | 22 +-
tests/scripts/task_lint.sh | 84 +-
tests/scripts/task_mypy.sh | 1 +
tests/scripts/task_python_hexagon.sh | 3 +
tests/scripts/task_web_wasm.sh | 3 +-
web/emcc/wasm_runtime.cc | 2 +-
web/tsconfig.json | 27 +-
web/typedoc.json | 2 +-
432 files changed, 18575 insertions(+), 5721 deletions(-)
create mode 100644 .github/dependabot.yml
create mode 100755 docker/install/ubuntu_install_cmsis.sh
create mode 100755 docker/install/ubuntu_install_googletest.sh
create mode 100644 include/tvm/runtime/crt/aot_executor.h
copy include/tvm/runtime/crt/{graph_executor_module.h =>
aot_executor_module.h} (73%)
create mode 100644 include/tvm/runtime/metadata_types.h
create mode 100644 python/tvm/contrib/ethosu/cascader/logging.py
copy python/tvm/contrib/{pipeline_executor.py => pipeline_executor_build.py}
(71%)
create mode 100644 python/tvm/relay/op/strategy/adreno.py
create mode 100644 python/tvm/relay/transform/flexible_shape.py
delete mode 100644 python/tvm/script/tir/utils.py
copy tests/scripts/task_build_hexagon_api.sh => python/tvm/testing/usmp.py
(51%)
mode change 100755 => 100644
copy apps/microtvm/reference-vm/zephyr/base-box/base_box_provision.sh =>
python/tvm/topi/adreno/__init__.py (72%)
create mode 100644 python/tvm/topi/adreno/conv2d_alter_op.py
create mode 100644 python/tvm/topi/adreno/conv2d_nchw.py
create mode 100644 python/tvm/topi/adreno/conv2d_nhwc.py
create mode 100644 python/tvm/topi/adreno/depthwise_conv2d_nchw.py
create mode 100644 python/tvm/topi/adreno/depthwise_conv2d_nhwc.py
create mode 100644 python/tvm/topi/adreno/pooling.py
create mode 100644 python/tvm/topi/adreno/utils.py
create mode 100644 python/tvm/topi/cuda/stft.py
create mode 100644 python/tvm/topi/stft.py
copy tests/scripts/task_web_wasm.sh => python/tvm/utils/__init__.py (79%)
mode change 100755 => 100644
create mode 100644 python/tvm/utils/roofline.py
copy include/tvm/runtime/crt/graph_executor_module.h =>
src/relay/backend/contrib/cublas/target.cc (53%)
copy include/tvm/runtime/crt/graph_executor_module.h =>
src/relay/backend/contrib/cudnn/target.cc (56%)
copy include/tvm/runtime/crt/graph_executor_module.h =>
src/relay/backend/contrib/cutlass/target.cc (55%)
copy include/tvm/runtime/crt/graph_executor_module.h =>
src/relay/backend/contrib/tensorrt/target.cc (56%)
create mode 100644 src/relay/quantize/realize.h
delete mode 100644 src/relay/transforms/inline_composites.cc
create mode 100644 src/runtime/crt/aot_executor/aot_executor.c
create mode 100644 src/runtime/crt/aot_executor_module/aot_executor_module.c
create mode 100644 src/runtime/minrpc/minrpc_interfaces.h
create mode 100644 src/runtime/minrpc/minrpc_logger.cc
create mode 100644 src/runtime/minrpc/minrpc_logger.h
create mode 100644 src/runtime/minrpc/minrpc_server_logging.h
create mode 100644 src/runtime/rpc/rpc_channel_logger.h
rename tests/{cpp/runtime/hexagon_buffer.cc =>
cpp-runtime/hexagon/hexagon_buffer_tests.cc} (99%)
create mode 100644 tests/cpp-runtime/hexagon/run_all_tests.cc
create mode 100644 tests/cpp-runtime/hexagon/run_unit_tests.cc
copy tests/{scripts/task_web_wasm.sh => lint/blocklint.sh} (60%)
create mode 100644 tests/python/contrib/test_ethosn/test_leaky_relu.py
copy tests/python/contrib/test_hexagon/{ => conv2d}/README.md (100%)
delete mode 100644 tests/python/contrib/test_hexagon/test_launcher.md
create mode 100644 tests/python/contrib/test_hexagon/test_run_unit_tests.py
create mode 100644 tests/python/contrib/test_hexagon/test_thread_pool.py
create mode 100644 tests/python/contrib/test_hexagon/test_usmp.py
create mode 100644
tests/python/contrib/test_hexagon/topi/test_conv2d_transpose.py
create mode 100644 tests/python/frontend/oneflow/test_vision_models.py
create mode 100644 tests/python/relay/test_conv2d_nchw_texture.py
create mode 100644 tests/python/relay/test_conv2d_nhwc_texture.py
create mode 100644 tests/python/relay/test_depthwise_conv2d_nchw_texture.py
create mode 100644 tests/python/relay/test_depthwise_conv2d_nhwc_texture.py
create mode 100644 tests/python/relay/test_pass_flexible_shape_dispatch.py
delete mode 100644 tests/python/relay/test_pass_inline_composites.py
create mode 100644 tests/python/relay/utils/adreno_utils.py
create mode 100644 tests/python/target/test_arm_target.py
create mode 100644 tests/python/unittest/test_index_map.py
create mode 100644
tests/python/unittest/test_tir_schedule_set_axis_separator.py
create mode 100644 tests/python/unittest/test_tir_texture_scope.py
