profile
viewpoint
Jared Roesch jroesch University of Washington Seattle, Washington OctoML, Apache TVM, leading Relay IR, PhD student @ UW. Previously Rustc, Lean & F* hacker.

ekmett/sparse 23

sparse matrices in Morton order

hubris-lang/hubris 17

A dependently typed programming language, and verification tool

grievejia/LLVMDynamicTools 11

An LLVM interpreter that aims to compute points-to sets dynamically

fishuyo/seer 5

audio visual interaction live coding in scala

jroesch/290c-reading-list 5

A reading list with citations for UCSB's CS 290G

andrewberls/crowdcode 4

Crowdsourced code review on rails.

icemelon9/tvm 2

End to end Tensor IR/DSL stack for deploying deep learning workloads to hardwares

jroesch/acetic 2

A playground for experimenting with Rust's type system

andrewberls/btable-haskell 1

Haskell bindings for the BTable binary serialization format

jroesch/Cb 1

A compiler for a C like language with ad-hoc polymorphism.

push eventapache/incubator-tvm

Robert Kimball

commit sha 1831c17998b29f3797f364410980809bfef554ca

Update include and src dir CHECK* to ICHECK* (#6745)

view details

push time in 9 hours

PR merged apache/incubator-tvm

Replace CHECK* with ICHECK*
  • I apologize for large size of this PR
  • To keep the size down some only /include and /src directories were updated
  • Instance of CHECK are placed with ICHECK
  • Instances of CHECK_GT, CHECK_LT, etc. replaced with ICHECK_GT, ICHECK_LT, etc.
  • The file include/tvm/support/logging.h is not modified since it is where ICHECK is defined and ICHECK ultimately calls CHECK
  • In include/tvm/support/logging.h the tvm:: namespace was explicitly added to kTVM_INTERNAL_ERROR_MESSAGE to address compile errors.
  • Files in runtime/crt not updated to use ICHECK
+4251 -4234

0 comment

429 changed files

rkimball

pr closed time in 9 hours

PullRequestReviewEvent

push eventjroesch/tvm

Jared Roesch

commit sha 62f3e39d51f4ca95cb9f9732ca65d0e5515879d6

Remove dead comment

view details

push time in a day

push eventjroesch/tvm

Jared Roesch

commit sha d086193cff3f4ef87a2854dac6ee60fc90f681d4

Format all

view details

push time in a day

push eventjroesch/tvm

Jared Roesch

commit sha 9a0e727b0f5aa6225ec6952ffb9dc601d9d2ad8f

Clean up PR

view details

push time in a day

push eventjroesch/tvm

Jared Roesch

commit sha 7e038e08c9c1aeda2b79d63bf7ce6214882b00f7

Update rust/tvm/src/ir/diagnostics/mod.rs Co-authored-by: Robert Kimball <bobkimball@gmail.com>

view details

push time in a day

push eventjroesch/tvm

Jared Roesch

commit sha 9aa1a091ad94fc257078aed897fced9784d9eeee

Update cmake/modules/RustExt.cmake Co-authored-by: Robert Kimball <bobkimball@gmail.com>

view details

push time in a day

push eventapache/incubator-tvm

Ramana Radhakrishnan

commit sha bb4179e2867a1d3cb9bd56c707681cc66c07d459

Add pytest-xdist and pytest-profiling to the base installation packages. (#6736) For building and testing some small portions of the python testsuite, I've been playing off and on with xdist and pytest-profiling. We know it's not safe for the entirity of CI yet but this could enable smaller parts of pipelines that folks use using the common scripts to be parallelized or indeed profiled for more insight into where time is spent in building and testing TVM

view details

push time in a day

PR merged apache/incubator-tvm

Add pytest-xdist and pytest-profiling to the base installation packages.

For building and testing and parallelizing some small portions of the python testsuite which are multicore safe and parallelizable I've been playing off and on with xdist and pytest-profiling.

We know it's not safe for the entirity of CI yet but this could enable smaller parts of pipelines that folks use using the common scripts to be parallelized or indeed profiled for more insight into where time is spent in building and testing TVM and while I'm not currently proposing the jenkins flow to parallelize more, having these in the base docker images will allow for more experimentation and thought.

if folks would like to try it with existing scripts the environment variable CI_PYTEST_ADD_OPTIONS will help in this regard.

So to use pytest-xdist one could provide

CI_PYTEST_ADD_OPTIONS="-n 4" to run 4 instances using pytest-xdist or CI_PYTEST_ADD_OPTIONS="--profiling --profiling-svg" to get profiling information for their build and test runs.

Alternatively if they were using pytest on the command line they could do so with the same options on their command line.

@tqchen @tmoreau89 - thoughts ?

+1 -1

1 comment

1 changed file

u99127

pr closed time in a day

PullRequestReviewEvent

push eventapache/incubator-tvm

Lily Orth-Smith

commit sha 372d7374d221fb98f7e7fe5d9d5c937059a35515

[RELAY] Refactor FoldConstant to skip TNonComputationalOps (#6720) * add TNonComputational to qnn ops and change FoldConstant * remove comments * check if op in nonComputational map * forgot to mark device_copy op as TNonComputational * hacky fix to fuseops pass * fix typo * manually skip device_copy in fold_constant * Update src/relay/transforms/fold_constant.cc Co-authored-by: Junru Shao <junrushao1994@gmail.com> Co-authored-by: Junru Shao <junrushao1994@gmail.com>

view details

push time in a day

PR merged apache/incubator-tvm

[RELAY] Refactor FoldConstant to skip TNonComputationalOps

This PR refactors FoldConstant to skip TNonComputational ops, instead of explicitly checking for those ops. It also makes all QNN ops TNonComputational. This allows the constant folder to be applied to graphs containing QNN ops.

@jroesch @jwfromm please take a look!

+13 -3

2 comments

8 changed files

electriclilies

pr closed time in a day

push eventapache/incubator-tvm

Jared Roesch

commit sha ee9b6d6de52accf6515d69c7f565aeefdf19bddc

[Docker] Turn on Rust docs and MxNet based ResNet (#6640) * Enable ResNet and Rust docs * Tweak * Format * Fix issue with overwriting

view details

push time in a day

PR merged apache/incubator-tvm

[Docker] Turn on Rust docs and MxNet based ResNet

Tweak a few things in order to build some Rust tests cc @tmoreau89 and @tqchen.

+16 -21

1 comment

6 changed files

jroesch

pr closed time in a day

PR opened apache/incubator-tvm

[Rust][IRModule] Flesh out IRModule methods

This PR fully expand the IRModule interface in Rust and fixes a whole host of sharp edges and improves the procedural macros for generating the boilerplate.

cc @mwillsey @altanh @imalsogreg @gussmith23

+287 -96

0 comment

13 changed files

pr created time in 2 days

create barnchjroesch/tvm

branch : rust-improve-ir-mod

created branch time in 2 days

push eventjroesch/tvm

Jared Roesch

commit sha 7ee00df5f7cf96a38607526654db0231c5905117

Fix issue with overwriting

view details

push time in 3 days

pull request commentapache/incubator-tvm

[COMMUNITY] junrushao1994 -> committer

@zhiics you will probably have to rebase due to auto scheduling CI corruption

zhiics

comment created time in 3 days

push eventjroesch/tvm

Cody Yu

commit sha 9ae386c1f437da368cc168f3d70de29e043b1b7d

[Bugfix] Auto scheduler tutorial failure on CI (#6723)

view details

Jared Roesch

commit sha 136784d8b74a17d2f67de9fcda64ceff3774464c

Enable ResNet and Rust docs

view details

Jared Roesch

commit sha 03f5bde9d63d5e1341637c1a7c8e350849ab7a81

Tweak

view details

Jared Roesch

commit sha 14ac6258da4ee11d0de155d61d6215c7c80a37be

Format

view details

push time in 3 days

push eventjroesch/tvm

Jared Roesch

commit sha db6b35503245e41924ddb4e5aaee806214866dfb

WIP more improvments

view details

push time in 4 days

create barnchjroesch/tvm-cli-rs

branch : master

created branch time in 4 days

created repositoryjroesch/tvm-cli-rs

An example of building a small CLI using TVM and Rust.

created time in 4 days

PullRequestReviewEvent

pull request commentapache/incubator-tvm

[Bugfix] Auto scheduler tutorial failure on CI

@comaniac thanks for doing this!

comaniac

comment created time in 4 days

push eventjroesch/tvm

Jared Roesch

commit sha 06cdc4753434d19a06f5c2702abd5d759d20ec0a

Add docs and address feedback

view details

push time in 4 days

PullRequestReviewEvent

push eventjroesch/tvm

Tristan Konolige

commit sha 82431457cb39807f599a70193fae6cb56463b26d

[FIX,MICROTVM] Skip microtvm tests if microtvm is not built (#6693)

view details

Trevor Morris

commit sha 08b08b175f0ebee89aa80a3296f50c52a9ef1672

[TFLite] Fix detection of crop in convert_batch_to_space_nd (#6670)

view details

Jared Roesch

commit sha 4c4d3dc2d8723a5b78d34f1bb0540ef66ee185a3

Fix tutorial broken by Docker build (#6694)

view details

masahi

commit sha cf49e8b5bf6635099ee691e5dc3cd005aecf3be6

[Torch, Quantization] Necessary workaround to prepare for 1.6 update (#6602) * add support for 1.6 quantized models * fix lint * move version check function to a common utils * fix lint Co-authored-by: masa <masa@pop-os.localdomain>

view details

lixiaoquan

commit sha e997185795480d24075a2e7d3fa42ccec425b5f6

[Relay] Change some passes to mix mode (#6695)

view details

Tianqi Chen

commit sha 461e75bd5ffaf45a0f270998514d444463d11261

[LLVM][WINDOWS] Recover windows support for the latest LLVM (#6698) Windows COFF requires comdat information to support weak-like linkage(via any). This patch fixes the windows LLVM support after LLVM-8.

view details

Tianqi Chen

commit sha 5b8e0c6b9176456e2f8a76a2dd76b474c1d9afdc

Resolve more warnings in msvc (#6702)

view details

Tristan Konolige

commit sha 61245527fd7e7976aa00cc09c266752cf0dab41c

Add cloudpickle dependency to docker images (#6701)

view details

Robert Kimball

commit sha 80cc4802369f12c3e6a4cc186978d0685521159d

Refactor diagnostic to avoid circular dependencies (#6692)

view details

Tianqi Chen

commit sha 66a5f8ce651e430d8e3136d41fe8f7d4ab037be6

[TEST] Address flaky error in test_any (#6705)

view details

Xingjian Shi

commit sha 7c2a2e5175ed8c2f038042fc9a2a188af0dc78e8

[Frontend][Relay] Fix MXNet frontend to support NLP backbones in GluonNLP (#6699) * update Update type_relations.cc Update transform.cc Update transform.cc Update transform.cc Update transform.cc Update transform.cc Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py update Update mxnet.py debug Update generic.py Update topi_integration.py fix bug update Update test_forward.py Update test_forward.py fix test case Update mxnet.py update Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py Update mxnet.py Update mxnet.py debug Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py * address comments * Update mxnet.py * Update mxnet.py * fix * improve where test * Update test_forward.py * Update test_forward.py * Update test_forward.py * update * Update mxnet.py * Update mxnet.py * Update mxnet.py debug Update common.py update Update mxnet.py update Update test_forward.py Update test_forward.py * update * fix lint * Update mxnet.py * Update test_op_level1.py * fix lint

view details

Tianqi Chen

commit sha 13251c31d5a1d007c15279cba87436262e21f792

[PYTHON][WINDOWS] More robust dll loading behavior after python3.8 (#6707) The dll search directories need to be manually added by os.add_dll_directory after python3.8.

view details

Lianmin Zheng

commit sha 94679b5cf46bd89872ae611995ff43f6dae78786

[AutoScheduler] Add task scheduler (#6663) * Add task scheduler * fix lint * fix tests * fix tests * fix tests * fix test cases * fix test cases * fix tests * address comments

view details

Cody Yu

commit sha c5dc98c1c336610eb2763583078312b6f3e07c13

[AutoSchedule] Support multiple cache read and fix bugs (#6686) * Add shape to DAG print * avoid useless cross-thread reduction * Fix stage order * support multiple cache_read * lint * fix * fix * address comment * fix ci * Trigger CI & Update doc strings Co-authored-by: Lianmin Zheng <lianminzheng@gmail.com>

view details

Tianqi Chen

commit sha cc89b76c4d359dcb154d9a0a5dc99cef7c4e094b

[CI] Update docker to latest (#6708)

view details

presburger

commit sha 28e9ab7a8d8faf8bb043994766caf9aa511968b0

Fix the Type bug in ConvertSSA. (#6709) Co-authored-by: YushengMa <yusheng.ma@streamcomputing.com>

view details

Trevor Morris

commit sha af8636a84314fc41134a3777c1a980435aa9663d

[BYOC][TensorRT] TensorRT BYOC integration (#6395) * TensorRT integration using JSONRuntime Support input nodes with multiple data entries Fix failing tests Support layout transform, add engine caching Add comment Add PruneSubgraph pass Use prune_subgraph pass, make params member of trt runtime class Hide deprecation warnings coming from TRT headers Remove general prune subgraph Save/load use_implicit_batch and workspace size Clean up Fix cpp lint Addressing review comments Refactor tests Use relay.bind instead of VarReplacer. Improve some annotation functions Add TRT docs Use DLOG, formatting Use logging.info instead of print also refactor integ tests also refactor integ tests Formatting Formatting Format python fix python format Fix pylint Fix sphinx precheck Add tensorrt.rst to toctree Allow codegen to be tested when TRT runtime is not available. Enable TRT codegen in CI linty Address more comments Formatting Formatting * Documentation changes * Address comments * Rename USE_TENSORRT->USE_TENSORRT_CODEGEN and USE_TENSORRT_GRAPH_RUNTIME->USE_TENSORRT_RUNTIME * Fix comment typo * Test CI without TRT codegen enabled * formatting * Enable USE_TENSORRT_CODEGEN in CI * Change file_util.h -> file_utils.h

view details

Josh Fromm

commit sha e1b053ac07bf6f52ec56713801de5dd040d7f4ac

[Relay][Frontend][Onnx] Loop Support (#6700) * Onnx loop almost working, checkpointing for safety. * Very close to working. * Last piece is fixing scan initialization. * snapshotting for debug. * Fix Josh's issue * Use subgraph proto class. * Loop with scan. * Simple loop test now working. * Scan outputs now working. * Added second loop test. * Removed unneeded helper functions. * Remove bad merge artifact. * Cleaned up scan output creation. * Cleaned up some style mistakes. * Add pylint skip for unused-argument. * Remove onnx dependency. * Remove now obsolete checks for 0 shaped tensors. Co-authored-by: Jared Roesch <jroesch@octoml.ai>

view details

Jared Roesch

commit sha e6f992e74534516084e77bea52985f1a37150bc9

Add initial boilerplate for Rust diagnostic interface.

view details

Jared Roesch

commit sha 4fe35b0444ac106c3bb6f91338e026f8c6d52f95

Codespan example almost working

view details

push time in 5 days

Pull request review commentapache/incubator-tvm

[Relay] Mix mode type inference

 Expr PostOrderRewrite(const Expr& expr, ExprRewriter* rewriter);  */ void PostOrderVisit(const Expr& node, std::function<void(const Expr&)> fvisit); +/*!+ * \brief A function to iteratively traverse dataflow regions of a graph+ *+ * ExpandDataflow manually manages a stack and performs DFS to determine the processing+ * order of nodes in an input graph.+ *+ * If it finds a dataflow node (Call, Tuple, TupleGetItem), it checks if the arguments to that node+ * need to be processed via fcheck_visited. If so, the function pushes those arguments to the stack+ * and continues iteratively to process the top of the stack. When it finds a node that doesn't+ * match the dataflow types, or a node who's inputs have all been processed, it visits the current+ * leaf via fvisit_leaf.+ *+ * This function should be used internally to other classes to implement mixed-mode traversals. The+ * expectation is that fvisit_leaf will perform recursive analysis within mixed-mode traversal if it+ * hits a non-dataflow node.+ *+ * fcheck_visited and fvisit_leaf are templated to encourage compiler inlining.+ */+template <typename FCheckVisited, typename FVisitLeaf>

Why do we need to expose this in the header?

lixiaoquan

comment created time in 5 days

PullRequestReviewEvent

push eventjroesch/tvm

Tristan Konolige

commit sha 82431457cb39807f599a70193fae6cb56463b26d

[FIX,MICROTVM] Skip microtvm tests if microtvm is not built (#6693)

view details

Trevor Morris

commit sha 08b08b175f0ebee89aa80a3296f50c52a9ef1672

[TFLite] Fix detection of crop in convert_batch_to_space_nd (#6670)

view details

Jared Roesch

commit sha 4c4d3dc2d8723a5b78d34f1bb0540ef66ee185a3

Fix tutorial broken by Docker build (#6694)

view details

masahi

commit sha cf49e8b5bf6635099ee691e5dc3cd005aecf3be6

[Torch, Quantization] Necessary workaround to prepare for 1.6 update (#6602) * add support for 1.6 quantized models * fix lint * move version check function to a common utils * fix lint Co-authored-by: masa <masa@pop-os.localdomain>

view details

lixiaoquan

commit sha e997185795480d24075a2e7d3fa42ccec425b5f6

[Relay] Change some passes to mix mode (#6695)

view details

Tianqi Chen

commit sha 461e75bd5ffaf45a0f270998514d444463d11261

[LLVM][WINDOWS] Recover windows support for the latest LLVM (#6698) Windows COFF requires comdat information to support weak-like linkage(via any). This patch fixes the windows LLVM support after LLVM-8.

view details

Tianqi Chen

commit sha 5b8e0c6b9176456e2f8a76a2dd76b474c1d9afdc

Resolve more warnings in msvc (#6702)

view details

Tristan Konolige

commit sha 61245527fd7e7976aa00cc09c266752cf0dab41c

Add cloudpickle dependency to docker images (#6701)

view details

Robert Kimball

commit sha 80cc4802369f12c3e6a4cc186978d0685521159d

Refactor diagnostic to avoid circular dependencies (#6692)

view details

Tianqi Chen

commit sha 66a5f8ce651e430d8e3136d41fe8f7d4ab037be6

[TEST] Address flaky error in test_any (#6705)

view details

Xingjian Shi

commit sha 7c2a2e5175ed8c2f038042fc9a2a188af0dc78e8

[Frontend][Relay] Fix MXNet frontend to support NLP backbones in GluonNLP (#6699) * update Update type_relations.cc Update transform.cc Update transform.cc Update transform.cc Update transform.cc Update transform.cc Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py Update mxnet.py update Update mxnet.py debug Update generic.py Update topi_integration.py fix bug update Update test_forward.py Update test_forward.py fix test case Update mxnet.py update Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py Update mxnet.py Update mxnet.py debug Update mxnet.py Update mxnet.py Update test_forward.py Update mxnet.py * address comments * Update mxnet.py * Update mxnet.py * fix * improve where test * Update test_forward.py * Update test_forward.py * Update test_forward.py * update * Update mxnet.py * Update mxnet.py * Update mxnet.py debug Update common.py update Update mxnet.py update Update test_forward.py Update test_forward.py * update * fix lint * Update mxnet.py * Update test_op_level1.py * fix lint

view details

Tianqi Chen

commit sha 13251c31d5a1d007c15279cba87436262e21f792

[PYTHON][WINDOWS] More robust dll loading behavior after python3.8 (#6707) The dll search directories need to be manually added by os.add_dll_directory after python3.8.

view details

Lianmin Zheng

commit sha 94679b5cf46bd89872ae611995ff43f6dae78786

[AutoScheduler] Add task scheduler (#6663) * Add task scheduler * fix lint * fix tests * fix tests * fix tests * fix test cases * fix test cases * fix tests * address comments

view details

Cody Yu

commit sha c5dc98c1c336610eb2763583078312b6f3e07c13

[AutoSchedule] Support multiple cache read and fix bugs (#6686) * Add shape to DAG print * avoid useless cross-thread reduction * Fix stage order * support multiple cache_read * lint * fix * fix * address comment * fix ci * Trigger CI & Update doc strings Co-authored-by: Lianmin Zheng <lianminzheng@gmail.com>

view details

Tianqi Chen

commit sha cc89b76c4d359dcb154d9a0a5dc99cef7c4e094b

[CI] Update docker to latest (#6708)

view details

presburger

commit sha 28e9ab7a8d8faf8bb043994766caf9aa511968b0

Fix the Type bug in ConvertSSA. (#6709) Co-authored-by: YushengMa <yusheng.ma@streamcomputing.com>

view details

Trevor Morris

commit sha af8636a84314fc41134a3777c1a980435aa9663d

[BYOC][TensorRT] TensorRT BYOC integration (#6395) * TensorRT integration using JSONRuntime Support input nodes with multiple data entries Fix failing tests Support layout transform, add engine caching Add comment Add PruneSubgraph pass Use prune_subgraph pass, make params member of trt runtime class Hide deprecation warnings coming from TRT headers Remove general prune subgraph Save/load use_implicit_batch and workspace size Clean up Fix cpp lint Addressing review comments Refactor tests Use relay.bind instead of VarReplacer. Improve some annotation functions Add TRT docs Use DLOG, formatting Use logging.info instead of print also refactor integ tests also refactor integ tests Formatting Formatting Format python fix python format Fix pylint Fix sphinx precheck Add tensorrt.rst to toctree Allow codegen to be tested when TRT runtime is not available. Enable TRT codegen in CI linty Address more comments Formatting Formatting * Documentation changes * Address comments * Rename USE_TENSORRT->USE_TENSORRT_CODEGEN and USE_TENSORRT_GRAPH_RUNTIME->USE_TENSORRT_RUNTIME * Fix comment typo * Test CI without TRT codegen enabled * formatting * Enable USE_TENSORRT_CODEGEN in CI * Change file_util.h -> file_utils.h

view details

Josh Fromm

commit sha e1b053ac07bf6f52ec56713801de5dd040d7f4ac

[Relay][Frontend][Onnx] Loop Support (#6700) * Onnx loop almost working, checkpointing for safety. * Very close to working. * Last piece is fixing scan initialization. * snapshotting for debug. * Fix Josh's issue * Use subgraph proto class. * Loop with scan. * Simple loop test now working. * Scan outputs now working. * Added second loop test. * Removed unneeded helper functions. * Remove bad merge artifact. * Cleaned up scan output creation. * Cleaned up some style mistakes. * Add pylint skip for unused-argument. * Remove onnx dependency. * Remove now obsolete checks for 0 shaped tensors. Co-authored-by: Jared Roesch <jroesch@octoml.ai>

view details

Jared Roesch

commit sha 35ea47e5a784c20f7077b0ef3840ad16dd8b39b4

Enable ResNet and Rust docs

view details

Jared Roesch

commit sha 7b955f376748a0bd960fbe6dd699fef1033d3b22

Tweak

view details

push time in 5 days

Pull request review commentapache/incubator-tvm

[Rust][Diagnostics] Add initial boilerplate for Rust diagnostic interface.

+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)

I am not a great CMake user so just let me know what the best way to do things is. I just have no clue what the "right way" in CMake is.

jroesch

comment created time in 8 days

PullRequestReviewEvent

Pull request review commentapache/incubator-tvm

[Rust][Diagnostics] Add initial boilerplate for Rust diagnostic interface.

+/*+ * Licensed to the Apache Software Foundation (ASF) under one+ * or more contributor license agreements.  See the NOTICE file+ * distributed with this work for additional information+ * regarding copyright ownership.  The ASF licenses this file+ * to you under the Apache License, Version 2.0 (the+ * "License"); you may not use this file except in compliance+ * with the License.  You may obtain a copy of the License at+ *+ *   http://www.apache.org/licenses/LICENSE-2.0+ *+ * Unless required by applicable law or agreed to in writing,+ * software distributed under the License is distributed on an+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY+ * KIND, either express or implied.  See the License for the+ * specific language governing permissions and limitations+ * under the License.+ */++use env_logger;+use tvm;+use tvm::runtime::function::register_override;++fn test_fn() -> Result<(), tvm::Error> {+    println!("Hello Greg from Rust!");+    Ok(())+}++fn test_fn2(message: tvm::runtime::string::String) -> Result<(), tvm::Error> {+    println!("The message: {}", message);+    Ok(())+}++tvm::export!(test_fn, test_fn2);++#[no_mangle]+fn compiler_ext_initialize() -> i32 {

I added it.

jroesch

comment created time in 8 days

PullRequestReviewEvent

Pull request review commentapache/incubator-tvm

[Rust][Diagnostics] Add initial boilerplate for Rust diagnostic interface.

+/*+ * Licensed to the Apache Software Foundation (ASF) under one+ * or more contributor license agreements.  See the NOTICE file+ * distributed with this work for additional information+ * regarding copyright ownership.  The ASF licenses this file+ * to you under the Apache License, Version 2.0 (the+ * "License"); you may not use this file except in compliance+ * with the License.  You may obtain a copy of the License at+ *+ *   http://www.apache.org/licenses/LICENSE-2.0+ *+ * Unless required by applicable law or agreed to in writing,+ * software distributed under the License is distributed on an+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY+ * KIND, either express or implied.  See the License for the+ * specific language governing permissions and limitations+ * under the License.+ */++use env_logger;+use tvm;+use tvm::runtime::function::register_override;++fn test_fn() -> Result<(), tvm::Error> {+    println!("Hello Greg from Rust!");+    Ok(())+}++fn test_fn2(message: tvm::runtime::string::String) -> Result<(), tvm::Error> {+    println!("The message: {}", message);+    Ok(())+}++tvm::export!(test_fn, test_fn2);++#[no_mangle]+fn compiler_ext_initialize() -> i32 {

I think extern just sets symbol visibility, seems to have worked.

jroesch

comment created time in 8 days

PullRequestReviewEvent

push eventjroesch/tvm

Jared Roesch

commit sha 22ec391d26cefc5c3f3d030b359a0a42c51475de

Fix the extension code

view details

push time in 8 days

push eventjroesch/tvm

Jared Roesch

commit sha 8a6ae776f723af37cfa71cb7297bccd6e16a8435

Remove type checker

view details

Jared Roesch

commit sha c4f21d2b9394047c76831ada249ab816e84d5d13

Format and cleanup

view details

push time in 8 days

Pull request review commentapache/incubator-tvm

[Rust][Diagnostics] Add initial boilerplate for Rust diagnostic interface.

+[package]+name = "compiler-ext"+version = "0.1.0"+authors = ["Jared Roesch <jroesch@octoml.ai>"]+edition = "2018"+# TODO(@jroesch): would be cool to figure out how to statically link instead.

I figured out how to do this.

jroesch

comment created time in 8 days

PullRequestReviewEvent
PullRequestReviewEvent

pull request commentapache/incubator-tvm

[Frontend][Relay] Fix MXNet frontend to support NLP backbones in GluonNLP

As we add more tests can we measure what kind of time increase this will induce in CI? integration tests are becoming increasingly slow and expensive to run. cc @areusch and @tkonolige

sxjscience

comment created time in 8 days

push eventapache/incubator-tvm

Tianqi Chen

commit sha 461e75bd5ffaf45a0f270998514d444463d11261

[LLVM][WINDOWS] Recover windows support for the latest LLVM (#6698) Windows COFF requires comdat information to support weak-like linkage(via any). This patch fixes the windows LLVM support after LLVM-8.

view details

push time in 8 days

PR merged apache/incubator-tvm

[LLVM][WINDOWS] Recover windows support for the latest LLVM

Windows COFF requires comdat information to support weak-like linkage(via any). This patch fixes the windows LLVM support after LLVM-8.

+31 -3

2 comments

4 changed files

tqchen

pr closed time in 8 days

PullRequestReviewEvent

push eventjroesch/tvm

Tristan Konolige

commit sha 6d0351a7f0e23eb5428c59a976edd2bfb8207c0d

Faster sparse_dense on GPUs (#6580) * Faster sparse_dense on GPUs. This new sparse_dense requires a padded matrix, so a new op `sparse_dense_padded` has been added. AlterOpLayout should transform `sparse_dense` to `sparse_dense_padded` when possible on the gpu. * formatting * more formatting * Check that alteroplayout is definedbefore using it * check if FTVMAlterOpLayout exists before using it * formatting * restore message passing * Fix sparse_dense and sparse_dense_padded docs * Fix old sparse_dense, autotvm and sparse_dense dont play well together * Remove unused imports * clarify warp count in cuda_transpose * Document multidimensional access * Warn users not to use sparse_dense_padded * rename nn.sparse_dense_padded to nn.internal.sparse_dense_padded

view details

masahi

commit sha 5a3377460b75336c937ff1cbc9fac3ff8b69af31

Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" (#6658) * Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" This reverts commit 782190e88b1941fdbe31101af260bee06b81bf72. * run infer type on test_sparse_dense_padded_alter_op() to fix CI Co-authored-by: masa <masa@pop-os.localdomain>

view details

Lianmin Zheng

commit sha f6657a6b87104eddd1fc46e5ddc490837fadc1b5

[AutoScheduler] Improve the GPU tutorial by deleting measure_ctx earlier (#6660) * del measurement process in the tutorial * fix * trigger CI

view details

Lianmin Zheng

commit sha dd60d249e50f29c3cb34704693ba54803ae75a36

[AutoScheduler] Improve test cases (#6657) * Improve test cases * update * fix lint * fix lint * trigger CI * address comments * trigger CI

view details

Jorn Tuyls

commit sha bf21371363fff978d35a27bc40d4f4fecf734aa2

[Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing (#6589) * Fix conv2d_transpose parsing in Tensorflow frontend for TF 1.15 * Add comments and convolution tests without AddShapesToGraphDef

view details

Dmitriy Smirnov

commit sha e561007f0c330e3d14c2bc8a3ef40fb741db9004

[BYOC][ACL] Support add operation (#6532) * [BYOC][ACL] Support add operation Added support for an "add" operation implemented via ACL for fp32 and quantized uint8 data types * Addressed lhutton1 comments * linter

view details

Chris Gyurgyik

commit sha d3ef1373dad8414b5825baa8012123d987e686d4

Fix typographical error. (#6664)

view details

Trevor Morris

commit sha 74b692269809a54b9f5a8160d90e401b9a9fc37d

[Relay][MXNet] Support broadcast_like (#6561)

view details

Tianqi Chen

commit sha 0cdd285abe58ac373c60c7544178888a47902d0d

[CI] Move to use main as the default (#6665)

view details

masahi

commit sha b277f181c4430ac705a533b016ac35fbf98ec2b8

[Torch] Object detection support update for PyTorch 1.6 (#6659) * update split * fix * cast nms output to int64 * add more comment and numel test * fix lint * also supported the latest master (1.7) Co-authored-by: masa <masa@pop-os.localdomain>

view details

Andrew Reusch

commit sha d955718cfe300d3ed294cacb0c74f8ea73ae6ad0

don't validate AttrInitEntry until a value has attempted to be set (#6672)

view details

Zhiqiang Wang

commit sha 4073adc15fc55d6232d00ff07dd1384bd9f936ff

[CI] Set main as default in github actions (#6669)

view details

Ritwik Das

commit sha d5728bdeb8e4eb3e633c608366a0c5795a236d93

[BYOC] Support control flow in annotate_target (#6641) * Change annotate target * Annotate_target * Revert namespace changes * Add tests for if-else node * Add while_let testcase * No merging in ifelse * Remove scope builder * Add ops * Replace < with less * Linter * Pass Tests * Change back to static const * Cpplinter * address PR comments' * PR Comments * Clang-format check * PR Comments * PR Comments * Change back to Insert Ann in AnnotateARgs Co-authored-by: Ritwik Das <dasritwi@3c22fb14d7c6.ant.amazon.com> Co-authored-by: Ubuntu <ubuntu@ip-172-31-3-223.us-west-2.compute.internal>

view details

Animesh Jain

commit sha d24634a7cc3c502176838d238af42cf7c94defbe

TF argmax - handling int64 datatype (#6674) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-202.us-west-2.compute.internal>

view details

MORITA Kazutaka

commit sha f196a81e73322776ca1c70d413cc266bacaa42cc

[CODEGEN][COREML] Call InferType explicitly in coreml test (#6676)

view details

Ishi Tatsuyuki

commit sha 8d9ca2ab4ce0be2ecf5fcfa5ea2b8b63a6de7d23

Adjust Vulkan queue selection and creation logic (#6662) * Adjust Vulkan queue selection and creation logic - The queue selection logic rewrite addresses a bug in the old implementation where the code would pass an invalid queue index when selecting any queues other than number 0. - The new implementation will attempt to use compute-only queues which are common on AMD GPUs. It's not clear how much difference will this make but hopefully it would lead to better scheduling. The queue changes were made as with the old configuration autotvm caused my system to hang, stutter or otherwise become unstable and crash. With the change I'm able to run autotvm tuning inside a desktop environment. * Return multiple queue family indexes from GetComputeQueues * Only create one queue

view details

Trevor Morris

commit sha d7fa9c1f2539e1be8527a88745fcd0bed9be2f0f

Install xgboost>=1.1.0 in CI container (#6679)

view details

qixiuai

commit sha 7d805b54d6adda82636d13bf7c46a2e9a933da5f

Fix format error in integrate.rst (#6677)

view details

Przemyslaw Tredak

commit sha e66b2e83db24c78c892308475687ff3bbb0173b9

Revert #5238 (#6680)

view details

Lianmin Zheng

commit sha 78bba3c74785433b0f29589d89112ef06e4bca75

[AutoScheduler] Fix a bug in thread binding (#6683) * fix for lstm use case * update

view details

push time in 8 days

push eventjroesch/tvm

Jared Roesch

commit sha 9feb2df8fdd119e9a331096f1aa934335388fda6

Fix

view details

push time in 9 days

push eventjroesch/tvm

Jared Roesch

commit sha 1cc234e3dadaaebc4803c6344af10f8bb9b7a315

Fix calling

view details

push time in 9 days

push eventjroesch/tvm

Tristan Konolige

commit sha 6d0351a7f0e23eb5428c59a976edd2bfb8207c0d

Faster sparse_dense on GPUs (#6580) * Faster sparse_dense on GPUs. This new sparse_dense requires a padded matrix, so a new op `sparse_dense_padded` has been added. AlterOpLayout should transform `sparse_dense` to `sparse_dense_padded` when possible on the gpu. * formatting * more formatting * Check that alteroplayout is definedbefore using it * check if FTVMAlterOpLayout exists before using it * formatting * restore message passing * Fix sparse_dense and sparse_dense_padded docs * Fix old sparse_dense, autotvm and sparse_dense dont play well together * Remove unused imports * clarify warp count in cuda_transpose * Document multidimensional access * Warn users not to use sparse_dense_padded * rename nn.sparse_dense_padded to nn.internal.sparse_dense_padded

view details

masahi

commit sha 5a3377460b75336c937ff1cbc9fac3ff8b69af31

Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" (#6658) * Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" This reverts commit 782190e88b1941fdbe31101af260bee06b81bf72. * run infer type on test_sparse_dense_padded_alter_op() to fix CI Co-authored-by: masa <masa@pop-os.localdomain>

view details

Lianmin Zheng

commit sha f6657a6b87104eddd1fc46e5ddc490837fadc1b5

[AutoScheduler] Improve the GPU tutorial by deleting measure_ctx earlier (#6660) * del measurement process in the tutorial * fix * trigger CI

view details

Lianmin Zheng

commit sha dd60d249e50f29c3cb34704693ba54803ae75a36

[AutoScheduler] Improve test cases (#6657) * Improve test cases * update * fix lint * fix lint * trigger CI * address comments * trigger CI

view details

Jorn Tuyls

commit sha bf21371363fff978d35a27bc40d4f4fecf734aa2

[Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing (#6589) * Fix conv2d_transpose parsing in Tensorflow frontend for TF 1.15 * Add comments and convolution tests without AddShapesToGraphDef

view details

Dmitriy Smirnov

commit sha e561007f0c330e3d14c2bc8a3ef40fb741db9004

[BYOC][ACL] Support add operation (#6532) * [BYOC][ACL] Support add operation Added support for an "add" operation implemented via ACL for fp32 and quantized uint8 data types * Addressed lhutton1 comments * linter

view details

Chris Gyurgyik

commit sha d3ef1373dad8414b5825baa8012123d987e686d4

Fix typographical error. (#6664)

view details

Trevor Morris

commit sha 74b692269809a54b9f5a8160d90e401b9a9fc37d

[Relay][MXNet] Support broadcast_like (#6561)

view details

Tianqi Chen

commit sha 0cdd285abe58ac373c60c7544178888a47902d0d

[CI] Move to use main as the default (#6665)

view details

masahi

commit sha b277f181c4430ac705a533b016ac35fbf98ec2b8

[Torch] Object detection support update for PyTorch 1.6 (#6659) * update split * fix * cast nms output to int64 * add more comment and numel test * fix lint * also supported the latest master (1.7) Co-authored-by: masa <masa@pop-os.localdomain>

view details

Andrew Reusch

commit sha d955718cfe300d3ed294cacb0c74f8ea73ae6ad0

don't validate AttrInitEntry until a value has attempted to be set (#6672)

view details

Zhiqiang Wang

commit sha 4073adc15fc55d6232d00ff07dd1384bd9f936ff

[CI] Set main as default in github actions (#6669)

view details

Ritwik Das

commit sha d5728bdeb8e4eb3e633c608366a0c5795a236d93

[BYOC] Support control flow in annotate_target (#6641) * Change annotate target * Annotate_target * Revert namespace changes * Add tests for if-else node * Add while_let testcase * No merging in ifelse * Remove scope builder * Add ops * Replace < with less * Linter * Pass Tests * Change back to static const * Cpplinter * address PR comments' * PR Comments * Clang-format check * PR Comments * PR Comments * Change back to Insert Ann in AnnotateARgs Co-authored-by: Ritwik Das <dasritwi@3c22fb14d7c6.ant.amazon.com> Co-authored-by: Ubuntu <ubuntu@ip-172-31-3-223.us-west-2.compute.internal>

view details

Animesh Jain

commit sha d24634a7cc3c502176838d238af42cf7c94defbe

TF argmax - handling int64 datatype (#6674) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-202.us-west-2.compute.internal>

view details

MORITA Kazutaka

commit sha f196a81e73322776ca1c70d413cc266bacaa42cc

[CODEGEN][COREML] Call InferType explicitly in coreml test (#6676)

view details

Ishi Tatsuyuki

commit sha 8d9ca2ab4ce0be2ecf5fcfa5ea2b8b63a6de7d23

Adjust Vulkan queue selection and creation logic (#6662) * Adjust Vulkan queue selection and creation logic - The queue selection logic rewrite addresses a bug in the old implementation where the code would pass an invalid queue index when selecting any queues other than number 0. - The new implementation will attempt to use compute-only queues which are common on AMD GPUs. It's not clear how much difference will this make but hopefully it would lead to better scheduling. The queue changes were made as with the old configuration autotvm caused my system to hang, stutter or otherwise become unstable and crash. With the change I'm able to run autotvm tuning inside a desktop environment. * Return multiple queue family indexes from GetComputeQueues * Only create one queue

view details

Trevor Morris

commit sha d7fa9c1f2539e1be8527a88745fcd0bed9be2f0f

Install xgboost>=1.1.0 in CI container (#6679)

view details

qixiuai

commit sha 7d805b54d6adda82636d13bf7c46a2e9a933da5f

Fix format error in integrate.rst (#6677)

view details

Przemyslaw Tredak

commit sha e66b2e83db24c78c892308475687ff3bbb0173b9

Revert #5238 (#6680)

view details

Lianmin Zheng

commit sha 78bba3c74785433b0f29589d89112ef06e4bca75

[AutoScheduler] Fix a bug in thread binding (#6683) * fix for lstm use case * update

view details

push time in 9 days

pull request commentapache/incubator-tvm

[Relay][Frontend] SparseTensorDenseMatMul support for Tensorflow

cc @tkonolige

ANSHUMAN87

comment created time in 9 days

PullRequestReviewEvent

push eventjroesch/tvm

Lianmin Zheng

commit sha 78bba3c74785433b0f29589d89112ef06e4bca75

[AutoScheduler] Fix a bug in thread binding (#6683) * fix for lstm use case * update

view details

Tianqi Chen

commit sha 60ed9261058c0f1faa2632cb97319f29b26b9573

[ARITH] Introduce iterator (quasi)affine map detection. (#6667) * [ARITH] Introduce iterator (quasi)affine map detection. The loop transformations (split, fuse) create bijective maps from a collection of source iterators to target iterators. DetectIterMap is a function that detects such bijective mappings from the lowered index expression. We choose the term quasi affine to be consistent with the terminology used by in polyhedral compilation. DetectIterMap can handle symbolic integers(in split/fuse) to some extent. The utility can be useful in detecting loop transformation patterns and data layout change patterns in TIR. * Update per feedback

view details

Tianqi Chen

commit sha 8439f3e9221f10534a4c9fa0a9a6777efe3b5d0c

[REFACTOR] util => utils for consistency in the project. (#6684) * [REFACTOR] util => utils for consistency in the project. * Update CMake

view details

Josh Fromm

commit sha 9564925ae22aab8392a23cd9afa5843d6c96bb48

[Relay][Frontend][Onnx] Allow A to B broadcasting of batch_matmul and reverse strided slice (#6681) * slice and batch_matmul fixes. * Bug fix in shape inference. * Test backwards strided slice. * Fix batch_matmul dynamic shape function. * formatting. * Fix edge case for implicit broadcast

view details

Andrew Reusch

commit sha c7ff88516f99ea9ee275e9975cbd30e8979054e9

Add µTVM Zephyr support + QEMU regression test (#6603) * Split transport classes into transport package. * Introduce transport timeouts. * black format * Add metadata-only artifacts * Simplify utvm rpc server API and ease handling of short packets. * add zephyr test against qemu * Add qemu build config * fix typo * cleanup zephyr main * fix nonblocking piping on some linux kernels * don't double-open transport * validate FD are in non-blocking mode * gitignore test debug files * cleanup zephyr compiler * re-comment serial until added * remove logging * add zephyr exclusions to check_file_type * add asf header * lint * black format * more pylint * kill utvm rpc_server bindings, which don't work anymore and fail pylint * fix compiler warning * fixes related to pylint * clang-format again * more black format * add qemu regression * Fix paths for qemu/ dir * fix typo * fix SETFL logic * export SessionTerminatedError and update except after moving * fix test_micro_artifact * retrigger staging CI * fix jenkins syntax hopefully * one last syntax error * Add ci_qemu to Jenkinsfile * build in qemu * address liangfu comments * fix new bug with list passing * retrigger CI

view details

Matthew Brookhart

commit sha b121278a451243630fb67bf978079a8930de6558

int32 pooling with int64 shapes (#6687) * Failing tests for Int32 avg_pooling with Int64 shapes * fix pooling implementations

view details

Jared Roesch

commit sha 3e8ba2a0c5f2cad9cd757476d7eec9b19602881e

[Docker] Update CI CPU and GPU images based on new Docker build files. (#6690) * Turn on Rust docs and MxNet based ResNet * Add deps needed for Rust examples and docs * Setup Rust path * Bump Jenkinsfile * Fix broken version setting, which instead redirects stdout and stderr * Update Jenkinsfile * Format * Disable Rust change for now * Completely disable ResNet * Reset test changes * Remove temporary labels * Remove patch needed for docs

view details

Jared Roesch

commit sha 78f9b98ebc3f9b7dcb734bc3ef0ba09b34a1cecc

Enable ResNet and Rust docs

view details

push time in 9 days

PR opened apache/incubator-tvm

[Docker] Fix tutorial broken by Docker build

cc @tqchen @tmoreau89 this change is required for CI to pass on the new images.

+5 -1

0 comment

1 changed file

pr created time in 9 days

push eventjroesch/tvm

Lianmin Zheng

commit sha 78bba3c74785433b0f29589d89112ef06e4bca75

[AutoScheduler] Fix a bug in thread binding (#6683) * fix for lstm use case * update

view details

Tianqi Chen

commit sha 60ed9261058c0f1faa2632cb97319f29b26b9573

[ARITH] Introduce iterator (quasi)affine map detection. (#6667) * [ARITH] Introduce iterator (quasi)affine map detection. The loop transformations (split, fuse) create bijective maps from a collection of source iterators to target iterators. DetectIterMap is a function that detects such bijective mappings from the lowered index expression. We choose the term quasi affine to be consistent with the terminology used by in polyhedral compilation. DetectIterMap can handle symbolic integers(in split/fuse) to some extent. The utility can be useful in detecting loop transformation patterns and data layout change patterns in TIR. * Update per feedback

view details

Tianqi Chen

commit sha 8439f3e9221f10534a4c9fa0a9a6777efe3b5d0c

[REFACTOR] util => utils for consistency in the project. (#6684) * [REFACTOR] util => utils for consistency in the project. * Update CMake

view details

Josh Fromm

commit sha 9564925ae22aab8392a23cd9afa5843d6c96bb48

[Relay][Frontend][Onnx] Allow A to B broadcasting of batch_matmul and reverse strided slice (#6681) * slice and batch_matmul fixes. * Bug fix in shape inference. * Test backwards strided slice. * Fix batch_matmul dynamic shape function. * formatting. * Fix edge case for implicit broadcast

view details

Andrew Reusch

commit sha c7ff88516f99ea9ee275e9975cbd30e8979054e9

Add µTVM Zephyr support + QEMU regression test (#6603) * Split transport classes into transport package. * Introduce transport timeouts. * black format * Add metadata-only artifacts * Simplify utvm rpc server API and ease handling of short packets. * add zephyr test against qemu * Add qemu build config * fix typo * cleanup zephyr main * fix nonblocking piping on some linux kernels * don't double-open transport * validate FD are in non-blocking mode * gitignore test debug files * cleanup zephyr compiler * re-comment serial until added * remove logging * add zephyr exclusions to check_file_type * add asf header * lint * black format * more pylint * kill utvm rpc_server bindings, which don't work anymore and fail pylint * fix compiler warning * fixes related to pylint * clang-format again * more black format * add qemu regression * Fix paths for qemu/ dir * fix typo * fix SETFL logic * export SessionTerminatedError and update except after moving * fix test_micro_artifact * retrigger staging CI * fix jenkins syntax hopefully * one last syntax error * Add ci_qemu to Jenkinsfile * build in qemu * address liangfu comments * fix new bug with list passing * retrigger CI

view details

Matthew Brookhart

commit sha b121278a451243630fb67bf978079a8930de6558

int32 pooling with int64 shapes (#6687) * Failing tests for Int32 avg_pooling with Int64 shapes * fix pooling implementations

view details

Jared Roesch

commit sha 3e8ba2a0c5f2cad9cd757476d7eec9b19602881e

[Docker] Update CI CPU and GPU images based on new Docker build files. (#6690) * Turn on Rust docs and MxNet based ResNet * Add deps needed for Rust examples and docs * Setup Rust path * Bump Jenkinsfile * Fix broken version setting, which instead redirects stdout and stderr * Update Jenkinsfile * Format * Disable Rust change for now * Completely disable ResNet * Reset test changes * Remove temporary labels * Remove patch needed for docs

view details

Jared Roesch

commit sha cf24ad59bf66b753d7bdc372137124a020bf6813

Fix tutorial broken by Docker build

view details

push time in 9 days

push eventapache/incubator-tvm

Jared Roesch

commit sha 3e8ba2a0c5f2cad9cd757476d7eec9b19602881e

[Docker] Update CI CPU and GPU images based on new Docker build files. (#6690) * Turn on Rust docs and MxNet based ResNet * Add deps needed for Rust examples and docs * Setup Rust path * Bump Jenkinsfile * Fix broken version setting, which instead redirects stdout and stderr * Update Jenkinsfile * Format * Disable Rust change for now * Completely disable ResNet * Reset test changes * Remove temporary labels * Remove patch needed for docs

view details

push time in 9 days

push eventjroesch/tvm

Jared Roesch

commit sha dbe7991c696118ecffe996cf536623936bd93e32

Fix Linux build

view details

Jared Roesch

commit sha af055cbbd0dd87ca889dbe3574c1cb6d20326394

Clean up exporting to show off new diagnostics

view details

push time in 9 days

Pull request review commentapache/incubator-tvm

Refactor diagnostic to avoid circular dependencies

+/*+ * Licensed to the Apache Software Foundation (ASF) under one+ * or more contributor license agreements.  See the NOTICE file+ * distributed with this work for additional information+ * regarding copyright ownership.  The ASF licenses this file+ * to you under the Apache License, Version 2.0 (the+ * "License"); you may not use this file except in compliance+ * with the License.  You may obtain a copy of the License at+ *+ *   http://www.apache.org/licenses/LICENSE-2.0+ *+ * Unless required by applicable law or agreed to in writing,+ * software distributed under the License is distributed on an+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY+ * KIND, either express or implied.  See the License for the+ * specific language governing permissions and limitations+ * under the License.+ */++/*!+ * \file diagnostic.h+ * \brief A new diagnostic interface for TVM error reporting.+ *+ * A prototype of the new diagnostic reporting interface for TVM.+ *+ * Eventually we hope to promote this file to the top-level and+ * replace the existing errors.h.+ */++#ifndef TVM_IR_DIAGNOSTIC_CONTEXT_H_+#define TVM_IR_DIAGNOSTIC_CONTEXT_H_++#include <tvm/ir/module.h>+#include <tvm/parser/source_map.h>++#include <sstream>+#include <string>++namespace tvm {++using tvm::parser::SourceMap;+using tvm::runtime::TypedPackedFunc;++extern const char* kTVM_INTERNAL_ERROR_MESSAGE;++class DiagnosticBuilder;++/*! \brief A compiler diagnostic. */+class Diagnostic;++/*! \brief A compiler diagnostic message. */+class DiagnosticNode : public Object {+ public:+  /*! \brief The level. */

We should improve this comment.

rkimball

comment created time in 9 days

PullRequestReviewEvent

push eventjroesch/tvm

Jared Roesch

commit sha 5f8d51015ba95e085fac1fa0c4558fb76dbe480a

Remove patch needed for docs

view details

push time in 9 days

pull request commentapache/incubator-tvm

[Docker] Update CI CPU and GPU images based on new Docker build files.

cc @tmoreau89 @tqchen @u99127 this should include all the newest changes.

jroesch

comment created time in 10 days

push eventjroesch/tvm

Jared Roesch

commit sha c5a790241307e2e186301f9b7fb7617ed0e47fb5

Remove temporary labels

view details

push time in 10 days

create barnchjroesch/tvm

branch : update-docker-cpu-gpu

created branch time in 10 days

push eventapache/incubator-tvm

Jared Roesch

commit sha 5691c3c0e00eb6d0e4e4290cfb3e6b52a3ab106e

Completely disable ResNet

view details

push time in 10 days

push eventjroesch/tvm

Jared Roesch

commit sha f26989525ea84673bad72271e1cf4f5706146511

Update CMake and delete old API

view details

push time in 10 days

push eventjroesch/tvm

Jared Roesch

commit sha f4993824e25957555187d5712df26f32215879e5

Borrow code from Egg

view details

push time in 10 days

push eventapache/incubator-tvm

Jared Roesch

commit sha d4e3f4d043aa35dcca355ce171a29c1c33e76d08

Disable Rust change for now

view details

push time in 10 days

push eventapache/incubator-tvm

Jared Roesch

commit sha ecf8b656abcfb0d73903aee9a0ee3dea039c0ff8

Format

view details

push time in 10 days

push eventapache/incubator-tvm

Tristan Konolige

commit sha 6d0351a7f0e23eb5428c59a976edd2bfb8207c0d

Faster sparse_dense on GPUs (#6580) * Faster sparse_dense on GPUs. This new sparse_dense requires a padded matrix, so a new op `sparse_dense_padded` has been added. AlterOpLayout should transform `sparse_dense` to `sparse_dense_padded` when possible on the gpu. * formatting * more formatting * Check that alteroplayout is definedbefore using it * check if FTVMAlterOpLayout exists before using it * formatting * restore message passing * Fix sparse_dense and sparse_dense_padded docs * Fix old sparse_dense, autotvm and sparse_dense dont play well together * Remove unused imports * clarify warp count in cuda_transpose * Document multidimensional access * Warn users not to use sparse_dense_padded * rename nn.sparse_dense_padded to nn.internal.sparse_dense_padded

view details

masahi

commit sha 5a3377460b75336c937ff1cbc9fac3ff8b69af31

Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" (#6658) * Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" This reverts commit 782190e88b1941fdbe31101af260bee06b81bf72. * run infer type on test_sparse_dense_padded_alter_op() to fix CI Co-authored-by: masa <masa@pop-os.localdomain>

view details

Lianmin Zheng

commit sha f6657a6b87104eddd1fc46e5ddc490837fadc1b5

[AutoScheduler] Improve the GPU tutorial by deleting measure_ctx earlier (#6660) * del measurement process in the tutorial * fix * trigger CI

view details

Lianmin Zheng

commit sha dd60d249e50f29c3cb34704693ba54803ae75a36

[AutoScheduler] Improve test cases (#6657) * Improve test cases * update * fix lint * fix lint * trigger CI * address comments * trigger CI

view details

Jorn Tuyls

commit sha bf21371363fff978d35a27bc40d4f4fecf734aa2

[Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing (#6589) * Fix conv2d_transpose parsing in Tensorflow frontend for TF 1.15 * Add comments and convolution tests without AddShapesToGraphDef

view details

Dmitriy Smirnov

commit sha e561007f0c330e3d14c2bc8a3ef40fb741db9004

[BYOC][ACL] Support add operation (#6532) * [BYOC][ACL] Support add operation Added support for an "add" operation implemented via ACL for fp32 and quantized uint8 data types * Addressed lhutton1 comments * linter

view details

Chris Gyurgyik

commit sha d3ef1373dad8414b5825baa8012123d987e686d4

Fix typographical error. (#6664)

view details

Trevor Morris

commit sha 74b692269809a54b9f5a8160d90e401b9a9fc37d

[Relay][MXNet] Support broadcast_like (#6561)

view details

Tianqi Chen

commit sha 0cdd285abe58ac373c60c7544178888a47902d0d

[CI] Move to use main as the default (#6665)

view details

masahi

commit sha b277f181c4430ac705a533b016ac35fbf98ec2b8

[Torch] Object detection support update for PyTorch 1.6 (#6659) * update split * fix * cast nms output to int64 * add more comment and numel test * fix lint * also supported the latest master (1.7) Co-authored-by: masa <masa@pop-os.localdomain>

view details

Andrew Reusch

commit sha d955718cfe300d3ed294cacb0c74f8ea73ae6ad0

don't validate AttrInitEntry until a value has attempted to be set (#6672)

view details

Zhiqiang Wang

commit sha 4073adc15fc55d6232d00ff07dd1384bd9f936ff

[CI] Set main as default in github actions (#6669)

view details

Ritwik Das

commit sha d5728bdeb8e4eb3e633c608366a0c5795a236d93

[BYOC] Support control flow in annotate_target (#6641) * Change annotate target * Annotate_target * Revert namespace changes * Add tests for if-else node * Add while_let testcase * No merging in ifelse * Remove scope builder * Add ops * Replace < with less * Linter * Pass Tests * Change back to static const * Cpplinter * address PR comments' * PR Comments * Clang-format check * PR Comments * PR Comments * Change back to Insert Ann in AnnotateARgs Co-authored-by: Ritwik Das <dasritwi@3c22fb14d7c6.ant.amazon.com> Co-authored-by: Ubuntu <ubuntu@ip-172-31-3-223.us-west-2.compute.internal>

view details

Animesh Jain

commit sha d24634a7cc3c502176838d238af42cf7c94defbe

TF argmax - handling int64 datatype (#6674) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-202.us-west-2.compute.internal>

view details

MORITA Kazutaka

commit sha f196a81e73322776ca1c70d413cc266bacaa42cc

[CODEGEN][COREML] Call InferType explicitly in coreml test (#6676)

view details

Ishi Tatsuyuki

commit sha 8d9ca2ab4ce0be2ecf5fcfa5ea2b8b63a6de7d23

Adjust Vulkan queue selection and creation logic (#6662) * Adjust Vulkan queue selection and creation logic - The queue selection logic rewrite addresses a bug in the old implementation where the code would pass an invalid queue index when selecting any queues other than number 0. - The new implementation will attempt to use compute-only queues which are common on AMD GPUs. It's not clear how much difference will this make but hopefully it would lead to better scheduling. The queue changes were made as with the old configuration autotvm caused my system to hang, stutter or otherwise become unstable and crash. With the change I'm able to run autotvm tuning inside a desktop environment. * Return multiple queue family indexes from GetComputeQueues * Only create one queue

view details

Trevor Morris

commit sha d7fa9c1f2539e1be8527a88745fcd0bed9be2f0f

Install xgboost>=1.1.0 in CI container (#6679)

view details

qixiuai

commit sha 7d805b54d6adda82636d13bf7c46a2e9a933da5f

Fix format error in integrate.rst (#6677)

view details

Przemyslaw Tredak

commit sha e66b2e83db24c78c892308475687ff3bbb0173b9

Revert #5238 (#6680)

view details

Jared Roesch

commit sha b06fa681eab41c92c8385067deee94e54b4e7d1b

Turn on Rust docs and MxNet based ResNet

view details

push time in 11 days

push eventjroesch/tvm

Jared Roesch

commit sha 7e264b80afcc5a879c145535c5c3a16b5bcd3956

Fix broken version setting, which instead redirects stdout and stderr

view details

Jared Roesch

commit sha 9ddcff91a0915a5b843e5fac05d5a77d7c2f8e3b

Update Jenkinsfile

view details

push time in 11 days

push eventjroesch/tvm

Tristan Konolige

commit sha 6d0351a7f0e23eb5428c59a976edd2bfb8207c0d

Faster sparse_dense on GPUs (#6580) * Faster sparse_dense on GPUs. This new sparse_dense requires a padded matrix, so a new op `sparse_dense_padded` has been added. AlterOpLayout should transform `sparse_dense` to `sparse_dense_padded` when possible on the gpu. * formatting * more formatting * Check that alteroplayout is definedbefore using it * check if FTVMAlterOpLayout exists before using it * formatting * restore message passing * Fix sparse_dense and sparse_dense_padded docs * Fix old sparse_dense, autotvm and sparse_dense dont play well together * Remove unused imports * clarify warp count in cuda_transpose * Document multidimensional access * Warn users not to use sparse_dense_padded * rename nn.sparse_dense_padded to nn.internal.sparse_dense_padded

view details

masahi

commit sha 5a3377460b75336c937ff1cbc9fac3ff8b69af31

Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" (#6658) * Revert "[Relay] Keep fixed dim when unifying dynamic shape (#5795)" This reverts commit 782190e88b1941fdbe31101af260bee06b81bf72. * run infer type on test_sparse_dense_padded_alter_op() to fix CI Co-authored-by: masa <masa@pop-os.localdomain>

view details

Lianmin Zheng

commit sha f6657a6b87104eddd1fc46e5ddc490837fadc1b5

[AutoScheduler] Improve the GPU tutorial by deleting measure_ctx earlier (#6660) * del measurement process in the tutorial * fix * trigger CI

view details

Lianmin Zheng

commit sha dd60d249e50f29c3cb34704693ba54803ae75a36

[AutoScheduler] Improve test cases (#6657) * Improve test cases * update * fix lint * fix lint * trigger CI * address comments * trigger CI

view details

Jorn Tuyls

commit sha bf21371363fff978d35a27bc40d4f4fecf734aa2

[Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing (#6589) * Fix conv2d_transpose parsing in Tensorflow frontend for TF 1.15 * Add comments and convolution tests without AddShapesToGraphDef

view details

Dmitriy Smirnov

commit sha e561007f0c330e3d14c2bc8a3ef40fb741db9004

[BYOC][ACL] Support add operation (#6532) * [BYOC][ACL] Support add operation Added support for an "add" operation implemented via ACL for fp32 and quantized uint8 data types * Addressed lhutton1 comments * linter

view details

Chris Gyurgyik

commit sha d3ef1373dad8414b5825baa8012123d987e686d4

Fix typographical error. (#6664)

view details

Trevor Morris

commit sha 74b692269809a54b9f5a8160d90e401b9a9fc37d

[Relay][MXNet] Support broadcast_like (#6561)

view details

Tianqi Chen

commit sha 0cdd285abe58ac373c60c7544178888a47902d0d

[CI] Move to use main as the default (#6665)

view details

masahi

commit sha b277f181c4430ac705a533b016ac35fbf98ec2b8

[Torch] Object detection support update for PyTorch 1.6 (#6659) * update split * fix * cast nms output to int64 * add more comment and numel test * fix lint * also supported the latest master (1.7) Co-authored-by: masa <masa@pop-os.localdomain>

view details

Andrew Reusch

commit sha d955718cfe300d3ed294cacb0c74f8ea73ae6ad0

don't validate AttrInitEntry until a value has attempted to be set (#6672)

view details

Zhiqiang Wang

commit sha 4073adc15fc55d6232d00ff07dd1384bd9f936ff

[CI] Set main as default in github actions (#6669)

view details

Ritwik Das

commit sha d5728bdeb8e4eb3e633c608366a0c5795a236d93

[BYOC] Support control flow in annotate_target (#6641) * Change annotate target * Annotate_target * Revert namespace changes * Add tests for if-else node * Add while_let testcase * No merging in ifelse * Remove scope builder * Add ops * Replace < with less * Linter * Pass Tests * Change back to static const * Cpplinter * address PR comments' * PR Comments * Clang-format check * PR Comments * PR Comments * Change back to Insert Ann in AnnotateARgs Co-authored-by: Ritwik Das <dasritwi@3c22fb14d7c6.ant.amazon.com> Co-authored-by: Ubuntu <ubuntu@ip-172-31-3-223.us-west-2.compute.internal>

view details

Animesh Jain

commit sha d24634a7cc3c502176838d238af42cf7c94defbe

TF argmax - handling int64 datatype (#6674) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-202.us-west-2.compute.internal>

view details

MORITA Kazutaka

commit sha f196a81e73322776ca1c70d413cc266bacaa42cc

[CODEGEN][COREML] Call InferType explicitly in coreml test (#6676)

view details

Ishi Tatsuyuki

commit sha 8d9ca2ab4ce0be2ecf5fcfa5ea2b8b63a6de7d23

Adjust Vulkan queue selection and creation logic (#6662) * Adjust Vulkan queue selection and creation logic - The queue selection logic rewrite addresses a bug in the old implementation where the code would pass an invalid queue index when selecting any queues other than number 0. - The new implementation will attempt to use compute-only queues which are common on AMD GPUs. It's not clear how much difference will this make but hopefully it would lead to better scheduling. The queue changes were made as with the old configuration autotvm caused my system to hang, stutter or otherwise become unstable and crash. With the change I'm able to run autotvm tuning inside a desktop environment. * Return multiple queue family indexes from GetComputeQueues * Only create one queue

view details

Trevor Morris

commit sha d7fa9c1f2539e1be8527a88745fcd0bed9be2f0f

Install xgboost>=1.1.0 in CI container (#6679)

view details

qixiuai

commit sha 7d805b54d6adda82636d13bf7c46a2e9a933da5f

Fix format error in integrate.rst (#6677)

view details

Przemyslaw Tredak

commit sha e66b2e83db24c78c892308475687ff3bbb0173b9

Revert #5238 (#6680)

view details

Jared Roesch

commit sha b06fa681eab41c92c8385067deee94e54b4e7d1b

Turn on Rust docs and MxNet based ResNet

view details

push time in 11 days

push eventjroesch/tvm

Jared Roesch

commit sha 56790de4f6ff5ef4c78374f43f231a0e7dc5cd58

Hacking on Rust inside of TVM

view details

push time in 11 days

create barnchjroesch/tvm

branch : rust-tyck

created branch time in 11 days

create barnchapache/incubator-tvm

branch : master

created branch time in 12 days

push eventapache/incubator-tvm

Andrew Reusch

commit sha 1c85047ce0d82a5bb421da57b3ba49454fc97dda

[apps/bundle_deploy] Link demo_* targets with LDFLAGS and also with -lm. (#6636)

view details

Andrew Reusch

commit sha e4b47a6637ce545a2218124ec5dbeae95657ed4b

Add qemu build step to CI (#6644)

view details

Matthew Brookhart

commit sha 455d4013499b504ea0c0a991540f6eba084e507b

add a test for assymetric padding in ONNX conv and fix importer (#6646)

view details

Qiang Xu

commit sha 35854c18b04cb6bee092fdc2aa8be5082ae0dd48

missing header for GraphRuntimeFactory in android_rpc (#6648)

view details

Ishi Tatsuyuki

commit sha dffdb23b64f81df758a2fd86107f18d8f63e35bf

Fix leakyReLU support for CoreML (#6651) The original implementation failed with the following error: File "../include/tvm/runtime/packed_func.h", line 372 TVMError: Check failed: type_code_ == kDLFloat (8 vs. 2) : expected float but get Object

view details

Leandro Nunes

commit sha a8c1273ea4382f68b7c930d012dfd283f2b50fcc

[CI] make sure graphviz is on both ci-cpu and ci-gpu images (#6645)

view details

Matthew Brookhart

commit sha a919824d13616afb0e698a440504a7e2b9cf2c55

Add Range op to ONNX, make tvm arange op support negative steps (#6647)

view details

Andrew Reusch

commit sha 1fc47cca27074e1aa881b671b633c1e7549a8473

[µTVM] Avoid use of builtin math functions (#6630)

view details

Tristan Konolige

commit sha 0922d170dd37b8da0ece2bbac26b3a52aaa4c68d

[FIX,AUTOTVM] More descriptive error message when an autotvm task is not (#6652) found.

view details

Leandro Nunes

commit sha f73a1f641680e61d2e4a0a081a98f01b7fca9689

[TEST][TEDD] improve TEDD tests to also run on CPU Docker image. (#6643) * Amend regular expressions to match with what is being reported by CPU Docker image Graphviz * Fix typo on dependency checking function * Organise imports

view details

Jared Roesch

commit sha 98c2096f4944bdbdbbb2b7b20ccd35c6c11dfbf6

[Diagnostics][Relay][InferType] Refactor InferType to work on whole module, and use new diagnostics. (#6274) * Refactor the type checker to use diagnostics Although this patch is very large and seemingly disjoint the fixes are required to get it working for the entire stack. I started with first changing InferType to use the diagnostics, these weren't yet in the pass manager so this required changes to module and module pass. InferType wasn't actually written correctly as a pass requring refactoring there, then in order to add spans to AST it required turning on AnnotateSpans which in term required changes to the parser, and module to make it possible to use the errors. These changes to parse and module required changes to diagnostics and InferType. Althought seemingly disconnected there are hidden cycles between the components which require simultaneous change in order to remove the old error reporting. A huge change due to this patch is that the module no longer implicitly type checks functions which are added. * Apply suggestions from code review Co-authored-by: Robert Kimball <bobkimball@gmail.com> Co-authored-by: Junru Shao <junrushao1994@gmail.com> * Apply suggestions from code review Co-authored-by: Tristan Konolige <tristan.konolige@gmail.com> * Clean up parser * CR feedback * Apply Bobs suggestions * Fix up Python interface for diagnostics * Fix test_ir_parser and formatting * Fix cpplint * Fix lint * Fix format * More lint * Fix format * Kill dead doc comment * Fix documentation comment * Rebase fixups * Add docs for type.h * Fix parser.cc * Fix unittests * Fix black * Skip previously typechecked functions * fix ACL * Fix numerous issues * Add repr method * Fix issue with Pytest, I am ready to cry * Fix the rest of tests * Kill dead code * Fix dignostic tests * Fix more tests * fix more tests (#11) * Fix diagnostic.py deinit bug * Fix deinit issue * Format * Tweak disabling of override * Format * Fix BYOC * Fix TensorArray stuff * Fix PyTorch * Format * Format Co-authored-by: Robert Kimball <bobkimball@gmail.com> Co-authored-by: Junru Shao <junrushao1994@gmail.com> Co-authored-by: Tristan Konolige <tristan.konolige@gmail.com> Co-authored-by: Cody Yu <comaniac0422@gmail.com> Co-authored-by: Zhi <5145158+zhiics@users.noreply.github.com>

view details

Jared Roesch

commit sha a04e8f6754e5054caed30b60889226cccde2044a

Turn on Rust docs and MxNet based ResNet

view details

Jared Roesch

commit sha 78585031353781f0a3717115b096fb5f17b5f66f

Add deps needed for Rust examples and docs

view details

Jared Roesch

commit sha 07047562452e66c84209259d2037666d516243da

Setup Rust path

view details

Jared Roesch

commit sha 074dbefd48a5db235dbfd987c29aad75c7c3c215

Bump Jenkinsfile

view details

push time in 15 days

push eventjroesch/tvm

Andrew Reusch

commit sha 1c85047ce0d82a5bb421da57b3ba49454fc97dda

[apps/bundle_deploy] Link demo_* targets with LDFLAGS and also with -lm. (#6636)

view details

Andrew Reusch

commit sha e4b47a6637ce545a2218124ec5dbeae95657ed4b

Add qemu build step to CI (#6644)

view details

Matthew Brookhart

commit sha 455d4013499b504ea0c0a991540f6eba084e507b

add a test for assymetric padding in ONNX conv and fix importer (#6646)

view details

Qiang Xu

commit sha 35854c18b04cb6bee092fdc2aa8be5082ae0dd48

missing header for GraphRuntimeFactory in android_rpc (#6648)

view details

Ishi Tatsuyuki

commit sha dffdb23b64f81df758a2fd86107f18d8f63e35bf

Fix leakyReLU support for CoreML (#6651) The original implementation failed with the following error: File "../include/tvm/runtime/packed_func.h", line 372 TVMError: Check failed: type_code_ == kDLFloat (8 vs. 2) : expected float but get Object

view details

Leandro Nunes

commit sha a8c1273ea4382f68b7c930d012dfd283f2b50fcc

[CI] make sure graphviz is on both ci-cpu and ci-gpu images (#6645)

view details

Matthew Brookhart

commit sha a919824d13616afb0e698a440504a7e2b9cf2c55

Add Range op to ONNX, make tvm arange op support negative steps (#6647)

view details

Andrew Reusch

commit sha 1fc47cca27074e1aa881b671b633c1e7549a8473

[µTVM] Avoid use of builtin math functions (#6630)

view details

Tristan Konolige

commit sha 0922d170dd37b8da0ece2bbac26b3a52aaa4c68d

[FIX,AUTOTVM] More descriptive error message when an autotvm task is not (#6652) found.

view details

Leandro Nunes

commit sha f73a1f641680e61d2e4a0a081a98f01b7fca9689

[TEST][TEDD] improve TEDD tests to also run on CPU Docker image. (#6643) * Amend regular expressions to match with what is being reported by CPU Docker image Graphviz * Fix typo on dependency checking function * Organise imports

view details

Jared Roesch

commit sha 98c2096f4944bdbdbbb2b7b20ccd35c6c11dfbf6

[Diagnostics][Relay][InferType] Refactor InferType to work on whole module, and use new diagnostics. (#6274) * Refactor the type checker to use diagnostics Although this patch is very large and seemingly disjoint the fixes are required to get it working for the entire stack. I started with first changing InferType to use the diagnostics, these weren't yet in the pass manager so this required changes to module and module pass. InferType wasn't actually written correctly as a pass requring refactoring there, then in order to add spans to AST it required turning on AnnotateSpans which in term required changes to the parser, and module to make it possible to use the errors. These changes to parse and module required changes to diagnostics and InferType. Althought seemingly disconnected there are hidden cycles between the components which require simultaneous change in order to remove the old error reporting. A huge change due to this patch is that the module no longer implicitly type checks functions which are added. * Apply suggestions from code review Co-authored-by: Robert Kimball <bobkimball@gmail.com> Co-authored-by: Junru Shao <junrushao1994@gmail.com> * Apply suggestions from code review Co-authored-by: Tristan Konolige <tristan.konolige@gmail.com> * Clean up parser * CR feedback * Apply Bobs suggestions * Fix up Python interface for diagnostics * Fix test_ir_parser and formatting * Fix cpplint * Fix lint * Fix format * More lint * Fix format * Kill dead doc comment * Fix documentation comment * Rebase fixups * Add docs for type.h * Fix parser.cc * Fix unittests * Fix black * Skip previously typechecked functions * fix ACL * Fix numerous issues * Add repr method * Fix issue with Pytest, I am ready to cry * Fix the rest of tests * Kill dead code * Fix dignostic tests * Fix more tests * fix more tests (#11) * Fix diagnostic.py deinit bug * Fix deinit issue * Format * Tweak disabling of override * Format * Fix BYOC * Fix TensorArray stuff * Fix PyTorch * Format * Format Co-authored-by: Robert Kimball <bobkimball@gmail.com> Co-authored-by: Junru Shao <junrushao1994@gmail.com> Co-authored-by: Tristan Konolige <tristan.konolige@gmail.com> Co-authored-by: Cody Yu <comaniac0422@gmail.com> Co-authored-by: Zhi <5145158+zhiics@users.noreply.github.com>

view details

Jared Roesch

commit sha a04e8f6754e5054caed30b60889226cccde2044a

Turn on Rust docs and MxNet based ResNet

view details

Jared Roesch

commit sha 78585031353781f0a3717115b096fb5f17b5f66f

Add deps needed for Rust examples and docs

view details

Jared Roesch

commit sha 07047562452e66c84209259d2037666d516243da

Setup Rust path

view details

Jared Roesch

commit sha 074dbefd48a5db235dbfd987c29aad75c7c3c215

Bump Jenkinsfile

view details

push time in 15 days

push eventapache/incubator-tvm

Tristan Konolige

commit sha 6d0351a7f0e23eb5428c59a976edd2bfb8207c0d

Faster sparse_dense on GPUs (#6580) * Faster sparse_dense on GPUs. This new sparse_dense requires a padded matrix, so a new op `sparse_dense_padded` has been added. AlterOpLayout should transform `sparse_dense` to `sparse_dense_padded` when possible on the gpu. * formatting * more formatting * Check that alteroplayout is definedbefore using it * check if FTVMAlterOpLayout exists before using it * formatting * restore message passing * Fix sparse_dense and sparse_dense_padded docs * Fix old sparse_dense, autotvm and sparse_dense dont play well together * Remove unused imports * clarify warp count in cuda_transpose * Document multidimensional access * Warn users not to use sparse_dense_padded * rename nn.sparse_dense_padded to nn.internal.sparse_dense_padded

view details

push time in 15 days

PR merged apache/incubator-tvm

Faster sparse_dense on GPUs

I've written a faster sparse_dense for GPUs using tir. This sparse_dense requires a padded matrix, so I've added a new op sparse_dense_padded. AlterOpLayout should transform sparse_dense to sparse_dense_padded when using a gpu.

This new sparse_dense improves prunebert performance from 155.41ms mean to 7.75ms mean. In general, this implementation is faster than cublas dense on matrices with density < 0.05 and is often faster than cusparse for machine learning workloads.

+537 -60

5 comments

16 changed files

tkonolige

pr closed time in 15 days

PullRequestReviewEvent

PR opened jwfromm/tvm

[Typeck] Fix typechecking by introducing let binding.

Fix

+13 -10

0 comment

2 changed files

pr created time in 15 days

create barnchjroesch/tvm

branch : onnx_loop

created branch time in 15 days

create barnchjroesch/tvm

branch : rust-diag

created branch time in 15 days

PR opened apache/incubator-tvm

[Rust][Diagnostics][WIP] Add initial boilerplate for Rust diagnostic interface.

This PR exposes new diagnostics to Rust so we can reuse the better libraries for diagnostics rendering via Rust.

+241 -1

1 comment

3 changed files

pr created time in 15 days

pull request commentapache/incubator-tvm

[Rust][Diagnostics][WIP] Add initial boilerplate for Rust diagnostic interface.

cc @gussmith23 and @mwillsey

jroesch

comment created time in 15 days

push eventjroesch/tvm

Tristan Konolige

commit sha 86122d125d9fc8d08003bf6a3fffeeca490dc634

[FIX,AUTOTVM] Print warning when all autotvm tasks fail with errors (#6612) * [FIX,AUTOTVM] Print warning when all autotvm tasks fail with errors. * formatting * write errors to tempfile * wording * wording * don't duplicate errors * Ensure we have a string for an error

view details

Matthew Brookhart

commit sha 311eca49a696f137e8cac6f4b9ba485b80bda0ee

Fix a bug with Alter Op Layout (#6626) * Regression test for a Scalar type issue in Alter Op Layout * fix the regression test by avoiding the Scalar optimization if types aren't defined

view details

Thierry Moreau

commit sha feb041d08c7658ee262430e6660435630e636f90

bump dockerfile (#6632)

view details

Max Willsey

commit sha 277bfc86aac94442625e8307aa55a9215e2fddd6

[Rust] Improve NDArray, GraphRt, and Relay bindings (#6563) * WIP WIP * Add support for loading Python packed functions * Flesh out Relay AST in Rust * More tweeks for getting functions out * Deploy Rust docs as part of build * Add some more types * Introduce NDArray 2.0 * Work on NDArray 2.0 before restoring tests * Formatting and code fixes to get it to compile * Add more Rust bindings - Converts Conv2d attrs to use tvm::String, so that we can add Rust binding - Uses Type for checked_type in Rust bindings - Fix type key in Rust bindings - Make data field contain NDArray in Rust bindings * Clean up object ptr passing. * WIP * Add debugging for NDArray and fix all test cases * Add breaking test * Dispatch some todos * Format * Fix ndarray size and len * Add BiasAddAttrs rust bindings * Add DenseAttrs rust bindings * Change to TVM string * Add more Rust bindings Add GlobalPool2DAttrs Rust binding Add ExpandDimsAttrs Rust bindings Add MaxPool2DAttrs rust bindings * Fix some test attributes * Improve the NDArray api * Fix some more ndarray stuff * Get the resnet demo kinda working * Add SoftmaxAttrs Rust bindings * Implement Hash and Eq for Relay Exprs * Add underscore to unused function * Fix broken ass resnet script * Improve some ndarray conversions * Make sure the build script runs correctly * Clean up ResNet example tremedously Expose C++ graph runtime via cleaner Rust API rewrite example. * Add ASF header * Format * Format * Format resnet rust python script * Add type files and refactor span * Format * Format * Change types from std::string to tvm::String in packed function * Add ASF header * Fix test w/ ndarray's API change * Fix array test * Fix anyhow import * Put back some anyhow stuff * Clean up * Try and fix tests/scripts/task_rust.sh * Disable ResNet for now * Turn off building of Rust docs until we update CI * Actually disable Co-authored-by: Jared Roesch <jroesch@octoml.ai> Co-authored-by: Gus Smith <guscomps@gmail.com>

view details

Robert Kimball

commit sha dec503029235b3cf8fc91fe6defafbea3307ab17

Fix example code (#6627)

view details

Leandro Nunes

commit sha 82289ea15b76e20184117a0da44d29d838e55be9

[TVMC] fail gracefully in case no subcommand is provided (#6625)

view details

Giuseppe Rossini

commit sha 8369adc9ce2d07f9369968eb49a391f78e5be4ca

Add dot product support for quantized convolution. (#6445) * Add dot product support for quantized convolution. We added two new intrinsics in: topi/arm_cpu/tensor_intrin.py, namely - mmla4x4: compute a matrix multiplication between tile A(4,4) and tile B(4,4) - mmla16x4: compute a matrix multiplication between tile A(rows,4) and tile B(4,16) Then we used those intrinsics in two separate strategies. We added the strategies in topi/arm_cpu/conv2d_int8.py and implemented the schedules in topi/arm_cpu/conv2d_gemm.py. In particular: - schedule_conv2d_gemm, when accelerated, packs matrix A, compute GEMM, and unpack the resulting matrix. This uses the mmla4x4 intrinsic - schedule_conv2d_gemm_hybrid doesn't do any packing on A and C which are in native form. This uses the mmla16x4 intrinsic Please note that for the limitations of `tensorize` we need to pad matrix A in both cases (when dimensions are not multiple of the tiling shape) Change-Id: Id0d818d84ffc458c6dad7983fd350a0f3d5db395 * Add back nhwc_spatial_pack strategy as default Change-Id: I8b1826a7ae1d742956296e8d157da19955a4942c * Fix linting through Black Change-Id: Ic74ef5461a90bca9f4d4980a214137e384d5f923 * Fix python linting Change-Id: I5fb8a2ae4467a87bd3470f6b3753c074f9b7cc78 * Addressing review comments Change-Id: I284b1f2c121051e672f548d6c6ee2a3267854e31 * Fix black linting issues Change-Id: I1813b0226b536aedee0dce9eeeba27aa2d95518b * Fixing failing test and adding tests for dot-product compilation Change-Id: Ic040722abd5538fccb85af4de922394c939e7000 * Fixing linting and review comments Change-Id: If09e3baa514c85dc78d3c27c2ac2fa2e01773d89 * Fixing black linting and address comments Change-Id: I857b28b6f9b23307d8c1eebc509de6ad2783c756 * Address review comments Change-Id: I63d1a639d4a72abeb33148fd2868cd356ef84122

view details

Josh Fromm

commit sha 889fac157beb2a07830b32008980b902a7feee28

[Topi] Allow batch_matmul to broadcast along batch dimension. (#6616) * Allow batch_matmul to broadcast along batch dimension. * Added typerel checking. * Fix style issue and respond to feedback. * Fix style. * More formatting issues :( * Fix issues after merge. * Comment update. * Small tweak.

view details

Zhi

commit sha 35793ae63bc43fffb6f2ef7e33126053b384b436

[COMMUNITY] areusch -> Reviewer (#6637)

view details

Leandro Nunes

commit sha 0cfda17611052603151a69fe7187f801588373da

[CI] fix Python dependency required by cpp tests to work standalone (#6639)

view details

Andrew Reusch

commit sha 1c85047ce0d82a5bb421da57b3ba49454fc97dda

[apps/bundle_deploy] Link demo_* targets with LDFLAGS and also with -lm. (#6636)

view details

Andrew Reusch

commit sha e4b47a6637ce545a2218124ec5dbeae95657ed4b

Add qemu build step to CI (#6644)

view details

Matthew Brookhart

commit sha 455d4013499b504ea0c0a991540f6eba084e507b

add a test for assymetric padding in ONNX conv and fix importer (#6646)

view details

Qiang Xu

commit sha 35854c18b04cb6bee092fdc2aa8be5082ae0dd48

missing header for GraphRuntimeFactory in android_rpc (#6648)

view details

Ishi Tatsuyuki

commit sha dffdb23b64f81df758a2fd86107f18d8f63e35bf

Fix leakyReLU support for CoreML (#6651) The original implementation failed with the following error: File "../include/tvm/runtime/packed_func.h", line 372 TVMError: Check failed: type_code_ == kDLFloat (8 vs. 2) : expected float but get Object

view details

Leandro Nunes

commit sha a8c1273ea4382f68b7c930d012dfd283f2b50fcc

[CI] make sure graphviz is on both ci-cpu and ci-gpu images (#6645)

view details

Matthew Brookhart

commit sha a919824d13616afb0e698a440504a7e2b9cf2c55

Add Range op to ONNX, make tvm arange op support negative steps (#6647)

view details

Andrew Reusch

commit sha 1fc47cca27074e1aa881b671b633c1e7549a8473

[µTVM] Avoid use of builtin math functions (#6630)

view details

Tristan Konolige

commit sha 0922d170dd37b8da0ece2bbac26b3a52aaa4c68d

[FIX,AUTOTVM] More descriptive error message when an autotvm task is not (#6652) found.

view details

Jared Roesch

commit sha 6361efc97227caf151b81cb69bba7624eaf0e71b

Refactor the type checker to use diagnostics Although this patch is very large and seemingly disjoint the fixes are required to get it working for the entire stack. I started with first changing InferType to use the diagnostics, these weren't yet in the pass manager so this required changes to module and module pass. InferType wasn't actually written correctly as a pass requring refactoring there, then in order to add spans to AST it required turning on AnnotateSpans which in term required changes to the parser, and module to make it possible to use the errors. These changes to parse and module required changes to diagnostics and InferType. Althought seemingly disconnected there are hidden cycles between the components which require simultaneous change in order to remove the old error reporting. A huge change due to this patch is that the module no longer implicitly type checks functions which are added.

view details

push time in 16 days

push eventjroesch/tvm

Jared Roesch

commit sha c0094fd82da726bd1875eab2fd735cb3e953d9b4

Format

view details

push time in 16 days

push eventjroesch/tvm

Jared Roesch

commit sha 5e880547382608b2229f18dfb78faeba34c55304

Format

view details

push time in 16 days

push eventjroesch/tvm

Jared Roesch

commit sha ceb725ad702699ea2259af3b912d53ad68beb53a

Fix TensorArray stuff

view details

Jared Roesch

commit sha c5e3718ef9d1109ccc34070b94bd465ddada8642

Fix PyTorch

view details

push time in 16 days

push eventjroesch/tvm

Jared Roesch

commit sha cd3f2480773ff87dc2f25d06e49436b204c5ca71

Fix BYOC

view details

push time in 17 days

push eventjroesch/tvm

Jared Roesch

commit sha 0f398b9ce49b5b8e36babced8760c2b876ce433f

Format

view details

push time in 17 days

push eventjroesch/tvm

Jared Roesch

commit sha 1063cf8ea93da6aa16aa5ed131bd0e51f132877a

Tweak disabling of override

view details

push time in 17 days

push eventjroesch/tvm

Jared Roesch

commit sha f83ec042dfe73b0411cfa268b7ed69375991d962

Fix deinit issue

view details

Jared Roesch

commit sha a1e64508c84987b404f9349a821374c955e23355

Format

view details

push time in 17 days

push eventjroesch/tvm

Jared Roesch

commit sha 4538fdecc6f95ae7e934afb6e24d1f2dc410e426

Fix diagnostic.py deinit bug

view details

push time in 17 days

push eventjroesch/tvm

Zhi

commit sha 79c9a94e941dc56ac5ae84ee0e2d9c190183c3f4

fix more tests (#11)

view details

push time in 18 days

PR merged jroesch/tvm

fix more tests

some fixes

+27 -0

0 comment

2 changed files

zhiics

pr closed time in 18 days

push eventapache/incubator-tvm

Jared Roesch

commit sha 7836654152e819252dd505818dac32f7aa6bc7a6

Add deps needed for Rust examples and docs

view details

Jared Roesch

commit sha 4b7aa5d777741174811a8a1cdc52b055d8da70cf

Setup Rust path

view details

push time in 18 days

more