[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041172051 @vpirogov Thank you so much! I will update the CI to the latest version. -- This is an automated message from the Apache Git Service. To respond to the message, please log on

[GitHub] [tvm] vpirogov commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
vpirogov commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041170017 @crazydemo, right, we stopped publishing binary releases on Github after v2.2. Binaries for later versions are available on

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041163172 The PR, updating CI has been submitted [#10266](https://github.com/apache/tvm/pull/10266) @masahi -- This is an automated message from the Apache Git Service. To respond to

[GitHub] [tvm] crazydemo opened a new pull request #10266: [CI] update dnnl version from v1.5 to v2.2

2022-02-15 Thread GitBox
crazydemo opened a new pull request #10266: URL: https://github.com/apache/tvm/pull/10266 This PR aims to update the dnnl version for CI from v1.5 to v2.2, and thus we can support more dnnl patterns and post-ops. -- This is an automated message from the Apache Git Service. To

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041156478 > The latest oneDNN release is [v2.5.2](https://github.com/oneapi-src/oneDNN/releases/tag/v2.5.2) Yes, the latest release version is v2.5.2, but only v2.2 has

[GitHub] [tvm] vpirogov commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
vpirogov commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041154312 The latest oneDNN release is [v2.5.2](https://github.com/oneapi-src/oneDNN/releases/tag/v2.5.2) -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] masahi commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
masahi commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041151961 No need for another rfc. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041151095 > Note that changing the docker script won't immediately update the CI environment (so the build will fail again).You need to send that change separately first like #10179. I

[GitHub] [tvm] masahi closed pull request #10179: [LIBXSMM] Add libxsmm to tvm ci

2022-02-15 Thread GitBox
masahi closed pull request #10179: URL: https://github.com/apache/tvm/pull/10179 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] masahi edited a comment on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
masahi edited a comment on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041149531 Note that changing the docker script won't immediately update the CI environment (so the build will fail again).You need to send that change separately first like

[GitHub] [tvm] masahi commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
masahi commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041149531 Note that changing the docker script won't immediately update the CI environment.You need to send that change separately first like https://github.com/apache/tvm/pull/10179. I can

[GitHub] [tvm] elvin-n commented on pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
elvin-n commented on pull request #10239: URL: https://github.com/apache/tvm/pull/10239#issuecomment-1041148103 > That is understood, my main concern is a lot of code looks similar Ideologically transformations look similar, but it is unable to extend current transformation exactly

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041139323 I will try to update to `OneDNN V2.2`. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to

[GitHub] [tvm] masahi commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
masahi commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041135514 https://github.com/apache/tvm/blob/813136401a11a49d6c15e6013c34dd822a5c4ff6/docker/install/ubuntu_install_dnnl.sh#L24 Probably too old, we should update that first. --

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041131131 Hi @masahi , I wonder which version of OneDNN is used for CI. I have one build error, which says `'eltwise_logsigmoid' is not a member of 'dnnl::algorithm'`, and this algo is

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
AndrewZhaoLuo commented on pull request #10239: URL: https://github.com/apache/tvm/pull/10239#issuecomment-1041126506 That is understood, my main concern is a lot of code looks similar and I'm wondering if this "optional" pass can supersede the original pass. -- This is an automated

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041057518 > I can merge this after you pass CI. Thank you! I will check! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to

[GitHub] [tvm] masahi commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
masahi commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041051727 I can merge this after you pass CI. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to

[GitHub] [tvm] crazydemo commented on pull request #9995: [BYOC-DNNL] add support for more ops and fusion patterns

2022-02-15 Thread GitBox
crazydemo commented on pull request #9995: URL: https://github.com/apache/tvm/pull/9995#issuecomment-1041046078 @tmoreau89 @mbs-octoml @mikepapadim @masahi Could you please help me review this PR? -- This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [tvm] slyubomirsky opened a new issue #10265: [Bug] Default TypeMutator creates a new Global Type Var but the type checker relies on pointer equality

2022-02-15 Thread GitBox
slyubomirsky opened a new issue #10265: URL: https://github.com/apache/tvm/issues/10265 The default behavior of the Relay type mutator when it encounters Global Type Variables is to [create a new Global Type

[GitHub] [tvm] alanmacd opened a new issue #10264: [microTVM] microTVM requires PyYAML version that conflicts with other python package requirements

2022-02-15 Thread GitBox
alanmacd opened a new issue #10264: URL: https://github.com/apache/tvm/issues/10264 This error appears to be benign, but it creates noise when building/deploying microTVM VM: ``` default: ERROR: pip's dependency resolver does not currently take into account all the packages that

[GitHub] [tvm] mehrdadh commented on pull request #10263: [Docker][Hexagon] Add docker file and scripts

2022-02-15 Thread GitBox
mehrdadh commented on pull request #10263: URL: https://github.com/apache/tvm/pull/10263#issuecomment-1041011879 cc @areusch @kparzysz-quic -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] mehrdadh opened a new pull request #10263: [Docker][Hexagon] Add docker file and scripts

2022-02-15 Thread GitBox
mehrdadh opened a new pull request #10263: URL: https://github.com/apache/tvm/pull/10263 This PR adds docker file for hexagon docker image and scripts to build this image. Also it adds script which are used to build TVM for testing hexagon. cc @areusch @kparzysz-quic -- This is

[GitHub] [tvm] vinx13 commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
vinx13 commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807362399 ## File path: src/tir/analysis/device_constraint_utils.cc ## @@ -223,6 +225,23 @@ class ApplyDeviceConstraintsMutator : public StmtExprMutator {

[GitHub] [tvm] electriclilies commented on pull request #10262: [BUGFIX] Define kTargetPoolReadWriteAccess globally

2022-02-15 Thread GitBox
electriclilies commented on pull request #10262: URL: https://github.com/apache/tvm/pull/10262#issuecomment-1040958236 @mbs-octoml @manupa-arm PTAL! thanks :) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [tvm-rfcs] ccjoechou commented on pull request #48: [RFC][BYOC] Marvell ML/AI Accelerator Integration

2022-02-15 Thread GitBox
ccjoechou commented on pull request #48: URL: https://github.com/apache/tvm-rfcs/pull/48#issuecomment-1040923708 @areusch: No worries. I saw lots TVM emails coming from you & others working on other also important stuffs. We will wait for your feedback. -- This is an automated message

[GitHub] [tvm-rfcs] areusch commented on pull request #48: [RFC][BYOC] Marvell ML/AI Accelerator Integration

2022-02-15 Thread GitBox
areusch commented on pull request #48: URL: https://github.com/apache/tvm-rfcs/pull/48#issuecomment-1040919020 @ccjoechou sorry for the delay--i've gotten pretty busy with something and will hopefully have some bandwidth towards the end of the week. cc @jroesch @mbs-octoml in case

[tvm] branch main updated (6de8fea -> 64e94ab)

2022-02-15 Thread masahi
This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 6de8fea add back supported tests (#10116) add 64e94ab [QNN] Register a bunch of unary elementwise ops (#10086) No

[GitHub] [tvm] masahi merged pull request #10086: [QNN] Register a bunch of unary elementwise ops

2022-02-15 Thread GitBox
masahi merged pull request #10086: URL: https://github.com/apache/tvm/pull/10086 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807377035 ## File path: tests/python/unittest/test_tir_transform_instrument_bound_checkers.py ## @@ -94,7 +94,7 @@ def test_out_of_bounds_vectorize_llvm(nn,

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807375704 ## File path: src/tir/transforms/merge_dynamic_shared_memory_allocations.cc ## @@ -294,22 +308,61 @@ class DynamicSharedMemoryRewriter : public

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807371988 ## File path: src/tir/transforms/split_host_device.cc ## @@ -155,10 +160,27 @@ class VarUseDefAnalysis : public StmtExprMutator { } PrimExpr

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807370425 ## File path: python/tvm/script/tir/special_stmt.py ## @@ -132,6 +132,7 @@ def match_buffer( align=-1, offset_factor=0,

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #10086: [QNN] Register a bunch of unary elementwise ops

2022-02-15 Thread GitBox
AndrewZhaoLuo commented on pull request #10086: URL: https://github.com/apache/tvm/pull/10086#issuecomment-1040857889 cc @masahi we cool with merging this? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807365129 ## File path: include/tvm/tir/function.h ## @@ -136,15 +157,31 @@ class PrimFunc : public BaseFunc { public: /*! * \brief Constructor + * *

[GitHub] [tvm] Lunderberg commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
Lunderberg commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r807356985 ## File path: include/tvm/tir/buffer.h ## @@ -98,10 +113,17 @@ class BufferNode : public Object { } bool SEqualReduce(const BufferNode* other,

[GitHub] [tvm] vinx13 commented on pull request #10066: [TIR] Add software pipelining

2022-02-15 Thread GitBox
vinx13 commented on pull request #10066: URL: https://github.com/apache/tvm/pull/10066#issuecomment-1040838036 comments are addressed -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] vinx13 closed issue #10210: [Bug] Tree reduction with shared memory is buggy where there is threadIdx.y

2022-02-15 Thread GitBox
vinx13 closed issue #10210: URL: https://github.com/apache/tvm/issues/10210 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] electriclilies edited a comment on pull request #9848: [Relay] [Virtual Device] Store function result virtual device in virtual_device_ field

2022-02-15 Thread GitBox
electriclilies edited a comment on pull request #9848: URL: https://github.com/apache/tvm/pull/9848#issuecomment-1040811596 I ended up just lifting the result virtual device out of the attributes, will discuss dealing with the text representation in another PR. -- This is an automated

[GitHub] [tvm] electriclilies commented on pull request #9848: [Relay] [Virtual Device] Store function result virtual device in virtual_device_ field

2022-02-15 Thread GitBox
electriclilies commented on pull request #9848: URL: https://github.com/apache/tvm/pull/9848#issuecomment-1040811596 I ended up just lifting the result virtual device out of the attributes, will discuss dealing with the on_device ops in a later PR. -- This is an automated message from

[GitHub] [tvm] electriclilies opened a new pull request #10262: [BUGFIX] Define kTargetPoolReadWriteAccess globally

2022-02-15 Thread GitBox
electriclilies opened a new pull request #10262: URL: https://github.com/apache/tvm/pull/10262 kTargetPoolReadWriteAccess was defined inside the PoolInfo class, leading to a compile time error for me. Making it global fixes the issue, and is also is consistent with how other attribute

[GitHub] [tvm] junrushao1994 commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

2022-02-15 Thread GitBox
junrushao1994 commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r807324302 ## File path: include/tvm/tir/transform.h ## @@ -29,6 +29,14 @@ #include #include +#include + +// Forward declarations from tvm::relay

[GitHub] [tvm] mehrdadh opened a new issue #10261: [Tracking Issue] Add Hexagon CI docker image to pre-commit CI

2022-02-15 Thread GitBox
mehrdadh opened a new issue #10261: URL: https://github.com/apache/tvm/issues/10261 We are adding a docker image to test hexagon tests in pre-commit CI. These are the steps: - Build ci-hexagon docker image - Test docker image in staging - Add ci-hexagon to tlcpack docker hub

[GitHub] [tvm] masahi commented on pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
masahi commented on pull request #10239: URL: https://github.com/apache/tvm/pull/10239#issuecomment-1040796147 > Need to read pass a little more carefully to understand difference between this and existing fq2i passes The key is that in the models that existing pass expects, there

[GitHub] [tvm] driazati commented on a change in pull request #10247: [docker] Update CI to Python 3.7 and Ubuntu 18

2022-02-15 Thread GitBox
driazati commented on a change in pull request #10247: URL: https://github.com/apache/tvm/pull/10247#discussion_r807306364 ## File path: docker/install/ubuntu1804_install_python.sh ## @@ -23,10 +23,10 @@ set -o pipefail # install python and pip, don't modify this, modify

[GitHub] [tvm] driazati commented on a change in pull request #10247: [docker] Update CI to Python 3.7 and Ubuntu 18

2022-02-15 Thread GitBox
driazati commented on a change in pull request #10247: URL: https://github.com/apache/tvm/pull/10247#discussion_r807296364 ## File path: docker/install/ubuntu_install_caffe.sh ## @@ -16,17 +16,15 @@ # specific language governing permissions and limitations # under the

[tvm] branch main updated (2b00835 -> 6de8fea)

2022-02-15 Thread masahi
This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 2b00835 [Relay] Make DeviceAnalyzer a mixed mode visitor (#10248) add 6de8fea add back supported tests (#10116)

[GitHub] [tvm] masahi merged pull request #10116: [Onnx] add back supported tests

2022-02-15 Thread GitBox
masahi merged pull request #10116: URL: https://github.com/apache/tvm/pull/10116 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] AndrewZhaoLuo commented on a change in pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
AndrewZhaoLuo commented on a change in pull request #10239: URL: https://github.com/apache/tvm/pull/10239#discussion_r807263764 ## File path: tests/python/relay/test_pass_fake_quantization_to_integer.py ## @@ -627,3 +628,217 @@ def conv2d(expr, type_map): # pylint:

[GitHub] [tvm] kparzysz-quic opened a new pull request #10259: [Hexagon] Don't use cmake glob for auto-generated source files

2022-02-15 Thread GitBox
kparzysz-quic opened a new pull request #10259: URL: https://github.com/apache/tvm/pull/10259 Glob treats inputs as patterns: if the file with a given name does not exist (is to be generated later), it won't be added to the output. -- This is an automated message from the Apache Git

[GitHub] [tvm] masahi commented on pull request #10229: [UnitTest] Disable ptx mma tests on unsupported nvcc versions.

2022-02-15 Thread GitBox
masahi commented on pull request #10229: URL: https://github.com/apache/tvm/pull/10229#issuecomment-1040734463 > Thank you, and more lint fixes applied. It looks like the issue in my local setup is that `black` can't be run from inside the `python/tvm/contrib`, because

[GitHub] [tvm] Lunderberg commented on pull request #10229: [UnitTest] Disable ptx mma tests on unsupported nvcc versions.

2022-02-15 Thread GitBox
Lunderberg commented on pull request #10229: URL: https://github.com/apache/tvm/pull/10229#issuecomment-1040730935 Thank you, and more lint fixes applied. It looks like the issue in my local setup is that `black` can't be run from inside the `python/tvm/contrib`, because

[GitHub] [tvm] tmoreau89 commented on a change in pull request #10256: [TVMC] Add an end_to_end benchmarking argument when benchmarking.

2022-02-15 Thread GitBox
tmoreau89 commented on a change in pull request #10256: URL: https://github.com/apache/tvm/pull/10256#discussion_r807264329 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -557,8 +569,11 @@ def run_module( module.run() times = [] else: -

[GitHub] [tvm] mei-ye commented on pull request #10248: [Relay] Make DeviceAnalyzer a mixed mode visitor

2022-02-15 Thread GitBox
mei-ye commented on pull request #10248: URL: https://github.com/apache/tvm/pull/10248#issuecomment-1040728058 I confirmed that this fixes the issue I saw on windows. Thanks! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[GitHub] [tvm] Lunderberg opened a new pull request #10258: [Cuda] Updated bfloat16 math defs.

2022-02-15 Thread GitBox
Lunderberg opened a new pull request #10258: URL: https://github.com/apache/tvm/pull/10258 Required to pass `test_cuda_bf16_vectorize_add` in `tests/python/unittest/test_target_codegen_cuda.py`. -- This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [tvm] masahi commented on pull request #10229: [UnitTest] Disable ptx mma tests on unsupported nvcc versions.

2022-02-15 Thread GitBox
masahi commented on pull request #10229: URL: https://github.com/apache/tvm/pull/10229#issuecomment-1040721744 @Lunderberg Need more lint fix -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] driazati commented on pull request #10246: [skip ci][ci] Use sccache in builds

2022-02-15 Thread GitBox
driazati commented on pull request #10246: URL: https://github.com/apache/tvm/pull/10246#issuecomment-1040706377 > Why do we need to have this public? Can't the instances assume a role to access it? Sorry public wasn't the right word, I meant OSS, we had spun this up in the

[GitHub] [tvm] driazati commented on a change in pull request #10246: [skip ci][ci] Use sccache in builds

2022-02-15 Thread GitBox
driazati commented on a change in pull request #10246: URL: https://github.com/apache/tvm/pull/10246#discussion_r807238655 ## File path: tests/scripts/task_build.sh ## @@ -15,5 +15,29 @@ # KIND, either express or implied. See the License for the # specific language

[GitHub] [tvm] driazati opened a new pull request #10257: [ci] Move pip dependencies to docker image

2022-02-15 Thread GitBox
driazati opened a new pull request #10257: URL: https://github.com/apache/tvm/pull/10257 Following on from #10246, this moves the `pip install`-at-runtime deps to the docker image install so they are baked in. Thanks for contributing to TVM! Please refer to guideline

[GitHub] [tvm-rfcs] TejashShah commented on pull request #46: Module Based Model Runtime for AOT

2022-02-15 Thread GitBox
TejashShah commented on pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46#issuecomment-1040676503 Thanks @manupa-arm for reviewing this RFC and providing crucial feedback in time-bound fashion. -- This is an automated message from the Apache Git Service. To respond to

[GitHub] [tvm-rfcs] manupa-arm merged pull request #46: Module Based Model Runtime for AOT

2022-02-15 Thread GitBox
manupa-arm merged pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] masahi commented on pull request #9802: [Runtime][ThreadPool]Refactor affinity function and support CPU affinity list setting.

2022-02-15 Thread GitBox
masahi commented on pull request #9802: URL: https://github.com/apache/tvm/pull/9802#issuecomment-1040673244 Please help reviewing @tqchen @junrushao1994 @FrozenGene -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

[GitHub] [tvm] masahi commented on a change in pull request #9802: [Runtime][ThreadPool]Refactor affinity function and support CPU affinity list setting.

2022-02-15 Thread GitBox
masahi commented on a change in pull request #9802: URL: https://github.com/apache/tvm/pull/9802#discussion_r807210048 ## File path: src/runtime/thread_pool.cc ## @@ -337,7 +339,7 @@ class ThreadPool { new tvm::runtime::threading::ThreadGroup(

[GitHub] [tvm] jwfromm commented on a change in pull request #10256: [TVMC] Add an end_to_end benchmarking argument when benchmarking.

2022-02-15 Thread GitBox
jwfromm commented on a change in pull request #10256: URL: https://github.com/apache/tvm/pull/10256#discussion_r807188490 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -557,8 +569,11 @@ def run_module( module.run() times = [] else: -

[GitHub] [tvm] ekalda commented on a change in pull request #10250: [microNPU] Refactor microNPU Codegen and Legalization tests

2022-02-15 Thread GitBox
ekalda commented on a change in pull request #10250: URL: https://github.com/apache/tvm/pull/10250#discussion_r807076836 ## File path: tests/python/contrib/test_ethosu/end_to_end/comparison_infra.py ## @@ -0,0 +1,158 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [tvm] dchauhan-arm commented on a change in pull request #10250: [microNPU] Refactor microNPU Codegen and Legalization tests

2022-02-15 Thread GitBox
dchauhan-arm commented on a change in pull request #10250: URL: https://github.com/apache/tvm/pull/10250#discussion_r806825481 ## File path: tests/python/contrib/test_ethosu/end_to_end/__init__.py ## @@ -0,0 +1,17 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [tvm] lhutton1 commented on a change in pull request #10254: [microNPU] Introduce a pass to remove redundant identity operations

2022-02-15 Thread GitBox
lhutton1 commented on a change in pull request #10254: URL: https://github.com/apache/tvm/pull/10254#discussion_r807133809 ## File path: tests/python/contrib/test_ethosu/test_identity_optimizer.py ## @@ -0,0 +1,245 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [tvm] manupa-arm commented on a change in pull request #10022: [microNPU] enable USMP

2022-02-15 Thread GitBox
manupa-arm commented on a change in pull request #10022: URL: https://github.com/apache/tvm/pull/10022#discussion_r806982398 ## File path: python/tvm/relay/backend/contrib/ethosu/tir_to_cs_translator.py ## @@ -105,22 +195,27 @@ def translate(tir_module, params):

[GitHub] [tvm] areusch commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

2022-02-15 Thread GitBox
areusch commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r807135033 ## File path: include/tvm/ir/module.h ## @@ -349,6 +386,9 @@ class IRModuleNode : public Object { */ std::unordered_set import_set_; friend class

[GitHub] [tvm] tmoreau89 commented on pull request #10226: [TVMC] End to end benchmarking by default

2022-02-15 Thread GitBox
tmoreau89 commented on pull request #10226: URL: https://github.com/apache/tvm/pull/10226#issuecomment-1040572698 @jwfromm great points, I'll close this PR in favor of your PR which seems less disruptive -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] Icemist commented on pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
Icemist commented on pull request #10239: URL: https://github.com/apache/tvm/pull/10239#issuecomment-1040447093 > Please also show the reference result (accuracy without fq2i). | Model type | Accuracy | | |

[GitHub] [tvm] tmoreau89 commented on a change in pull request #10256: [TVMC] Add an end_to_end benchmarking argument when benchmarking.

2022-02-15 Thread GitBox
tmoreau89 commented on a change in pull request #10256: URL: https://github.com/apache/tvm/pull/10256#discussion_r807119745 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -557,8 +569,11 @@ def run_module( module.run() times = [] else: -

[GitHub] [tvm] ekalda commented on a change in pull request #10022: [microNPU] enable USMP

2022-02-15 Thread GitBox
ekalda commented on a change in pull request #10022: URL: https://github.com/apache/tvm/pull/10022#discussion_r806892717 ## File path: python/tvm/relay/backend/contrib/ethosu/tir_to_cs_translator.py ## @@ -105,22 +195,27 @@ def translate(tir_module, params):

[GitHub] [tvm] tmoreau89 closed pull request #10226: [TVMC] End to end benchmarking by default

2022-02-15 Thread GitBox
tmoreau89 closed pull request #10226: URL: https://github.com/apache/tvm/pull/10226 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] mbrookhart commented on a change in pull request #10239: Add a conversion of individual operations in FQ2I pass.

2022-02-15 Thread GitBox
mbrookhart commented on a change in pull request #10239: URL: https://github.com/apache/tvm/pull/10239#discussion_r807010832 ## File path: src/relay/transforms/fake_quantization_to_integer.cc ## @@ -270,8 +293,233 @@ class FakeQuantizationRewriter : public MixedModeMutator {

[GitHub] [tvm] ekalda commented on a change in pull request #10254: [microNPU] Introduce a pass to remove redundant identity operations

2022-02-15 Thread GitBox
ekalda commented on a change in pull request #10254: URL: https://github.com/apache/tvm/pull/10254#discussion_r807032862 ## File path: tests/python/contrib/test_ethosu/test_identity_optimizer.py ## @@ -0,0 +1,245 @@ +# Licensed to the Apache Software Foundation (ASF) under one

[GitHub] [tvm] tmoreau89 commented on pull request #10248: [Relay] Make DeviceAnalyzer a mixed mode visitor

2022-02-15 Thread GitBox
tmoreau89 commented on pull request #10248: URL: https://github.com/apache/tvm/pull/10248#issuecomment-1040495770 Thank you @mbrookhart and @masahi - PR has been merged. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and

[GitHub] [tvm] jwfromm commented on pull request #10226: [TVMC] End to end benchmarking by default

2022-02-15 Thread GitBox
jwfromm commented on pull request #10226: URL: https://github.com/apache/tvm/pull/10226#issuecomment-1040571793 Thanks for making this change @tmoreau89. I think changing the default behavior is likely going to cause issues for entrenched users. I just put up PR #10256 which exposes an

[GitHub] [tvm] tmoreau89 merged pull request #10248: [Relay] Make DeviceAnalyzer a mixed mode visitor

2022-02-15 Thread GitBox
tmoreau89 merged pull request #10248: URL: https://github.com/apache/tvm/pull/10248 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm-rfcs] manupa-arm commented on pull request #46: Module Based Model Runtime for AOT

2022-02-15 Thread GitBox
manupa-arm commented on pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46#issuecomment-1040600782 Thanks @areusch . Looking forward for AoT being enabled in more places :). -- This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #10116: [Onnx] add back supported tests

2022-02-15 Thread GitBox
AndrewZhaoLuo commented on pull request #10116: URL: https://github.com/apache/tvm/pull/10116#issuecomment-1040570086 @masahi yeah this should be good to go -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [tvm] lhutton1 commented on pull request #10254: [microNPU] Introduce a pass to remove redundant identity operations

2022-02-15 Thread GitBox
lhutton1 commented on pull request #10254: URL: https://github.com/apache/tvm/pull/10254#issuecomment-1040589105 Thanks for the review @ekalda, I added a test case and removed the print among a couple more things I spotted :) -- This is an automated message from the Apache Git Service.

[GitHub] [tvm] masahi commented on pull request #10234: [Runtime][Pipeline Executor] multiple threads management and the data forwarding notification mechanism.

2022-02-15 Thread GitBox
masahi commented on pull request #10234: URL: https://github.com/apache/tvm/pull/10234#issuecomment-1039884305 Thanks @huajsj. This is interesting and tricky code, I'll try review carefully. Where does this PR sit in the feature road map? -- This is an automated message from the

[GitHub] [tvm] tmoreau89 closed issue #10135: [Bug] Update DevicePlanner to use MixedModeVisitor

2022-02-15 Thread GitBox
tmoreau89 closed issue #10135: URL: https://github.com/apache/tvm/issues/10135 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] areusch closed issue #10150: [CI Problem] Tensorflow frontend test CUDA errors in CI

2022-02-15 Thread GitBox
areusch closed issue #10150: URL: https://github.com/apache/tvm/issues/10150 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] vinx13 commented on a change in pull request #9727: [TE][TIR] Implement layout transformations, non-flat memory buffers

2022-02-15 Thread GitBox
vinx13 commented on a change in pull request #9727: URL: https://github.com/apache/tvm/pull/9727#discussion_r805065309 ## File path: tests/python/unittest/test_tir_transform_instrument_bound_checkers.py ## @@ -94,7 +94,7 @@ def test_out_of_bounds_vectorize_llvm(nn, index_a,

[GitHub] [tvm] Lunderberg commented on pull request #10229: [UnitTest] Disable ptx mma tests on unsupported nvcc versions.

2022-02-15 Thread GitBox
Lunderberg commented on pull request #10229: URL: https://github.com/apache/tvm/pull/10229#issuecomment-1040347110 Thank you, and the lint errors should be resolved now. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and

[GitHub] [tvm] mbs-octoml commented on a change in pull request #10248: [Relay] Make DeviceAnalyzer a mixed mode visitor

2022-02-15 Thread GitBox
mbs-octoml commented on a change in pull request #10248: URL: https://github.com/apache/tvm/pull/10248#discussion_r807093747 ## File path: src/relay/transforms/device_planner.cc ## @@ -22,9 +22,6 @@ * \brief Determines a unique \p VirtualDevice to hold the result of every

[GitHub] [tvm] ekalda commented on pull request #10251: [microNPU] Add support for conv2d running on two cores on U65

2022-02-15 Thread GitBox
ekalda commented on pull request #10251: URL: https://github.com/apache/tvm/pull/10251#issuecomment-1040115827 cc: @manupa-arm @lhutton1 @NicolaLancellotti @mbaret -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the

[GitHub] [tvm] huajsj commented on pull request #10234: [Runtime][Pipeline Executor] multiple threads management and the data forwarding notification mechanism.

2022-02-15 Thread GitBox
huajsj commented on pull request #10234: URL: https://github.com/apache/tvm/pull/10234#issuecomment-1039411862 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To

[GitHub] [tvm] masahi commented on pull request #10230: [TOPI] VNNI support for int8 dense

2022-02-15 Thread GitBox
masahi commented on pull request #10230: URL: https://github.com/apache/tvm/pull/10230#issuecomment-1039588965 Another perf result, this time on a desktop CPU `i5-11400 @ 2.60GHz`, 6 threads. TVM showing excellent performance! M|N|K|TVM|FBGEMM -- | -- | -- | --| -- |

[GitHub] [tvm] lhutton1 commented on a change in pull request #10242: [microNPU] Add MobileNetv2 test to network tests

2022-02-15 Thread GitBox
lhutton1 commented on a change in pull request #10242: URL: https://github.com/apache/tvm/pull/10242#discussion_r806313919 ## File path: tests/python/contrib/test_ethosu/test_networks.py ## @@ -65,5 +65,31 @@ def test_forward_mobilenet_v1(accel_type):

[GitHub] [tvm] junrushao1994 commented on pull request #10235: [Torch] Fix conv2d transpose with group

2022-02-15 Thread GitBox
junrushao1994 commented on pull request #10235: URL: https://github.com/apache/tvm/pull/10235#issuecomment-1039759287 right. I noticed the same issue very recently. Thanks for fixing this!! -- This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [tvm] jacobbohlin commented on a change in pull request #10242: [microNPU] Add MobileNetv2 test to network tests

2022-02-15 Thread GitBox
jacobbohlin commented on a change in pull request #10242: URL: https://github.com/apache/tvm/pull/10242#discussion_r806653030 ## File path: tests/python/contrib/test_ethosu/test_networks.py ## @@ -65,5 +65,31 @@ def test_forward_mobilenet_v1(accel_type):

[GitHub] [tvm] masahi commented on issue #8596: [RFC][Tracking Issue] Pipeline Executor For Compute graph pipeline

2022-02-15 Thread GitBox
masahi commented on issue #8596: URL: https://github.com/apache/tvm/issues/8596#issuecomment-1039909509 At one point can we expect a non-trivial demo? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to

[GitHub] [tvm] junrushao1994 closed issue #10223: [Bug] Conv2DTranspose with groups not working correctly

2022-02-15 Thread GitBox
junrushao1994 closed issue #10223: URL: https://github.com/apache/tvm/issues/10223 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

[GitHub] [tvm] mshr-h commented on a change in pull request #10247: [docker] Update CI to Python 3.7 and Ubuntu 18

2022-02-15 Thread GitBox
mshr-h commented on a change in pull request #10247: URL: https://github.com/apache/tvm/pull/10247#discussion_r806643624 ## File path: docker/install/ubuntu_install_boost.sh ## @@ -0,0 +1,30 @@ +#!/bin/bash +# Licensed to the Apache Software Foundation (ASF) under one +# or

[GitHub] [tvm] JCBrouwer commented on issue #10223: [Bug] Conv2DTranspose with groups not working correctly

2022-02-15 Thread GitBox
JCBrouwer commented on issue #10223: URL: https://github.com/apache/tvm/issues/10223#issuecomment-1039250272 Yes using CUDNN is fine with me, that would be great :) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the

[GitHub] [tvm] tqchen commented on a change in pull request #9748: Overload get() function for `Optional` type.

2022-02-15 Thread GitBox
tqchen commented on a change in pull request #9748: URL: https://github.com/apache/tvm/pull/9748#discussion_r805954346 ## File path: include/tvm/runtime/container/optional.h ## @@ -93,6 +93,11 @@ class Optional : public ObjectRef { ICHECK(data_ != nullptr); return

[GitHub] [tvm] junrushao1994 merged pull request #10235: [Torch] Fix conv2d transpose with group

2022-02-15 Thread GitBox
junrushao1994 merged pull request #10235: URL: https://github.com/apache/tvm/pull/10235 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail:

  1   2   3   >