[GitHub] [tvm] manupa-arm commented on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

manupa-arm commented on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920580279 Our proposal is to add a check to that loop whether it has 'local' storage_scope before we place them into the stack as it is the solution that works for the wider definition of the

[GitHub] [tvm] manupa-arm commented on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

manupa-arm commented on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920577170 @tqchen @mbs-octoml , This is not specific to Arm(R) Ethos(TM)-U codegen and its generally applicable for any micro controller where we would want to avoid creating allocation

[tvm] branch main updated (89bcc79 -> 4c77bae)

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 89bcc79 fix (#9021) add 4c77bae [Onnx] Add momentum (#9000) No new revisions were added by this update. Summary

[tvm] branch main updated (ff0868f -> 89bcc79)

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 ff0868f [Community] @AndrewZhaoLuo -> Reviewer (#9020) add 89bcc79 fix (#9021) No new revisions were added by

[GitHub] [tvm] masahi merged pull request #9000: [Onnx] Add momentum

masahi merged pull request #9000: URL: https://github.com/apache/tvm/pull/9000 -- 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:

[tvm] branch main updated (148ddca -> ff0868f)

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 148ddca [Hexagon] Implement model launcher (#8986) add ff0868f [Community] @AndrewZhaoLuo -> Reviewer (#9020) No

[GitHub] [tvm] masahi merged pull request #9021: [TOPI] Fix more pooling schedule

masahi merged pull request #9021: URL: https://github.com/apache/tvm/pull/9021 -- 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 merged pull request #9020: [Community] @AndrewZhaoLuo -> Reviewer

masahi merged pull request #9020: URL: https://github.com/apache/tvm/pull/9020 -- 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:

[tvm] branch main updated (777ace3 -> 148ddca)

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 777ace3 [Relay][Pass] Add ExtractOperators pass (#8996) add 148ddca [Hexagon] Implement model launcher (#8986) No

[GitHub] [tvm] masahi merged pull request #8986: [Hexagon] Implement model launcher

masahi merged pull request #8986: URL: https://github.com/apache/tvm/pull/8986 -- 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] junrushao1994 edited a comment on pull request #9023: [TE] Support negative indices

junrushao1994 edited a comment on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920494401 I think I sort of understand the usecase here: some negative indices are not known to be negative until in runtime. This forces us to defer the conversion from

[GitHub] [tvm] junrushao1994 edited a comment on pull request #9023: [TE] Support negative indices

junrushao1994 edited a comment on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920494401 I think I sort of understand the usecase here: some negative indices are known to be negative until in runtime. This forces us to defer the conversion from

[GitHub] [tvm] huajsj commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

huajsj commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709691953 ## File path: include/tvm/runtime/logging.h ## @@ -409,6 +416,68 @@ inline bool DebugLoggingEnabled() { return state == 1; } +/*! \brief Helpers for \p

[GitHub] [tvm] huajsj commented on a change in pull request #9018: [microTVM][autoTVM] Follow up fixes to #9003

huajsj commented on a change in pull request #9018: URL: https://github.com/apache/tvm/pull/9018#discussion_r709680012 ## File path: python/tvm/micro/testing.py ## @@ -0,0 +1,36 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license

[GitHub] [tvm] huajsj commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

huajsj commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709691953 ## File path: include/tvm/runtime/logging.h ## @@ -409,6 +416,68 @@ inline bool DebugLoggingEnabled() { return state == 1; } +/*! \brief Helpers for \p

[GitHub] [tvm] mehrdadh opened a new pull request #9026: [microtvm][Zephyr] Add MAIN_STACK_SIZE option to API server

mehrdadh opened a new pull request #9026: URL: https://github.com/apache/tvm/pull/9026 - When using API server in for external project, sometime we need to set main stack size for zephyr board if model is large. This PR adds this option to Zephyr project API. - In addition, this PR

[GitHub] [tvm] junrushao1994 commented on pull request #9023: [TE] Support negative indices

junrushao1994 commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920494401 I think I sort of understand the usecase here, where some indices are known negative only in runtime, which forces us to check things on runtime. On the other hand, I am not

[GitHub] [tvm] masahi merged pull request #8996: [Relay][Pass] Add ExtractOperators pass

masahi merged pull request #8996: URL: https://github.com/apache/tvm/pull/8996 -- 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:

[tvm] branch main updated (6f5b674 -> 777ace3)

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 6f5b674 [BYOC][TensorRT] Add TensorRT own int8 calibration support to TensorRT BYOC integration (#8808) add

[GitHub] [tvm] tqchen commented on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen commented on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920486224 Right, this the gets to the target dependent generation regime where TargetKind attribute is indeed the right solution. We should also send a PR to add comments to that code block so we

[GitHub] [tvm] mbs-octoml edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

mbs-octoml edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920485102 Thanks so much for for the context. I'll try to capture that in a comment. This is a 'bug' only in the sense that this heuristic is not working for the EthosU AOT codegen, I

[GitHub] [tvm] mbs-octoml edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

mbs-octoml edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920485102 Thanks so much for for the context. I'll try to capture that in a comment. This is a 'bug' only in the sense that this heuristic is not working for the EthosU AOT codegen, I

[GitHub] [tvm] mbs-octoml edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

mbs-octoml edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920485102 Thanks so much for for the context. I'll try to capture that in a comment. This is a 'bug' only in the sense that this heuristic is not working for the EthosU AOT codegen, I

[GitHub] [tvm] mbs-octoml commented on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

mbs-octoml commented on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920485102 Thank so much for for the context. I'll try to capture that in a comment. This is a 'bug' only in the sense that this heuristic is not working for the EthosU AOT codegen, I think

[GitHub] [tvm] csullivan commented on a change in pull request #8986: [Hexagon] Implement model launcher

csullivan commented on a change in pull request #8986: URL: https://github.com/apache/tvm/pull/8986#discussion_r709658089 ## File path: src/runtime/hexagon/launcher/README.md ## @@ -0,0 +1,173 @@ + + + + + + + + + + + + + + + + +# Hexagon Graph Launcher + +## Compilation +

[GitHub] [tvm] tqchen commented on pull request #9023: [TE] Support negative indices

tqchen commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920469704 @yzh119 I believe the intended usage was for the case where the value of indices are not known at compile time, otherwise compiler will be able to prove and simplify the

[GitHub] [tvm] tqchen edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920463412 @mbs-octoml To give a bit of context In the context of CPU, we want to preserve small alloca until the code generation point. And then the code will generate the stack

[GitHub] [tvm] tqchen edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920463412 @mbs-octoml To give a bit of context In the context of CPU, we want to preserve small alloca until the code generation point. And then the code will generate the stack

[GitHub] [tvm] tqchen edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920463412 @mbs-octoml To give a bit of context In the context of CPU, we want to preserve small alloca until the code generation point. And then the code will generate the stack

[GitHub] [tvm] tqchen edited a comment on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen edited a comment on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920463412 @mbs-octoml To give a bit of context In the context of CPU, we want to preserve small alloca until the code generation point. And then the code will generate the stack

[GitHub] [tvm] yzh119 commented on pull request #9023: [TE] Support negative indices

yzh119 commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920465162 Is the negative indices used inside TE? If so I wonder can we make it a transformation pass rather than runtime behavior? -- This is an automated message from the Apache Git

[GitHub] [tvm] tqchen commented on issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

tqchen commented on issue #9022: URL: https://github.com/apache/tvm/issues/9022#issuecomment-920463412 @mbs-octoml I believe the current behavior is intended. In the context of CPU, we want to preserve small alloca until the code generation point. And then the code will generate

[GitHub] [tvm] tqchen edited a comment on pull request #9023: [TE] Support negative indices

tqchen edited a comment on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920459959 I think this can be useful for some of the indexing operations. We could try to make it more explicit though. e.g. introduce a new API.

[GitHub] [tvm] tqchen commented on pull request #9023: [TE] Support negative indices

tqchen commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920459959 I think this can be useful for some of the indexing operations. We could try to make it more explicit though. e.g. introduce a new API.

[GitHub] [tvm] tmoreau89 commented on a change in pull request #8955: [Hexagon] Pytestify Hexagon unit test

tmoreau89 commented on a change in pull request #8955: URL: https://github.com/apache/tvm/pull/8955#discussion_r709644590 ## File path: tests/python/unittest/test_target_codegen_hexagon.py ## @@ -26,8 +26,10 @@ import tvm.contrib.hexagon as hexagon -# Register a phony

[GitHub] [tvm] tqchen commented on a change in pull request #9023: [TE] Support negative indices

tqchen commented on a change in pull request #9023: URL: https://github.com/apache/tvm/pull/9023#discussion_r709643444 ## File path: src/te/tensor.cc ## @@ -39,15 +39,26 @@ IterVar reduce_axis(Range dom, std::string name) { return IterVar(dom, Var(name) Var var(std::string

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709643127 ## File path: include/tvm/runtime/logging.h ## @@ -409,6 +416,68 @@ inline bool DebugLoggingEnabled() { return state == 1; } +/*! \brief Helpers for

[GitHub] [tvm] mbrookhart commented on issue #8978: [Bug][VTA][OpenCL] If allowed to allocate in stack, VTA multiple target test will fail

mbrookhart commented on issue #8978: URL: https://github.com/apache/tvm/issues/8978#issuecomment-920445448 @mbs-octoml opened this issue to track an approach to better distinguish multiple targets https://github.com/apache/tvm/issues/9022 -- This is an automated message from the Apache

[GitHub] [tvm] mbrookhart commented on issue #8978: [Bug][VTA][OpenCL] If allowed to allocate in stack, VTA multiple target test will fail

mbrookhart commented on issue #8978: URL: https://github.com/apache/tvm/issues/8978#issuecomment-920445094 FYI - this PR renables the test that this change breaks https://github.com/apache/tvm/pull/9019 -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] mbrookhart commented on issue #8977: [Bug][VM] If not allocate on stack, VM runtime cannot work？

mbrookhart commented on issue #8977: URL: https://github.com/apache/tvm/issues/8977#issuecomment-920444851 This is the test that fails in the VM without this code: https://github.com/apache/tvm/pull/9019 -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] mbrookhart commented on pull request #9023: [TE] Support negative indices

mbrookhart commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920426352 I like this idea. I've been slowly throwing `relay.where` around the ops and the importers to solve this problem when I hit it, but that adds some complication to ops, fusion,

[GitHub] [tvm] huajsj commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

huajsj commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709585141 ## File path: include/tvm/runtime/logging.h ## @@ -409,6 +416,68 @@ inline bool DebugLoggingEnabled() { return state == 1; } +/*! \brief Helpers for \p

[GitHub] [tvm] mehrdadh commented on pull request #9018: [microTVM][autoTVM] Follow up fixes to #9003

mehrdadh commented on pull request #9018: URL: https://github.com/apache/tvm/pull/9018#issuecomment-920391097 cc @leandron @areusch for possible review. -- 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

[GitHub] [tvm] kparzysz-quic opened a new pull request #9025: [Hexagon] Disable `thread_local` on Hexagon

kparzysz-quic opened a new pull request #9025: URL: https://github.com/apache/tvm/pull/9025 This is specific to running code on hardware: libc++abi can create TLS keys with destructors in the libc++abi library. Despite that, the library gets unloaded before the keys are destroyed,

[GitHub] [tvm] kparzysz-quic opened a new pull request #9024: [Hexagon] Allow undefined symbols in libtvm_runtime.so on Hexagon

kparzysz-quic opened a new pull request #9024: URL: https://github.com/apache/tvm/pull/9024 The shared library `libtvm_runtime.so` (or any other shared library built for Hexagon) will not contain definitions of symbols from libc. To avoid undefined symbol errors, turn that check off when

[GitHub] [tvm] masahi edited a comment on pull request #9023: [TE] Support negative indices

masahi edited a comment on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920369621 For the context, this PR was split from another PR following the discussion https://github.com/apache/tvm/pull/8971#discussion_r707850237, since this could be potentially a

[GitHub] [tvm] masahi edited a comment on pull request #9023: [TE] Support negative indices

masahi edited a comment on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920369621 For the context, this PR was split from another PR following the discussion https://github.com/apache/tvm/pull/8971#discussion_r707850237 cc @tqchen @junrushao1994

[GitHub] [tvm] masahi commented on pull request #9023: [TE] Support negative indices

masahi commented on pull request #9023: URL: https://github.com/apache/tvm/pull/9023#issuecomment-920369621 For the context, this PR was split from another PR following the discussion https://github.com/apache/tvm/pull/8971#discussion_r707850237 -- This is an automated message from the

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #8971: [Onnx] Fix NLL Loss tests

AndrewZhaoLuo commented on pull request #8971: URL: https://github.com/apache/tvm/pull/8971#issuecomment-920359675 https://github.com/apache/tvm/pull/9023 <-- discussion about making negative indices simpler -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] AndrewZhaoLuo opened a new pull request #9023: [TE] Support negative indices

AndrewZhaoLuo opened a new pull request #9023: URL: https://github.com/apache/tvm/pull/9023 Negative indices are a pretty common feature in most modern programming languages and libraries. This proposed PR would add optional support for negative indices out of the box for tensors in TE.

[GitHub] [tvm] comaniac commented on pull request #8808: [BYOC][TensorRT] Add TensorRT own int8 calibration support to TensorRT BYOC integration

comaniac commented on pull request #8808: URL: https://github.com/apache/tvm/pull/8808#issuecomment-920348214 Thanks @tiandiao123 @trevor-m @FrozenGene @jcf94 @vinx13 @Laurawly -- This is an automated message from the Apache Git Service. To respond to the message, please log on to

[tvm] branch main updated (e44f6c0 -> 6f5b674)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from e44f6c0 [ONNX] Add Einsum converter (#8985) add 6f5b674 [BYOC][TensorRT] Add TensorRT own int8 calibration

[GitHub] [tvm] comaniac merged pull request #8808: [BYOC][TensorRT] Add TensorRT own int8 calibration support to TensorRT BYOC integration

comaniac merged pull request #8808: URL: https://github.com/apache/tvm/pull/8808 -- 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] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709543170 ## File path: include/tvm/runtime/logging.h ## @@ -395,10 +399,13 @@ class LogMessageVoidify { inline bool DebugLoggingEnabled() { static int state =

[GitHub] [tvm] comaniac commented on a change in pull request #8808: [BYOC][TensorRT] Add TensorRT own int8 calibration support to TensorRT BYOC integration

comaniac commented on a change in pull request #8808: URL: https://github.com/apache/tvm/pull/8808#discussion_r709538959 ## File path: src/runtime/contrib/tensorrt/tensorrt_builder.h ## @@ -153,6 +153,9 @@ class TensorRTBuilder { /*! \brief Whether to automatically convert

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709540519 ## File path: include/tvm/runtime/logging.h ## @@ -409,6 +416,68 @@ inline bool DebugLoggingEnabled() { return state == 1; } +/*! \brief Helpers for

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709539416 ## File path: include/tvm/runtime/logging.h ## @@ -395,10 +399,13 @@ class LogMessageVoidify { inline bool DebugLoggingEnabled() { static int state =

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709538852 ## File path: src/runtime/logging.cc ## @@ -166,10 +167,127 @@ namespace tvm { namespace runtime { namespace detail { +std::unordered_map

[GitHub] [tvm] areusch commented on a change in pull request #8960: [Hexagon] Add contrib tests for blocked conv2d and maxpool2d

areusch commented on a change in pull request #8960: URL: https://github.com/apache/tvm/pull/8960#discussion_r709528557 ## File path: tests/python/contrib/test_hexagon/test_conv2d_blocked.py ## @@ -0,0 +1,473 @@ +# Licensed to the Apache Software Foundation (ASF) under one +#

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709537682 ## File path: include/tvm/runtime/logging.h ## @@ -129,8 +132,9 @@ * a = ... * b = ... * // if quit_on_assertion is true, if a==b, continue,

[GitHub] [tvm] mbs-octoml commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

mbs-octoml commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709537186 ## File path: CMakeLists.txt ## @@ -575,7 +575,7 @@ endif() # Create the `cpptest` target if we can find GTest. If not, we create dummy # targets that

[GitHub] [tvm] masahi commented on pull request #8985: [ONNX] Add Einsum converter

masahi commented on pull request #8985: URL: https://github.com/apache/tvm/pull/8985#issuecomment-920338681 Thanks @anwang2009 @AndrewZhaoLuo @mbrookhart @junrushao1994 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and

[tvm] branch main updated (2aebd33 -> e44f6c0)

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 2aebd33 Add standalone_crt/ to be part of the wheel package, when available. (#9005) add e44f6c0 [ONNX] Add

[GitHub] [tvm] masahi merged pull request #8985: [ONNX] Add Einsum converter

masahi merged pull request #8985: URL: https://github.com/apache/tvm/pull/8985 -- 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 a change in pull request #8985: [ONNX] Add Einsum converter

masahi commented on a change in pull request #8985: URL: https://github.com/apache/tvm/pull/8985#discussion_r709533999 ## File path: python/tvm/relay/op/strategy/cuda.py ## @@ -1210,3 +1210,16 @@ def invert_permutation_strategy_cuda(attrs, inputs, out_type, target):

[GitHub] [tvm] mbs-octoml commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

mbs-octoml commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709531817 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] sergey-grovety commented on pull request #8990: [microTVM] Update support for ARMv7m intrinsic

sergey-grovety commented on pull request #8990: URL: https://github.com/apache/tvm/pull/8990#issuecomment-920335684 > for the test image maybe you could reuse this image: https://github.com/apache/tvm/blob/main/tests/micro/testdata/mnist/digit-2.jpg Yes, sure. -- This is an

[GitHub] [tvm] mbs-octoml commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

mbs-octoml commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709531817 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] sergey-grovety commented on a change in pull request #8990: [microTVM] Update support for ARMv7m intrinsic

sergey-grovety commented on a change in pull request #8990: URL: https://github.com/apache/tvm/pull/8990#discussion_r709531309 ## File path: tests/micro/zephyr/test_zephyr_armv7m.py ## @@ -0,0 +1,293 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [tvm] mbs-octoml opened a new issue #9022: [Bug] BuiltinLower does not use alloca for storage on kDLCPU target devices

mbs-octoml opened a new issue #9022: URL: https://github.com/apache/tvm/issues/9022 https://github.com/apache/tvm/blob/2aebd3335d89bb32d330b0f851ddaf2d551fc56e/src/tir/transforms/lower_tvm_builtin.cc#L115 This particular code has a complex history but the upshot is we need finer

[GitHub] [tvm] sergey-grovety commented on a change in pull request #8990: [microTVM] Update support for ARMv7m intrinsic

sergey-grovety commented on a change in pull request #8990: URL: https://github.com/apache/tvm/pull/8990#discussion_r709527891 ## File path: apps/microtvm/reference-vm/zephyr/base-box/base_box_test.sh ## @@ -37,3 +37,5 @@ if [ $board == "stm32f746xx" ]; then else pytest

[GitHub] [tvm] areusch commented on a change in pull request #8423: Implementation of relay_to_tir target hook

areusch commented on a change in pull request #8423: URL: https://github.com/apache/tvm/pull/8423#discussion_r709518575 ## File path: cmake/modules/contrib/ExampleTargetHooks.cmake ## @@ -0,0 +1,19 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [tvm] areusch commented on a change in pull request #8423: Implementation of relay_to_tir target hook

areusch commented on a change in pull request #8423: URL: https://github.com/apache/tvm/pull/8423#discussion_r709518464 ## File path: cmake/modules/contrib/ExampleTargetHooks.cmake ## @@ -0,0 +1,19 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #8971: [Onnx] Fix NLL Loss tests

AndrewZhaoLuo commented on pull request #8971: URL: https://github.com/apache/tvm/pull/8971#issuecomment-920304017 Ok folks, I've removed the controversial changes and did an alternate work around. PTAL when you have time. -- This is an automated message from the Apache Git Service. To

[GitHub] [tvm] huajsj commented on a change in pull request #9012: [Relay] VLOG for finer grained control of hyper-detailed logging

huajsj commented on a change in pull request #9012: URL: https://github.com/apache/tvm/pull/9012#discussion_r709444137 ## File path: include/tvm/runtime/logging.h ## @@ -395,10 +399,13 @@ class LogMessageVoidify { inline bool DebugLoggingEnabled() { static int state = 0;

[GitHub] [tvm] icemelon commented on pull request #9021: [TOPI] Fix more pooling schedule

icemelon commented on pull request #9021: URL: https://github.com/apache/tvm/pull/9021#issuecomment-920295600 cc @comaniac @junrushao1994 -- 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] icemelon opened a new pull request #9021: [TOPI] Fix more pooling schedule

icemelon opened a new pull request #9021: URL: https://github.com/apache/tvm/pull/9021 It's a follow-up update for #8957. In previous PR, I forgot to other pooling schedule in cuda and x86. -- This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [tvm] comaniac opened a new pull request #9020: [Community] @AndrewZhaoLuo -> Reviewer

comaniac opened a new pull request #9020: URL: https://github.com/apache/tvm/pull/9020 Please join us to welcome @AndrewZhaoLuo as a new reviewer to TVM. Andrew has mainly contributed to automatic mixed precision (AMP) support and ONNX frontend improvements. - [Commits

[tvm] branch comaniac-patch-1 created (now afb3c28)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch comaniac-patch-1 in repository https://gitbox.apache.org/repos/asf/tvm.git. at afb3c28 [Community] @AndrewZhaoLuo -> Reviewer This branch includes the following new commits: new

[tvm] 01/01: [Community] @AndrewZhaoLuo -> Reviewer

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a commit to branch comaniac-patch-1 in repository https://gitbox.apache.org/repos/asf/tvm.git commit afb3c289a6ee1c74d6d850248e168bc6c04c051b Author: Cody Yu AuthorDate: Wed Sep 15 11:42:57 2021 -0700

[GitHub] [tvm] manupa-arm commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

manupa-arm commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709445247 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] manupa-arm commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

manupa-arm commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709445247 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] manupa-arm commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

manupa-arm commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709445247 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] manupa-arm commented on issue #8978: [Bug][VTA][OpenCL] If allowed to allocate in stack, VTA multiple target test will fail

manupa-arm commented on issue #8978: URL: https://github.com/apache/tvm/issues/8978#issuecomment-920268526 Also for some reason, if we really need the tir.allocate to be translated down to as a stack placement, we should probably make the tir.allocate with storage_scope = "local" to

[GitHub] [tvm] kparzysz-quic commented on a change in pull request #8986: [Hexagon] Implement model launcher

kparzysz-quic commented on a change in pull request #8986: URL: https://github.com/apache/tvm/pull/8986#discussion_r709459893 ## File path: src/runtime/hexagon/launcher/README.md ## @@ -0,0 +1,173 @@ + + + + + + + + + + + + + + + + +# Hexagon Graph Launcher + +## Compilation +

[GitHub] [tvm] mehrdadh commented on a change in pull request #8990: [microTVM] Update support for ARMv7m intrinsic

mehrdadh commented on a change in pull request #8990: URL: https://github.com/apache/tvm/pull/8990#discussion_r709458696 ## File path: tests/micro/zephyr/test_zephyr_armv7m.py ## @@ -0,0 +1,293 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [tvm] kparzysz-quic commented on a change in pull request #8986: [Hexagon] Implement model launcher

kparzysz-quic commented on a change in pull request #8986: URL: https://github.com/apache/tvm/pull/8986#discussion_r709457724 ## File path: src/runtime/hexagon/launcher/README.md ## @@ -0,0 +1,173 @@ + + + + + + + + + + + + + + + + +# Hexagon Graph Launcher + +## Compilation +

[GitHub] [tvm] kparzysz-quic commented on a change in pull request #8986: [Hexagon] Implement model launcher

kparzysz-quic commented on a change in pull request #8986: URL: https://github.com/apache/tvm/pull/8986#discussion_r709450729 ## File path: src/runtime/hexagon/launcher/README.md ## @@ -0,0 +1,173 @@ + + + + + + + + + + + + + + + + +# Hexagon Graph Launcher + +## Compilation +

[GitHub] [tvm] manupa-arm commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

manupa-arm commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709445247 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); //

[GitHub] [tvm] csullivan commented on a change in pull request #8986: [Hexagon] Implement model launcher

csullivan commented on a change in pull request #8986: URL: https://github.com/apache/tvm/pull/8986#discussion_r709380322 ## File path: src/runtime/hexagon/launcher/README.md ## @@ -0,0 +1,173 @@ + + + + + + + + + + + + + + + + +# Hexagon Graph Launcher + +## Compilation +

[GitHub] [tvm] jroesch commented on a change in pull request #8849: [5/6] Arm(R) Ethos(TM)-U NPU codegen integration

jroesch commented on a change in pull request #8849: URL: https://github.com/apache/tvm/pull/8849#discussion_r709441749 ## File path: src/tir/transforms/lower_tvm_builtin.cc ## @@ -113,16 +113,6 @@ class BuiltinLower : public StmtExprMutator { op = stmt.as(); // Get

[tvm] branch main updated (98ecefb -> 2aebd33)

This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 98ecefb [Relay] Remove memory planing from LowerTEPass (#8974) add 2aebd33 Add standalone_crt/ to be part of the

[GitHub] [tvm] areusch commented on a change in pull request #8990: [microTVM] Update support for ARMv7m intrinsic

areusch commented on a change in pull request #8990: URL: https://github.com/apache/tvm/pull/8990#discussion_r709435990 ## File path: tests/micro/zephyr/test_zephyr_armv7m.py ## @@ -0,0 +1,293 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more

[GitHub] [tvm] jroesch merged pull request #9005: Add standalone_crt/ to be part of the wheel package, when available.

jroesch merged pull request #9005: URL: https://github.com/apache/tvm/pull/9005 -- 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] jroesch commented on pull request #9005: Add standalone_crt/ to be part of the wheel package, when available.

jroesch commented on pull request #9005: URL: https://github.com/apache/tvm/pull/9005#issuecomment-920244206 @manupa-arm I agree we can land but might be worth revisiting, from my perspective library should only point to `.so`, `.dylib`, and so on. I think it would be good to at minimum

[GitHub] [tvm] areusch commented on a change in pull request #8990: [microTVM] Update support for ARMv7m intrinsic

areusch commented on a change in pull request #8990: URL: https://github.com/apache/tvm/pull/8990#discussion_r709432323 ## File path: apps/microtvm/reference-vm/zephyr/base-box/base_box_test.sh ## @@ -37,3 +37,5 @@ if [ $board == "stm32f746xx" ]; then else pytest

[GitHub] [tvm] vinx13 commented on pull request #8983: [Bugfix] Fix other div zero errors also in rewrite_simplify

vinx13 commented on pull request #8983: URL: https://github.com/apache/tvm/pull/8983#issuecomment-920236818 There are some flaky tests on CI, could you push again to retrigger CI? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to

[GitHub] [tvm] AndrewZhaoLuo commented on a change in pull request #9017: [ONNX] QLinearAveragePool and QLinearGlobalAveragePool contrib op

AndrewZhaoLuo commented on a change in pull request #9017: URL: https://github.com/apache/tvm/pull/9017#discussion_r709416154 ## File path: python/tvm/relay/frontend/onnx.py ## @@ -654,6 +676,40 @@ def _impl_v1(cls, inputs, attr, params): ) +class

[GitHub] [tvm] manupa-arm commented on pull request #9005: Add standalone_crt/ to be part of the wheel package, when available.

manupa-arm commented on pull request #9005: URL: https://github.com/apache/tvm/pull/9005#issuecomment-920228506 @areusch Shall we merge this ? I think we agree that this step is required eitherway. -- This is an automated message from the Apache Git Service. To respond to the

[tvm] branch main updated (57386a2 -> 98ecefb)

This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 57386a2 [Hexagon] Treat floats as float32 when passing args to offloaded kernels (#9010) add 98ecefb [Relay]

  1   2   >