[GitHub] [tvm] masahi commented on a diff in pull request #12911: [Hexagon] vrmpy tensorization for e2e compilation of int8 models

masahi commented on code in PR #12911: URL: https://github.com/apache/tvm/pull/12911#discussion_r980779348 ## python/tvm/topi/hexagon/conv2d_alter_op.py: ## @@ -0,0 +1,111 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license

[GitHub] [tvm] sfvaroglu commented on pull request #12906: [ONNX] Fix test_roi_align failure

sfvaroglu commented on PR #12906: URL: https://github.com/apache/tvm/pull/12906#issuecomment-1258983241 Please review @driazati, 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 above to go to the

[tvm] branch last-successful updated (830ebc4ec8 -> 5ddd35c377)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from 830ebc4ec8 [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse (#12904) add 5ddd35c377

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258872092 BTW. Special kudos to @lhotari for flagging it and to @tobked for the dashboard we have :) -- This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258871699 Thanks for super-quick reaction @Mousius @gigiblender ! I also asked INFRA to see if we can work out some kind of escape hatch in case similar thing happens in the future (not

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258870347 We are back in business :) https://user-images.githubusercontent.com/595491/192414685-a085d44c-c6bb-4a91-a998-830a40c915e4.png;> -- This is an automated message from the

[GitHub] [tvm] vinx13 opened a new pull request, #12913: [TOPI] Implement Einsum with reduction axes

vinx13 opened a new pull request, #12913: URL: https://github.com/apache/tvm/pull/12913 The current Einsum doesn't utilize `ReduceOp` for reduction. It instead unrolls the whole reduction when defining the computation. It makes the generated result not suitable for analysis as it doesn't

[tvm] branch last-successful updated (f25a702a1f -> 830ebc4ec8)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from f25a702a1f [TOPI][Hexagon] Add schedule and test for maxpool uint8 layout (#12826) add d4fb957ae1

[tvm] branch nightly-docker-update updated (bc655e8b66 -> 60752ba428)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch nightly-docker-update in repository https://gitbox.apache.org/repos/asf/tvm.git discard bc655e8b66 [ci][docker] Nightly Docker image update add a61c1ad0f0 [TIR] Fix plan buffer

[GitHub] [tvm] janetsc commented on a diff in pull request #12905: [Hexagon] [runtime] Add thread manager to resource management

janetsc commented on code in PR #12905: URL: https://github.com/apache/tvm/pull/12905#discussion_r980606854 ## src/runtime/hexagon/hexagon_device_api.h: ## @@ -164,6 +176,12 @@ class HexagonDeviceAPI final : public DeviceAPI { //! \brief Current buffer manager

[GitHub] [tvm] janetsc commented on a diff in pull request #12905: [Hexagon] [runtime] Add thread manager to resource management

janetsc commented on code in PR #12905: URL: https://github.com/apache/tvm/pull/12905#discussion_r980599945 ## src/runtime/hexagon/hexagon_device_api.h: ## @@ -54,17 +55,26 @@ class HexagonDeviceAPI final : public DeviceAPI { void AcquireResources() {

[tvm] branch main updated (5ddd35c377 -> 4d5ed07325)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 5ddd35c377 [Relay][TE] Add default param name if needed (#12912) add 4d5ed07325 [TIR] Fix GetProducer/Consumer for

[GitHub] [tvm] junrushao merged pull request #12910: [TIR] Fix GetProducer/Consumer for duplicating dep edges

junrushao merged PR #12910: URL: https://github.com/apache/tvm/pull/12910 -- 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 (830ebc4ec8 -> 5ddd35c377)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 830ebc4ec8 [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse (#12904) add 5ddd35c377 [Relay][TE]

[GitHub] [tvm] junrushao merged pull request #12912: [Relay][TE] Add default param name if needed

junrushao merged PR #12912: URL: https://github.com/apache/tvm/pull/12912 -- 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 last-successful updated (e1f3f90588 -> f25a702a1f)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from e1f3f90588 [TOPI][Hexagon] Implement quantize op for hexagon (#12820) add f25a702a1f [TOPI][Hexagon]

[GitHub] [tvm] JosephTheOctonaut commented on a diff in pull request #12905: [Hexagon] [runtime] Add thread manager to resource management

JosephTheOctonaut commented on code in PR #12905: URL: https://github.com/apache/tvm/pull/12905#discussion_r980583486 ## src/runtime/hexagon/hexagon_device_api.h: ## @@ -164,6 +176,12 @@ class HexagonDeviceAPI final : public DeviceAPI { //! \brief Current buffer manager

[tvm] branch last-successful updated (fd26813723 -> e1f3f90588)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from fd26813723 [TVMScript] Infer T.match_buffer parameters for region (#12890) add e1f3f90588

[GitHub] [tvm] areusch commented on a diff in pull request #12818: [microTVM] Use default Project Options in template projects and add Makefile for Arduino template project

areusch commented on code in PR #12818: URL: https://github.com/apache/tvm/pull/12818#discussion_r980560302 ## tests/micro/project_api/test_project_api.py: ## @@ -0,0 +1,104 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license

[GitHub] [tvm] junrushao commented on pull request #12904: [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse

junrushao commented on PR #12904: URL: https://github.com/apache/tvm/pull/12904#issuecomment-1258712551 Oh ooops! Sorry I merged it in prematurely without a regression test...Feel free to add one as a follow-up! -- This is an automated message from the Apache Git Service. To respond to

[tvm] branch main updated (d4fb957ae1 -> 830ebc4ec8)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from d4fb957ae1 [microTVM][ARM] Improve dense DSP micro kernel (#12908) add 830ebc4ec8 [TIR] Refactor IndexMap::Inverse

[GitHub] [tvm] junrushao merged pull request #12904: [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse

junrushao merged PR #12904: URL: https://github.com/apache/tvm/pull/12904 -- 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 commented on pull request #12880: [usmp] Also remap VarNode to USMP-allocated buffer

areusch commented on PR #12880: URL: https://github.com/apache/tvm/pull/12880#issuecomment-1258706889 @manupak great, deleted the LOG lines and added a test. please take a look! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[tvm] branch main updated (f25a702a1f -> d4fb957ae1)

This is an automated email from the ASF dual-hosted git repository. areusch pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from f25a702a1f [TOPI][Hexagon] Add schedule and test for maxpool uint8 layout (#12826) add d4fb957ae1 [microTVM][ARM]

[GitHub] [tvm] areusch merged pull request #12908: [microTVM][ARM] Improve dense DSP micro kernel

areusch merged PR #12908: URL: https://github.com/apache/tvm/pull/12908 -- 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 pull request #12863: [TIR][Arith] Prove conditionals by transitively applying knowns

Lunderberg commented on PR #12863: URL: https://github.com/apache/tvm/pull/12863#issuecomment-1258687309 Current failures are timeouts during `apps/microtvm/ethosu/run_demo.sh`. The demo passes locally, but the runtime has increased from 1m22s to 5m48s, which would make it trip the

[tvm] branch last-successful updated (8711ba44b9 -> fd26813723)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from 8711ba44b9 [TVMScript] Import TIR methods into the IRBuilder (#12900) add fd26813723 [TVMScript] Infer

[GitHub] [tvm] comaniac opened a new pull request, #12912: [Relay][TE] Add default param name if needed

comaniac opened a new pull request, #12912: URL: https://github.com/apache/tvm/pull/12912 #10516 used the Relay parameter name when lowering to TE. However, this creates an issue when the parameter name is empty. This is legal in Relay, but results in errors during code generation. For

[tvm] branch fix_placeholder created (now 2677a7c536)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch fix_placeholder in repository https://gitbox.apache.org/repos/asf/tvm.git at 2677a7c536 [Relay][TE] Add default param name if needed This branch includes the following new commits:

[tvm] 01/01: [Relay][TE] Add default param name if needed

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a commit to branch fix_placeholder in repository https://gitbox.apache.org/repos/asf/tvm.git commit 2677a7c536078595e28eb4f58b834ca940d50882 Author: Cody Yu AuthorDate: Mon Sep 26 20:47:30 2022 +

[tvm] branch main updated (e1f3f90588 -> f25a702a1f)

This is an automated email from the ASF dual-hosted git repository. kparzysz pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from e1f3f90588 [TOPI][Hexagon] Implement quantize op for hexagon (#12820) add f25a702a1f [TOPI][Hexagon] Add schedule

[GitHub] [tvm] kparzysz-quic merged pull request #12826: [TOPI][Hexagon] Add schedule and test for maxpool uint8 layout

kparzysz-quic merged PR #12826: URL: https://github.com/apache/tvm/pull/12826 -- 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 (fd26813723 -> e1f3f90588)

This is an automated email from the ASF dual-hosted git repository. kparzysz pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from fd26813723 [TVMScript] Infer T.match_buffer parameters for region (#12890) add e1f3f90588 [TOPI][Hexagon]

[GitHub] [tvm] kparzysz-quic merged pull request #12820: [TOPI][Hexagon] Implement quantize op for hexagon

kparzysz-quic merged PR #12820: URL: https://github.com/apache/tvm/pull/12820 -- 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] junrushao commented on a diff in pull request #12910: [TIR] Fix GetProducer/Consumer for duplicating dep edges

junrushao commented on code in PR #12910: URL: https://github.com/apache/tvm/pull/12910#discussion_r980463447 ## src/tir/schedule/primitive/get_block_loop.cc: ## @@ -80,27 +80,27 @@ Array GetChildBlocks(const ScheduleState& self, const StmtSRef& parent Array

[GitHub] [tvm] masahi opened a new pull request, #12911: [Hexagon] vrmpy tensorization for e2e compilation of int8 models

masahi opened a new pull request, #12911: URL: https://github.com/apache/tvm/pull/12911 This PR lets us run int8 resnet50 in 146 msec. All convolutions and the final dense op are tensorized. The current bottleneck is `requantize`-related operations. -- This is an automated

[tvm] branch last-successful updated (4ef1465d40 -> 8711ba44b9)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from 4ef1465d40 [skip ci] Temporarily disable comments bot (#12903) add b6a660be58 [BUILD] Re-enable ccache

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258508500 Down to 64% for today :)  -- This is an automated message from

[GitHub] [tvm] trahman-quic commented on pull request #12820: [TOPI][Hexagon] Implement quantize op for hexagon

trahman-quic commented on PR #12820: URL: https://github.com/apache/tvm/pull/12820#issuecomment-1258484311 @kparzysz-quic @mehrdadh @jverma-quic If the PR looks good, can we please merge it? -- This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [tvm] trahman-quic commented on pull request #12826: [TOPI][Hexagon] Add schedule and test for maxpool uint8 layout

trahman-quic commented on PR #12826: URL: https://github.com/apache/tvm/pull/12826#issuecomment-1258484508 @kparzysz-quic @mehrdadh @jverma-quic If the PR looks good, can we please merge it? -- This is an automated message from the Apache Git Service. To respond to the message, please

[tvm] branch main updated: [TVMScript] Infer T.match_buffer parameters for region (#12890)

This is an automated email from the ASF dual-hosted git repository. wuwei pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/main by this push: new fd26813723 [TVMScript] Infer T.match_buffer parameters

[GitHub] [tvm] vinx13 merged pull request #12890: [TVMScript] Infer T.match_buffer parameters for region

vinx13 merged PR #12890: URL: https://github.com/apache/tvm/pull/12890 -- 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] YuchenJin commented on a diff in pull request #89: [RFC] Relax Upstreaming

YuchenJin commented on code in PR #89: URL: https://github.com/apache/tvm-rfcs/pull/89#discussion_r980379409 ## rfcs/0089-relax-upstreaming.md: ## @@ -0,0 +1,701 @@ +- Feature Name: Relax Upstreaming +- Start Date: 2022-08-17 +- RFC PR: 

[GitHub] [tvm] areusch commented on issue #12832: [release] v0.10.0 Release Schedule

areusch commented on issue #12832: URL: https://github.com/apache/tvm/issues/12832#issuecomment-1258419300 I'm hoping to land https://github.com/apache/tvm/pull/12880 shortly, will update this thread when it does so we can cherry-pick it. -- This is an automated message from the Apache

[GitHub] [tvm] vinx13 opened a new pull request, #12910: [TIR] Fix GetProducer/Consumer for duplicating dep edges

vinx13 opened a new pull request, #12910: URL: https://github.com/apache/tvm/pull/12910 cc @junrushao -- 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,

[GitHub] [tvm] driazati opened a new pull request, #12909: [ci] Initialize git during deploys

driazati opened a new pull request, #12909: URL: https://github.com/apache/tvm/pull/12909 We rely on some utilities scripts in the deploy steps so they also need a git checkout. Fixes issues like https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/main/4345/pipeline --

[GitHub] [tvm] cconvey commented on a diff in pull request #12794: [TIR] add unit-tests for upcoming primfunc-slicing

cconvey commented on code in PR #12794: URL: https://github.com/apache/tvm/pull/12794#discussion_r980317496 ## tests/python/unittest/test_slice_tir.py: ## @@ -0,0 +1,178 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements.

[GitHub] [tvm] mkatanbaf commented on pull request #12495: [microTVM] add the option to open a saved micro project for debugging

mkatanbaf commented on PR #12495: URL: https://github.com/apache/tvm/pull/12495#issuecomment-1258367564 @tvm-bot rerun -- 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] mehrdadh opened a new pull request, #12908: [microTVM][ARM] Improve dense DSP micro kernel

mehrdadh opened a new pull request, #12908: URL: https://github.com/apache/tvm/pull/12908 Currently, dense DSP schedule has a hard coded condition where it forces the schedule to not to use the SIMD instructions. Our investigations on using this schedule with anomaly detection model (which

[GitHub] [tvm] junrushao commented on issue #12907: [Bug] Metaschedule ThreadedTraceApply::Apply failed with error ScheduleError: An error occurred in the schedule primitive 'parallel'

junrushao commented on issue #12907: URL: https://github.com/apache/tvm/issues/12907#issuecomment-1258341639 Thanks! Uninitialized buffer is not supported by TIR as it’s out of the scope of TE compute. We might want to generate a better warning message for such cases -- This is an

[tvm] branch main updated: [TVMScript] Import TIR methods into the IRBuilder (#12900)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/main by this push: new 8711ba44b9 [TVMScript] Import TIR methods into the

[GitHub] [tvm] junrushao merged pull request #12900: [TVMScript] Import TIR methods into the IRBuilder

junrushao merged PR #12900: URL: https://github.com/apache/tvm/pull/12900 -- 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] tkonolige commented on issue #12907: [Bug] Metaschedule ThreadedTraceApply::Apply failed with error ScheduleError: An error occurred in the schedule primitive 'parallel'

tkonolige commented on issue #12907: URL: https://github.com/apache/tvm/issues/12907#issuecomment-1258339135 In this case, `C` is already initialized by the user. I don't want to zero it in this TIR function. -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] junrushao commented on pull request #12904: [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse

junrushao commented on PR #12904: URL: https://github.com/apache/tvm/pull/12904#issuecomment-1258337915 CC @vinx13 @zxybazh -- 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.

[GitHub] [tvm] junrushao commented on issue #12907: [Bug] Metaschedule ThreadedTraceApply::Apply failed with error ScheduleError: An error occurred in the schedule primitive 'parallel'

junrushao commented on issue #12907: URL: https://github.com/apache/tvm/issues/12907#issuecomment-1258337167 Looking a closer look at the TVMScript you provided - looks like you forgot to write an init block for the reduction, which is the cause of the problem -- This is an automated

[GitHub] [tvm] driazati commented on a diff in pull request #12898: Issue Templates Edits

driazati commented on code in PR #12898: URL: https://github.com/apache/tvm/pull/12898#discussion_r980283143 ## .github/ISSUE_TEMPLATE/bug-report.md: ## @@ -1,8 +1,8 @@ --- name: "\U0001F41B Bug report" -about: To help the developer act on the issues, please include a

[tvm] branch main updated (4ef1465d40 -> b6a660be58)

This is an automated email from the ASF dual-hosted git repository. areusch pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 4ef1465d40 [skip ci] Temporarily disable comments bot (#12903) add b6a660be58 [BUILD] Re-enable ccache by default

[GitHub] [tvm] areusch merged pull request #12839: [BUILD] Re-enable ccache by default

areusch merged PR #12839: URL: https://github.com/apache/tvm/pull/12839 -- 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] tkonolige opened a new issue, #12907: [Bug] Metaschedule ThreadedTraceApply::Apply failed with error ScheduleError: An error occurred in the schedule primitive 'parallel'

tkonolige opened a new issue, #12907: URL: https://github.com/apache/tvm/issues/12907 Running this script: ``` import tvm import tvm.meta_schedule from tvm.script import tir as T @T.prim_func def matmul( A: T.Buffer[(512, 512), "float32"], B:

[GitHub] [tvm] driazati commented on pull request #12869: [ONNX] Fix test_roi_align failure

driazati commented on PR #12869: URL: https://github.com/apache/tvm/pull/12869#issuecomment-1258303312 Since we need to update the image to fix the test and fix the test to update the image, https://github.com/apache/tvm/tree/sevin/test_roi_align_fix rolls them all into one (CI jobs that

[GitHub] [tvm] Lunderberg commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Lunderberg commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1258300786 > I figured that was why it was in draft, being a bit nosey as I remember wondering about it when I was writing MakeUnpackedAPI :smile_cat: No worries at all, and I wouldn't

[GitHub] [tvm] Lunderberg commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Lunderberg commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1258298073 > The fix makes sense to me, can we either make use of

[GitHub] [tvm] driazati commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

driazati commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258284106 Thanks everyone for flagging this and disabling the job quickly! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and

[GitHub] [tvm] Lunderberg commented on a diff in pull request #12892: [TIR][Transform] Remove num_unpacked_args from MakePackedAPI

Lunderberg commented on code in PR #12892: URL: https://github.com/apache/tvm/pull/12892#discussion_r980232185 ## python/tvm/tir/transform/transform.py: ## @@ -402,7 +402,9 @@ def MakePackedAPI(num_unpacked_params: int = -1): fpass : tvm.transform.Pass The result

[GitHub] [tvm] janetsc commented on a diff in pull request #12905: [Hexagon] [runtime] Add thread manager to resource management

janetsc commented on code in PR #12905: URL: https://github.com/apache/tvm/pull/12905#discussion_r980227500 ## src/runtime/hexagon/hexagon_device_api.h: ## @@ -164,6 +176,12 @@ class HexagonDeviceAPI final : public DeviceAPI { //! \brief Current buffer manager

[GitHub] [tvm] janetsc opened a new pull request, #12905: [Hexagon] [runtime] Add thread manager to resource management

janetsc opened a new pull request, #12905: URL: https://github.com/apache/tvm/pull/12905 Adds a thread manager to Acquire/Release Resources. Added a small unit test. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

[GitHub] [tvm] Mousius commented on a diff in pull request #12892: [TIR][Transform] Remove num_unpacked_args from MakePackedAPI

Mousius commented on code in PR #12892: URL: https://github.com/apache/tvm/pull/12892#discussion_r980211903 ## python/tvm/tir/transform/transform.py: ## @@ -402,7 +402,9 @@ def MakePackedAPI(num_unpacked_params: int = -1): fpass : tvm.transform.Pass The result

[GitHub] [tvm] Mousius commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Mousius commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1258232368 > To me, the `buffer_map` is an indicator for how a function should later be lowered, and what buffers should be defined as part of the lowered function. Once it has been used, then it

[GitHub] [tvm] Lunderberg commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Lunderberg commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1258229042 @Mousius Can you check the latest commit on this PR? I didn't try to re-order the passes, but reading from the `BufferStore`/`BufferLoad` objects instead of the `buffer_map` resolved

[GitHub] [tvm] Lunderberg commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Lunderberg commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1258215543 @Mousius Thank you, and it looks like that's also being caught in the tests, as this breaks pretty much every ethos-u test at [this

[GitHub] [tvm] junrushao commented on issue #12852: [Bug] Meta Schedule Layout Rewrite Failure

junrushao commented on issue #12852: URL: https://github.com/apache/tvm/issues/12852#issuecomment-1258195689 > @tkonolige I tested it, and had the same behavior. I was able to get better error messages by changing [this

[GitHub] [tvm] Lunderberg commented on pull request #12892: [TIR][Transform] Remove num_unpacked_args from MakePackedAPI

Lunderberg commented on PR #12892: URL: https://github.com/apache/tvm/pull/12892#issuecomment-1258096805 @Mousius No problem! As far as I could tell from inspection (and maybe @tqchen could verify), it looks like an early-development feature that dates back at least to 2017

[GitHub] [tvm] ashutosh-arm commented on a diff in pull request #12887: [ETHOSN] Support conversion of add/mul to requantize where possible

ashutosh-arm commented on code in PR #12887: URL: https://github.com/apache/tvm/pull/12887#discussion_r980029874 ## python/tvm/relay/op/contrib/ethosn.py: ## @@ -227,11 +227,13 @@ def qnn_add_pattern(): is_constant(), is_constant(), ) -

[GitHub] [tvm] Lunderberg commented on issue #12852: [Bug] Meta Schedule Layout Rewrite Failure

Lunderberg commented on issue #12852: URL: https://github.com/apache/tvm/issues/12852#issuecomment-1258062633 @tkonolige I tested it, and had the same behavior. I was able to get better error messages by changing [this

[GitHub] [tvm] Lunderberg opened a new pull request, #12904: [TIR] Refactor IndexMap::Inverse in terms of NonSurjectiveInverse

Lunderberg opened a new pull request, #12904: URL: https://github.com/apache/tvm/pull/12904 The two implementations were largely identical, and had implementations that drifted apart, resulting in bugs such as https://github.com/apache/tvm/issues/12852. This commit removes this

[tvm] branch last-successful updated (87085b0e0d -> 4ef1465d40)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from 87085b0e0d [frontend][pytorch]Support aten::Tensor_split operator (#12871) add 4ef1465d40 [skip ci]

[GitHub] [tvm] Lunderberg commented on issue #12852: [Bug] Meta Schedule Layout Rewrite Failure

Lunderberg commented on issue #12852: URL: https://github.com/apache/tvm/issues/12852#issuecomment-1258024355 > In general, IndexMap::Inverse doesn’t guarantee to generate the inverse map even if the index map is (theoretically) bijective because the underlying DetectIterMap can only

[GitHub] [tvm] Mousius commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

Mousius commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1258009414 @gigiblender / @potiuk I've started to take care of the remaining jobs  -- This is an automated message from the Apache Git Service. To respond to the message, please log on to

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257957063 Thanks. I see that it already started to drop slighlty - > I see that it's down from 87 % of today's workflows to 80% -- This is an automated message from the Apache Git Service. To

[tvm] branch last-successful updated (46ea2ed42e -> 87085b0e0d)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git from 46ea2ed42e [MetaSchedule][UX] User Interface for Jupyter Notebook (#12866) add cc6e01edc6

[GitHub] [tvm] gigiblender commented on pull request #12903: [skip ci] Temporarily disable comments bot

gigiblender commented on PR #12903: URL: https://github.com/apache/tvm/pull/12903#issuecomment-1257859272 > > Just a comment as well - all the pending workflow should be cancelled as well, otherwise it might take days to clean the queue (I think @lhotari gave some examples how they did it

[GitHub] [tvm] gigiblender commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

gigiblender commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257857648 Thank you for raising this, and we sincerely apologize for causing this inconvenience. #12903 should disable the bot. Regarding skipping the already queued/in-progress

[tvm] branch main updated (87085b0e0d -> 4ef1465d40)

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 87085b0e0d [frontend][pytorch]Support aten::Tensor_split operator (#12871) add 4ef1465d40 [skip ci] Temporarily

[GitHub] [tvm] masahi merged pull request #12903: [skip ci] Temporarily disable comments bot

masahi merged PR #12903: URL: https://github.com/apache/tvm/pull/12903 -- 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] gigiblender commented on pull request #12903: [skip ci] Temporarily disable comments bot

gigiblender commented on PR #12903: URL: https://github.com/apache/tvm/pull/12903#issuecomment-1257828560 > Just a comment as well - all the pending workflow should be cancelled as well, otherwise it might take days to clean the queue (I think @lhotari gave some examples how they did it in

[GitHub] [tvm] potiuk commented on pull request #12903: [skip ci] Temporarily disable comments bot

potiuk commented on PR #12903: URL: https://github.com/apache/tvm/pull/12903#issuecomment-1257826465 Just a comment as well - all the pending workflow should be cancelled as well, otherwise it might take days to clean the queue (I think @lhotari gave some examples how they did it in Pulsar

[GitHub] [tvm] github-actions[bot] commented on pull request #12903: [skip ci] Temporarily disable comments bot

github-actions[bot] commented on PR #12903: URL: https://github.com/apache/tvm/pull/12903#issuecomment-1257806822 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code

[GitHub] [tvm] gigiblender opened a new pull request, #12903: [skip ci] Temporarily disable comments bot

gigiblender opened a new pull request, #12903: URL: https://github.com/apache/tvm/pull/12903 Since September 14th, 2022, tvm has been consuming about 75% of the shared Apache org GitHub Actions resources. This seems to be due to the comments bot. Check #12902 and #12695 for details.

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257796206 This one has started to heavily affect other problems (the "fairness" algorithms of GitHub likely get a very tough time with the project taking already almost 90% of the queueu

[tvm] branch main updated: [frontend][pytorch]Support aten::Tensor_split operator (#12871)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/main by this push: new 87085b0e0d [frontend][pytorch]Support

[GitHub] [tvm] masahi merged pull request #12871: [frontend][pytorch]Support aten::Tensor_split operator

masahi merged PR #12871: URL: https://github.com/apache/tvm/pull/12871 -- 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 (46ea2ed42e -> cc6e01edc6)

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 46ea2ed42e [MetaSchedule][UX] User Interface for Jupyter Notebook (#12866) add cc6e01edc6 [frontend][pytorch]support

[GitHub] [tvm] masahi merged pull request #12872: [frontend][pytorch]support aten::zero_ operator

masahi merged PR #12872: URL: https://github.com/apache/tvm/pull/12872 -- 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] Mousius commented on pull request #12892: [TIR][Transform] Remove num_unpacked_args from MakePackedAPI

Mousius commented on PR #12892: URL: https://github.com/apache/tvm/pull/12892#issuecomment-1257733229 Thanks for doing this @Lunderberg, I've often wondered what this was for 樂 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257731705 For me it looks like your comment bot triggers itself in recursive fashion :( -- This is an automated message from the Apache Git Service. To respond to the message, please log on

[GitHub] [tvm] Mousius commented on pull request #12891: [TIR][Transform] Clear buffer_map during MakeUnpackedAPI

Mousius commented on PR #12891: URL: https://github.com/apache/tvm/pull/12891#issuecomment-1257728044 @Lunderberg I think we make use of this in following passes and I'm unsure why make packed API decides to clear it? -- This is an automated message from the Apache Git Service. To

[GitHub] [tvm] potiuk commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

potiuk commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257727684 Oh yeah. It's not nice neither for your users nor for other apache projects and seems it just went out of control for you -- This is an automated message from the Apache Git

[GitHub] [tvm] masahi commented on pull request #12877: [LLVM] Emit fp16/fp32 builtins directly into target module

masahi commented on PR #12877: URL: https://github.com/apache/tvm/pull/12877#issuecomment-1257724910 Thanks @kparzysz-quic, I'll verify tomorrow. -- 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] manupak commented on pull request #12880: [usmp] Also remap VarNode to USMP-allocated buffer

manupak commented on PR #12880: URL: https://github.com/apache/tvm/pull/12880#issuecomment-1257720475 Thanks for the fix -- yea I double checked whether buffer info extraction captures this scenario as well -- it seems so. If we can get the tests and LOG(INFO)s removed -- I think we cant

[GitHub] [tvm] nicoloboschi commented on issue #12902: [CI Problem] apache/tvm consumes too much of Apache org shared GitHub Actions resources

nicoloboschi commented on issue #12902: URL: https://github.com/apache/tvm/issues/12902#issuecomment-1257719169 You could use this script to cleanup your pending actions: https://github.com/nicoloboschi/pulsar-validation-tool/blob/master/pulsar-scripts/pulsar-gha/cancel-workflows.js

  1   2   >