[tvm] branch main updated (4b67dac -> 887324f)

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 4b67dac [CUDA] Support multiple TIR-level dynamic shared memory allocations (#8571) add 887324f [TOPI][CUDA] Impro

[GitHub] [tvm] masahi merged pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

masahi merged pull request #8479: URL: https://github.com/apache/tvm/pull/8479 -- 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: commits-unsubscr...

[GitHub] [tvm] junrushao1994 opened a new pull request #8615: [M3a] Instruction and Trace

junrushao1994 opened a new pull request #8615: URL: https://github.com/apache/tvm/pull/8615 -- 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: commit

[GitHub] [tvm] vvchernov edited a comment on issue #8608: [TEST] test_lstms taking 20+ mins on CI runs

vvchernov edited a comment on issue #8608: URL: https://github.com/apache/tvm/issues/8608#issuecomment-890343844 Warnings are a part of the code in TVM. The first one is related to that LSTM with projection supported by pytorch starting from 1.8.0 version. But TVM works stably with pytorch

[GitHub] [tvm] YuchenJin edited a comment on pull request #8586: [Tutorial][Executor] Fix the usage of executors in tutorials

YuchenJin edited a comment on pull request #8586: URL: https://github.com/apache/tvm/pull/8586#issuecomment-890373873 > Hi @YuchenJin, regarding this PR, what do you think can be further improved or we should have another PR to systematically refactor the API usage? Thanks. Hi @ganl

[GitHub] [tvm] YuchenJin commented on pull request #8586: [Tutorial][Executor] Fix the usage of executors in tutorials

YuchenJin commented on pull request #8586: URL: https://github.com/apache/tvm/pull/8586#issuecomment-890373873 > Hi @YuchenJin, regarding this PR, what do you think can be further improved or we should have another PR to systematically refactor the API usage? Thanks. Hi @ganler, I t

[GitHub] [tvm] CaptainDuke commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

CaptainDuke commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680379446 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -764,6 +764,9 @@ def scatter_nd(data, indices, updates, mode): """ _verify_scatter_nd_inputs

[GitHub] [tvm] CaptainDuke commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

CaptainDuke commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680379100 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -787,44 +791,94 @@ def gen_ir(data_ptr, indices_ptr, updates_ptr, out_ptr): for i in data_ptr

[GitHub] [tvm] CaptainDuke commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

CaptainDuke commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680378831 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -787,44 +791,94 @@ def gen_ir(data_ptr, indices_ptr, updates_ptr, out_ptr): for i in data_ptr

[GitHub] [tvm] vinx13 merged pull request #8571: [CUDA] Support multiple TIR-level dynamic shared memory allocations

vinx13 merged pull request #8571: URL: https://github.com/apache/tvm/pull/8571 -- 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: commits-unsubscr...

[tvm] branch main updated (28de742 -> 4b67dac)

This is an automated email from the ASF dual-hosted git repository. wuwei pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 28de742 [Refactor] Unify the shared pass prefix between vm and graph (#8526) add 4b67dac [CUDA] Support multiple TI

[GitHub] [tvm] vvchernov commented on issue #8608: [TEST] test_lstms taking 20+ mins on CI runs

vvchernov commented on issue #8608: URL: https://github.com/apache/tvm/issues/8608#issuecomment-890343844 Warnings are a part of the code in TVM. The first one is related to that LSTM with projection supported by pytorch starting from 1.8.0 version. But TVM stable works with pytorch 1.7.0.

[GitHub] [tvm] tqchen commented on pull request #8606: [BugFix][TensorIR] `specialize()` updates the `attrs` of PrimFuncs

tqchen commented on pull request #8606: URL: https://github.com/apache/tvm/pull/8606#issuecomment-890343007 We should find a better way to implement the feature(e.g. flop counting) without relying on this pattern. Additionally, it would be great if we can add a well-form checker to ensure

[GitHub] [tvm] vvchernov commented on issue #8608: [TEST] test_lstms taking 20+ mins on CI runs

vvchernov commented on issue #8608: URL: https://github.com/apache/tvm/issues/8608#issuecomment-890342675 Hello! It was resolved in PR #8583 -- 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 main updated (7d8a774 -> 28de742)

This is an automated email from the ASF dual-hosted git repository. tqchen pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 7d8a774 [VTA] Recover rpc server support (#8604) add 28de742 [Refactor] Unify the shared pass prefix between vm and

[GitHub] [tvm] tqchen merged pull request #8526: [Refactor] Unify the shared pass prefix between vm and graph

tqchen merged pull request #8526: URL: https://github.com/apache/tvm/pull/8526 -- 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: commits-unsubscr...

[tvm] branch main updated (2a8950b -> 7d8a774)

This is an automated email from the ASF dual-hosted git repository. tqchen pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 2a8950b [TensorIR] Support for match_buffer from subregion (#8585) add 7d8a774 [VTA] Recover rpc server support (#8

[GitHub] [tvm] tqchen merged pull request #8604: [VTA] Recover rpc server support

tqchen merged pull request #8604: URL: https://github.com/apache/tvm/pull/8604 -- 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: commits-unsubscr...

[GitHub] [tvm] tqchen commented on a change in pull request #8373: [BUILD] Add caching to CMake

tqchen commented on a change in pull request #8373: URL: https://github.com/apache/tvm/pull/8373#discussion_r680354402 ## File path: CMakeLists.txt ## @@ -635,3 +635,32 @@ if(APPLE AND TVM_IS_DEBUG_BUILD) VERBATIM ) endif() + +#Caches the build. +#Note

[GitHub] [tvm] ganler commented on pull request #8586: [Tutorial][Executor] Fix the usage of executors in tutorials

ganler commented on pull request #8586: URL: https://github.com/apache/tvm/pull/8586#issuecomment-890309576 Hi @YuchenJin, regarding this PR, what do you think can be further improved or we should have another PR to systematically refactor the API usage? Thanks. -- This is an automated m

[GitHub] [tvm] MasterJH5574 commented on pull request #8606: [BugFix][TensorIR] `specialize()` updates the `attrs` of PrimFuncs

MasterJH5574 commented on pull request #8606: URL: https://github.com/apache/tvm/pull/8606#issuecomment-890307069 Currently only meta-schedule needs this feature of `specialize`. If we cannot decide whether to support it now, perhaps we can support it in TIR sub-branch, and think of other

[GitHub] [tvm] masahi commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

masahi commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680322918 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -764,6 +764,9 @@ def scatter_nd(data, indices, updates, mode): """ _verify_scatter_nd_inputs(data

[GitHub] [tvm] masahi commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

masahi commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680322863 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -787,44 +791,94 @@ def gen_ir(data_ptr, indices_ptr, updates_ptr, out_ptr): for i in data_ptr.shap

[GitHub] [tvm] masahi commented on a change in pull request #8479: [TOPI][CUDA] Improve the performance of scatter_nd

masahi commented on a change in pull request #8479: URL: https://github.com/apache/tvm/pull/8479#discussion_r680322633 ## File path: python/tvm/topi/cuda/scatter.py ## @@ -787,44 +791,94 @@ def gen_ir(data_ptr, indices_ptr, updates_ptr, out_ptr): for i in data_ptr.shap

[GitHub] [tvm-rfcs] huajsj opened a new pull request #14: pipeline compute RFC.

huajsj opened a new pull request #14: URL: https://github.com/apache/tvm-rfcs/pull/14 Split relay graph into Subgraph then doing Subgraph Pipeline :RFC In this RFC, we present a new framework for subgraph pipelining that first split relay graph into multiple subgraph/group operator s