masahi closed pull request #11620: [CI] Disable TFLite rsqrt test
URL: https://github.com/apache/tvm/pull/11620
--
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,
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 e9b331831a [COMMUNITY] Min Chen -> Reviewer (#13628)
e9
masahi merged PR #13628:
URL: https://github.com/apache/tvm/pull/13628
--
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.apache.or
masahi commented on PR #13532:
URL: https://github.com/apache/tvm/pull/13532#issuecomment-1354323713
Found a bug https://github.com/apache/tvm/pull/13631
@shingjan This may fix your issue.
--
This is an automated message from the Apache Git Service.
To respond to the message, please
tvm-bot commented on PR #13631:
URL: https://github.com/apache/tvm/pull/13631#issuecomment-1354321344
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 reviews from
masahi opened a new pull request, #13631:
URL: https://github.com/apache/tvm/pull/13631
Since https://github.com/apache/tvm/pull/12515, const folding on fp16, when
both lhs and rhs are constant, is disabled. This is the case even when one of
the arguments are one or zero. This leads to a re
lightzhan-intellif commented on code in PR #13630:
URL: https://github.com/apache/tvm/pull/13630#discussion_r1050386527
##
tests/python/unittest/test_tvmscript_regression.py:
##
@@ -45,5 +45,20 @@ def test_multi_element_array_in_outmost_namespace():
tvm.ir.assert_structural
echuraev commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050386160
##
tests/python/unittest/test_runtime_module_based_interface.py:
##
@@ -688,6 +688,36 @@ def test_num_threads():
assert reported == hardware_threads or reported
echuraev commented on code in PR #13627:
URL: https://github.com/apache/tvm/pull/13627#discussion_r1050370999
##
docs/conf.py:
##
@@ -249,76 +343,76 @@ def git_describe_version(original_version):
# The unlisted files are sorted by filenames.
# The unlisted files always appear
junrushao commented on code in PR #13630:
URL: https://github.com/apache/tvm/pull/13630#discussion_r1050377117
##
python/tvm/script/parser/core/parser.py:
##
@@ -150,7 +150,7 @@ def add(self, var: str, value: Any, allow_shadowing: bool =
False):
The options of whet
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from cc0f27a8b1 [Relay] Remove overwriting of matmul shapes when they are
static (#13615)
add 06be0b3b26 [Frontend]
Hzfengsy commented on code in PR #13630:
URL: https://github.com/apache/tvm/pull/13630#discussion_r1050371682
##
tests/python/unittest/test_tvmscript_regression.py:
##
@@ -45,5 +45,20 @@ def test_multi_element_array_in_outmost_namespace():
tvm.ir.assert_structural_equal(fun
tvm-bot commented on PR #13630:
URL: https://github.com/apache/tvm/pull/13630#issuecomment-1354201379
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 reviews from
lightzhan-intellif opened a new pull request, #13630:
URL: https://github.com/apache/tvm/pull/13630
This PR tries to fix the crash of parser when the old value of a var is an
array but the new value is not. For example:
```python
from tvm.script import tir as T
def func_wrapper(shap
lightzhan-intellif closed pull request #13629: [BugFix][TVMScript] Parser crash
URL: https://github.com/apache/tvm/pull/13629
--
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
shingjan commented on PR #13623:
URL: https://github.com/apache/tvm/pull/13623#issuecomment-1354175289
@tvm-bot re-run
--
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 un
tvm-bot commented on PR #13629:
URL: https://github.com/apache/tvm/pull/13629#issuecomment-1354167482
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 reviews from
lightzhan-intellif opened a new pull request, #13629:
URL: https://github.com/apache/tvm/pull/13629
This PR tries to fix the crash of parser when the old value of a var is an
array but the new value is not. For example:
```python
from tvm.script import tir as T
def func_wrapper(shap
tvm-bot commented on PR #13628:
URL: https://github.com/apache/tvm/pull/13628#issuecomment-1354164079
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 reviews from
junrushao opened a new pull request, #13628:
URL: https://github.com/apache/tvm/pull/13628
Please join us to welcome @multiverstack-intellif as a new reviewer to TVM.
Min contributed key features in TIR scheduling, specifically, cache_read/write
that are aware of cache location. These are h
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 37f6aa0c7e [MetaSchedule] Fix tensorcore winograd task
masahi merged PR #13625:
URL: https://github.com/apache/tvm/pull/13625
--
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.apache.or
masahi merged PR #13613:
URL: https://github.com/apache/tvm/pull/13613
--
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.apache.or
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 7674ea84fe Add check for non-contiguous memory access w
tvm-bot commented on PR #13627:
URL: https://github.com/apache/tvm/pull/13627#issuecomment-135402
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 reviews from
guberti opened a new pull request, #13627:
URL: https://github.com/apache/tvm/pull/13627
Details to be announced at Hackathon presentation 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 t
slyubomirsky opened a new issue, #13626:
URL: https://github.com/apache/tvm/issues/13626
The following Relay program causes the debug interpreter to crash before it
even starts evaluation:
```python
import tvm
from tvm import relay
modtext = """
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 ce97138ebe [TVMScript] Fix print round-tripable multi thread env
binding (#13622)
add cdb4eea138 [TOPI][Hexagon] Im
kparzysz-quic merged PR #13614:
URL: https://github.com/apache/tvm/pull/13614
--
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.ap
tvm-bot commented on PR #13625:
URL: https://github.com/apache/tvm/pull/13625#issuecomment-1353801977
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 reviews from
masahi opened a new pull request, #13625:
URL: https://github.com/apache/tvm/pull/13625
When both AutoTVM and AS/MS winograd are applicable, the latter is never
picked because the AutoTVM has a higher plevel
https://github.com/masahi/tvm/blob/500cfcf84a4aeece12cad555cf09682073ad9de2/pyt
nverke commented on code in PR #13613:
URL: https://github.com/apache/tvm/pull/13613#discussion_r1050165016
##
src/tir/transforms/lower_async_dma.cc:
##
@@ -146,13 +153,32 @@ class AsyncDMALowerer : public StmtExprMutator {
// map loop variable to zero for the store ind
nverke commented on PR #13613:
URL: https://github.com/apache/tvm/pull/13613#issuecomment-1353742695
We are still discussing the pros and cons of the potential options but the
possible solutions are.
1. Lower the copy into n x 1D dma copies.
2. Lower to 2D dma copies.
3. Use tenso
masahi commented on PR #13613:
URL: https://github.com/apache/tvm/pull/13613#issuecomment-1353734805
> once support is added.
What support is being planed? 2D DMA or breaking up into contiguous copies?
--
This is an automated message from the Apache Git Service.
To respond to the me
masahi commented on PR #13532:
URL: https://github.com/apache/tvm/pull/13532#issuecomment-1353715208
Turns out, for winograd fp16, the issue was due to the following `cse_vars`
```
let cse_var_11 = (0h*2h)
let cse_var_10 = (0h*2.5h)
let cse_var_9 = (0h*1h)
let cse_var_8
cbalint13 commented on code in PR #13624:
URL: https://github.com/apache/tvm/pull/13624#discussion_r1050131591
##
src/relay/backend/contrib/uma/targets.cc:
##
@@ -62,16 +61,23 @@
TVM_REGISTER_GLOBAL("relay.backend.contrib.uma.RegisterTarget")
manupak commented on code in PR #13624:
URL: https://github.com/apache/tvm/pull/13624#discussion_r1050114968
##
src/relay/backend/contrib/uma/targets.cc:
##
@@ -62,16 +61,23 @@
TVM_REGISTER_GLOBAL("relay.backend.contrib.uma.RegisterTarget")
cgerum commented on PR #13624:
URL: https://github.com/apache/tvm/pull/13624#issuecomment-1353622186
Thank you very much @cbalint13 for taking over some UMA maintenance. The
patch looks fine to me.
--
This is an automated message from the Apache Git Service.
To respond to the message, ple
shingjan commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050055038
##
python/tvm/contrib/graph_executor.py:
##
@@ -195,6 +197,47 @@ def set_input(self, key=None, value=None, **params):
if val:
self.
vinx13 commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050047034
##
python/tvm/contrib/graph_executor.py:
##
@@ -195,6 +197,47 @@ def set_input(self, key=None, value=None, **params):
if val:
self._g
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 c5911a6f23 [ETHOSN] Add support for experimental compiler option
(#13410)
add ce97138ebe [TVMScript] Fix print round-t
vinx13 merged PR #13622:
URL: https://github.com/apache/tvm/pull/13622
--
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.apache.or
vinx13 commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050014668
##
python/tvm/contrib/graph_executor.py:
##
@@ -212,7 +212,17 @@ def set_input_zero_copy(self, key, value, **params):
Additional arguments
"""
shingjan commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050009039
##
python/tvm/contrib/graph_executor.py:
##
@@ -195,6 +198,49 @@ def set_input(self, key=None, value=None, **params):
if val:
self.
vinx13 commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1050002262
##
python/tvm/contrib/graph_executor.py:
##
@@ -195,6 +198,49 @@ def set_input(self, key=None, value=None, **params):
if val:
self._g
shingjan commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1049998823
##
python/tvm/contrib/graph_executor.py:
##
@@ -195,6 +198,49 @@ def set_input(self, key=None, value=None, **params):
if val:
self.
cbalint13 opened a new pull request, #13624:
URL: https://github.com/apache/tvm/pull/13624
This PR address fixes for UMA target registration.
* Fix the doc issue #13304
* Continues stalled PR #12731
---
Changes:
* Incorporates all proposed fixes from mentioned
tvm-bot commented on PR #13624:
URL: https://github.com/apache/tvm/pull/13624#issuecomment-1353517316
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 reviews from
vinx13 commented on code in PR #13623:
URL: https://github.com/apache/tvm/pull/13623#discussion_r1049990723
##
src/runtime/graph_executor/graph_executor.cc:
##
@@ -160,6 +160,16 @@ void GraphExecutor::SetInput(int index, DLTensor* data_in)
{
uint32_t eid = this->entry_id(inp
shingjan opened a new pull request, #13623:
URL: https://github.com/apache/tvm/pull/13623
As titled, this apis is in cpp but not python. Added this apis for
GraphModule and some tests.
cc: @vinx13
--
This is an automated message from the Apache Git Service.
To respond to the messa
tvm-bot commented on PR #13623:
URL: https://github.com/apache/tvm/pull/13623#issuecomment-1353504164
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 reviews from
junrushao commented on issue #8577:
URL: https://github.com/apache/tvm/issues/8577#issuecomment-1353448789
@AndrewZhaoLuo TVMScript is tested to support python 3.6-3.10, the part that
doesn’t support python 3.9 is the legacy hybrid script. We need to migrate them
to TVMScript
--
This is
adstraw commented on code in PR #13613:
URL: https://github.com/apache/tvm/pull/13613#discussion_r1049933426
##
src/tir/transforms/lower_async_dma.cc:
##
@@ -146,13 +153,32 @@ class AsyncDMALowerer : public StmtExprMutator {
// map loop variable to zero for the store in
AndrewZhaoLuo commented on issue #8577:
URL: https://github.com/apache/tvm/issues/8577#issuecomment-1353401506
Some dynamic (i.e. the shape is not known until runtime) kernels will not
work due to incompatibility with TVMscript.
--
This is an automated message from the Apache Git Service.
quic-sanirudh commented on issue #10317:
URL: https://github.com/apache/tvm/issues/10317#issuecomment-1353346179
hexagon @quic-sanirudh
metaschedule @quic-sanirudh
tir @quic-sanirudh
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
trahman-quic commented on PR #13614:
URL: https://github.com/apache/tvm/pull/13614#issuecomment-1353264453
Hi @mehrdadh, this PR passed all the checks.
--
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
tvm-bot commented on PR #13622:
URL: https://github.com/apache/tvm/pull/13622#issuecomment-1352972425
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 reviews from
wrongtest-intellif opened a new pull request, #13622:
URL: https://github.com/apache/tvm/pull/13622
Fix the following script with multiple kernel launch (each take standalone
thread binding) which is not round-trip-able.
```python
import tvm
from tvm.script import tir as T
@T
Mousius commented on PR #13606:
URL: https://github.com/apache/tvm/pull/13606#issuecomment-1352929237
> @Mousius The new function simply composes the existing interface to the
TECompiler, `LowerTECompute` and `CreatePrimFunc` that converts lowered TE
compute to TensorIR. `LowerTECompute` wa
This is an automated email from the ASF dual-hosted git repository.
ashutoshp 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 c5911a6f23 [ETHOSN] Add support for experimental com
ashutosh-arm merged PR #13410:
URL: https://github.com/apache/tvm/pull/13410
--
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.apa
tvm-bot commented on PR #13621:
URL: https://github.com/apache/tvm/pull/13621#issuecomment-1352898295
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 reviews from
vvchernov opened a new pull request, #13621:
URL: https://github.com/apache/tvm/pull/13621
It looks like despite of some avx512 intrinsics were supported (see topi/x86
and tir), they are not used during tuning model on skylake-avx512 target.
The aim is end-to-end support of Skylake X arch
leandron commented on PR #13598:
URL: https://github.com/apache/tvm/pull/13598#issuecomment-1352766843
Hi all. I think it's important to echo @gromero's comment regarding the
[Commit Message
Guideline](https://github.com/apache/tvm/blob/main/docs/contribute/pull_request.rst#commit-message-g
64 matches
Mail list logo