quic-sanirudh commented on code in PR #16557:
URL: https://github.com/apache/tvm/pull/16557#discussion_r1495450159
##
src/runtime/hexagon/hexagon_device_api.h:
##
@@ -199,6 +215,9 @@ class HexagonDeviceAPI final : public DeviceAPI {
//! \brief Hexagon power manager
std::
felix-ro commented on PR #16610:
URL: https://github.com/apache/tvm/pull/16610#issuecomment-1953756108
> LGTM, but is there a reason for the `tvm.runtime.String` addition. the FFI
call should automatically handle if I'm not mistaken.
I just checked the FFI call, and it is handled ther
lhutton1 merged PR #16523:
URL: https://github.com/apache/tvm/pull/16523
--
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.
This is an automated email from the ASF dual-hosted git repository.
lukhut 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 a6157a6369 [SVE] Change the dtype of Ramp and Broadcast
quic-sanirudh commented on PR #16557:
URL: https://github.com/apache/tvm/pull/16557#issuecomment-1953795374
> Thanks @quic-sanirudh . I did another round provided all the feedbacks,
let me know if they make sense
Thanks a lot @tqchen for taking the time to review. I've made all the
c
lhutton1 commented on PR #16523:
URL: https://github.com/apache/tvm/pull/16523#issuecomment-1953777295
Thanks @ekalda @tqchen @Lunderberg
--
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 speci
lhutton1 opened a new pull request, #16612:
URL: https://github.com/apache/tvm/pull/16612
This commit adds support for representing scalable vectors using the string
data type format. For example, "float32xvscalex4" may be used to represent the
following scalable type: `DataType(kDLFloat, 3
ekalda commented on PR #16611:
URL: https://github.com/apache/tvm/pull/16611#issuecomment-1953961974
Thanks @lhutton1, would you mind adding a test for this change? There is
currently no path that exercises this option.
--
This is an automated message from the Apache Git Service.
To respo
Liam-Sturge opened a new issue, #16613:
URL: https://github.com/apache/tvm/issues/16613
### The Problem
Intermittently, when running `./build_image.sh cpu`, the function
`detect_llvm_version()` does not run correctly due to exceeding Github API rate
limits. This then causes a complet
Liam-Sturge closed issue #16613: [CI Problem] TLCPack Docker image unable to
set LLVM version
URL: https://github.com/apache/tvm/issues/16613
--
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 spec
Liam-Sturge commented on issue #16613:
URL: https://github.com/apache/tvm/issues/16613#issuecomment-1954039000
Raised issue on TLCPack repo instead:
https://github.com/tlc-pack/tlcpack/issues/190
--
This is an automated message from the Apache Git Service.
To respond to the message, pleas
lhutton1 commented on PR #18:
URL: https://github.com/apache/tvm-rfcs/pull/18#issuecomment-1954039440
Closing as superseded by: https://github.com/apache/tvm-rfcs/pull/104
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and us
lhutton1 closed pull request #18: [RFC] Adding initial SVE implementation
URL: https://github.com/apache/tvm-rfcs/pull/18
--
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 uns
Liam-Sturge closed issue #16613: [CI Problem] TLCPack Docker image unable to
set LLVM version
URL: https://github.com/apache/tvm/issues/16613
--
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 spec
quic-sanirudh commented on PR #16557:
URL: https://github.com/apache/tvm/pull/16557#issuecomment-1954267571
@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.
T
tqchen commented on PR #16557:
URL: https://github.com/apache/tvm/pull/16557#issuecomment-1954405459
happy to get it in @quic-sanirudh to unblock folks, let us followup and try
remove the phyiscal map by embedding them in buffer descriptor
--
This is an automated message from the Apache G
quic-sanirudh commented on PR #16557:
URL: https://github.com/apache/tvm/pull/16557#issuecomment-1954410751
> happy to get it in @quic-sanirudh to unblock folks, let us followup and
try remove the phyiscal map by embedding them in buffer descriptor
Thanks a lot for understanding. I'll
Lunderberg commented on PR #16588:
URL: https://github.com/apache/tvm/pull/16588#issuecomment-1955024693
Good point on checking the performance. I did a benchmark, with shown
results shown in the plot below. The x-axis is the time required to run the
analyzer with the `BoundUsingReciproca
Lunderberg commented on PR #16563:
URL: https://github.com/apache/tvm/pull/16563#issuecomment-1955086643
All CI tests passing, and thank you for the review @slyubomirsky ! I'll
follow up with another PR to add inspection of the remainder of the `DLTensor*`
fields.
--
This is an automate
Lunderberg merged PR #16563:
URL: https://github.com/apache/tvm/pull/16563
--
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.apach
This is an automated email from the ASF dual-hosted git repository.
lunderberg 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 b21855758e [Relax] Implement operators to read runt
tqchen merged PR #16610:
URL: https://github.com/apache/tvm/pull/16610
--
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.
tqchen 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 e5bfb028b9 [Doc] Fixed Docstring usage example in `tvm.
tqchen commented on PR #16588:
URL: https://github.com/apache/tvm/pull/16588#issuecomment-1955244162
I still think in this case the overall constraining the behavior of
ConstIntBound is more predictable and readable.
My main worry is we open a flood gate of introducing many recursive
sdalvi-quic commented on issue #16577:
URL: https://github.com/apache/tvm/issues/16577#issuecomment-1955530546
Thank you eric for pointing to the location which might be causing issue.
Yes, the additional predicate is not getting appended to the constraints.
--
This is an automated messa
vinx13 commented on code in PR #16598:
URL: https://github.com/apache/tvm/pull/16598#discussion_r1496719908
##
src/relax/transform/fuse_ops.cc:
##
@@ -1238,10 +1238,14 @@ class CompositeFunctionAnnotator : public ExprMutator {
Expr VisitExpr_(const FunctionNode* func_node)
Lunderberg commented on code in PR #16598:
URL: https://github.com/apache/tvm/pull/16598#discussion_r1496737460
##
src/relax/transform/fuse_ops.cc:
##
@@ -1238,10 +1238,14 @@ class CompositeFunctionAnnotator : public ExprMutator {
Expr VisitExpr_(const FunctionNode* func_no
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 e5bfb028b9 [Doc] Fixed Docstring usage example in `tvm.ir.make_node`
(#16610)
add d91fe450c8 [Transform][Bugfix] Handl
vinx13 merged PR #16598:
URL: https://github.com/apache/tvm/pull/16598
--
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
Lunderberg commented on code in PR #16598:
URL: https://github.com/apache/tvm/pull/16598#discussion_r1496740578
##
src/relax/transform/fuse_ops.cc:
##
@@ -1238,10 +1238,14 @@ class CompositeFunctionAnnotator : public ExprMutator {
Expr VisitExpr_(const FunctionNode* func_no
vinx13 commented on code in PR #16598:
URL: https://github.com/apache/tvm/pull/16598#discussion_r1496741848
##
src/relax/transform/fuse_ops.cc:
##
@@ -1238,10 +1238,14 @@ class CompositeFunctionAnnotator : public ExprMutator {
Expr VisitExpr_(const FunctionNode* func_node)
vinx13 commented on PR #16595:
URL: https://github.com/apache/tvm/pull/16595#issuecomment-1955701786
`As a result, expressions such as R.zeros([16], "int32") would be extracted
out into the parameter transformation, even though they do not depend on any
parameters.
`
Does this affect
Lunderberg commented on PR #16562:
URL: https://github.com/apache/tvm/pull/16562#issuecomment-1955842490
> I'm surprised that the previous logic failed.
Yeah, I was pretty surprised at it, too. I think the core issue is the
mismatch between the internal representation of a `relax::If
Lunderberg commented on PR #16595:
URL: https://github.com/apache/tvm/pull/16595#issuecomment-1955860414
> Does this affect the result? If a weight transformation depends on some
values like `R.zeros`, such transformation will no longer be lifted if
`R.zeros` is not lifted, maybe better we
Lunderberg commented on code in PR #16598:
URL: https://github.com/apache/tvm/pull/16598#discussion_r1496878192
##
src/relax/transform/fuse_ops.cc:
##
@@ -1238,10 +1238,14 @@ class CompositeFunctionAnnotator : public ExprMutator {
Expr VisitExpr_(const FunctionNode* func_no
LeiWang1999 commented on code in PR #16560:
URL: https://github.com/apache/tvm/pull/16560#discussion_r1496882770
##
src/tir/schedule/primitive/blockize_tensorize.cc:
##
@@ -738,6 +739,28 @@ StmtSRef Blockize(ScheduleState self, const
Array& blocks, bool preser
return result;
slyubomirsky commented on code in PR #16602:
URL: https://github.com/apache/tvm/pull/16602#discussion_r1496856295
##
python/tvm/relax/transform/lazy_transform_params.py:
##
@@ -157,24 +159,60 @@ def transform(self, func: relax.Function) ->
relax.Function:
self.memory_f
zxybazh opened a new issue, #16614:
URL: https://github.com/apache/tvm/issues/16614
### Expected behavior
MetaSchedule Tuning Works for the given Conv2d workload
### Actual behavior
Triggers an error `ValueError: The block no longer exists in the IRModule`
during applica
quic-sanirudh commented on PR #16557:
URL: https://github.com/apache/tvm/pull/16557#issuecomment-1955892762
@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.
T
Lunderberg commented on PR #16588:
URL: https://github.com/apache/tvm/pull/16588#issuecomment-1955895172
> My main worry is we open a flood gate of introducing many recursive
rewrite patterns to `ConstIntBound` itself.
Ah, I think I see where I may have miscommunicated. There isn't a
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 46202e [KVCache] Support passing in attn_score_scaling_factor into
KV cache (#16606)
add 2066ce9612 [Unity
41 matches
Mail list logo
