ysh329 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2044378573
> > #16770 The sage continues...
>
> Hi Steven, I'm little confused. Does these three cases been fixed?
>
> I tried main with commit: 240326,
ysh329 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2044136386
> > #16770 The sage continues...
>
> Hi Steven, I'm little confused. Does these three cases been fixed?
>
> I tried main with commit: 240326,
ysh329 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2042366352
> > #16770 The sage continues...
>
> Hi Steven, I'm little confused. Does these three cases been fixed?
>
> I tried main with commit: 240326,
ysh329 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2019299851
> #16770 The sage continues...
Hi Steven, I'm little confused. Does these three cases been fixed?
I tried main with commit: 240326, b2204ae6 and update status as below. It
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2015639445
https://github.com/apache/tvm/pull/16770 The sage continues...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-2014165365
These bugs should all be fixed now that #16569 was merged. I recommend you
close unless there's a change in here that wasn't covered there.
--
This is an automated message from the
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1989743075
I'm not sure all of them actually contain the string "gpu"
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
Lunderberg commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1981273199
> You need to update `IsScheduledOnGPU()` in `default_gpu_schedule.cc` to
also include Metal in its definition of GPU.
Should we instead check whether the target tags include
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1979832554
Fix for the `add_on_metal` bug: You need to update `IsScheduledOnGPU()` in
`default_gpu_schedule.cc` to also include Metal in its definition of GPU.
```c++
bool
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1979755734
Bisect for `test_transform_default_gpu_schedule.py::test_add_on_metal`:
First failing commit is `1c35c392648e4336fc5e00ab91abb37af997cd59`. It looks to
have made changes to test
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1974035630
I did a git bisect for `test_tir_transform_hoist_if.py`:
`eb15d04c3bff76062e26d5647fb8af0323de1bed` is the first bad commit for
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1973891325
I've figured out, thanks to @Lunderberg's comments in my PR, that the
failure in
`tests/python/tir-transform/test_tir_transform_force_narrow_index_to_i32.py::test_thread_axis2`
is
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1967669197
> Run these test cases in local with this patch, below 3 test cases still
will be failed because of TIR structure isn't equal, @Hzfengsy @tqchen can you
help to see them? Thanks.
>
slyubomirsky commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1967667539
I had this issue come up in https://github.com/apache/tvm/pull/16569. I
think these tests probably were not being executed before. I am also pretty
sure it's safe to just remove those
Hzfengsy commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1966380114
cc @tqchen
--
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
Johnson9009 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1963242851
The reason of why the crashed test cases won't be caught by CI is the pytest
exist code is 5, it means no tests is collected, in file
"tests/scripts/setup-pytest-env.sh" we haven't
Johnson9009 commented on PR #16634:
URL: https://github.com/apache/tvm/pull/16634#issuecomment-1961546682
Run these test cases in local with this patch, below 3 test cases still will
be failed because of TIR structure isn't equal, @Hzfengsy @tqchen can you help
to see them? Thanks.
17 matches
Mail list logo