Skip to content

Conversation

@manuelcandales
Copy link
Contributor

@manuelcandales manuelcandales commented Jul 8, 2025

@manuelcandales manuelcandales requested a review from swolchok as a code owner July 8, 2025 18:33
@pytorch-bot
Copy link

pytorch-bot bot commented Jul 8, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/12277

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (1 Unrelated Failure)

As of commit c18c05e with merge base ba19c75 (image):

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Jul 8, 2025
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D77759060

@manuelcandales manuelcandales added the release notes: none Do not include this in the release notes label Jul 8, 2025
Copy link
Contributor

@swolchok swolchok left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

accepted with minor revision, see comments

double rsqrt(double x) {
template <typename T>
T rsqrt(T x) {
return 1.0 / std::sqrt(x);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

think this should be T(1.0) otherwise we're forcing computation in double, right?

double rsqrt(double x) {
template <typename T>
T rsqrt(T x) {
return 1.0 / std::sqrt(x);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ditto T(1.0)

Comment on lines +93 to +98
#define DEFINE_UNARY_UFUNC_REALHBBF16_TO_FLOATHBF16(op_name, fn) \
Tensor& op_name(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) { \
return internal::unary_ufunc_realhbbf16_to_floathbf16( \
fn, fn, ctx, in, out); \
}

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why not define this macro to use the full namespace (prefixed with :: so it works from anywhere) so that backends/cadence can use it too?

@facebook-github-bot facebook-github-bot merged commit d146c80 into gh/manuelcandales/130/base Jul 8, 2025
98 of 102 checks passed
@facebook-github-bot facebook-github-bot deleted the gh/manuelcandales/130/head branch July 8, 2025 22:10
manuelcandales added a commit that referenced this pull request Jul 9, 2025
…2289)

This PR was created by the merge bot to help merge the original PR into
the main branch.
ghstack PR number: #12277 by
@manuelcandales
^ Please use this as the source of truth for the PR details, comments,
and reviews
ghstack PR base:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/base
ghstack PR head:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/main
Merge bot PR head:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/orig
@diff-train-skip-merge

Co-authored-by: Manuel Candales <manuelcandales@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported release notes: none Do not include this in the release notes

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants