Skip to content

Fix neon64/swizzle on MSVC. - #1410

Open
degasus wants to merge 1 commit into
xtensor-stack:masterfrom
degasus:neon_convert
Open

Fix neon64/swizzle on MSVC.#1410
degasus wants to merge 1 commit into
xtensor-stack:masterfrom
degasus:neon_convert

Conversation

@degasus

@degasus degasus commented Sep 4, 2026

Copy link
Copy Markdown
Contributor

vreinterpretq is a no-op here, not even converting the type. As we provide the batch into vreinterpretq, we keep its strong typing and on return, CL tells me that it cannot convert to the output type.

Please don't treat this is high quality code - more as a bug report - even if it fix compiling my kernels.
After looking at this code for roughly 30 minutes, I assume a real fix might be to encapsulate all vreinterpretq calls into something strongly typed, even if this sounds annoying.

I actually wonder, is there a test which checks if swizzle compiles?

vreinterpretq is a no-op here, not even converting the type.
As we provide the batch into vreinterpretq, we keep its strong typing
and on return, CL tells me that it cannot convert to the output type.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant