Add restriction to "vec::as" #447
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This change was prompted by a discussion in KhronosGroup/SYCL-CTS#756, relating to a test of `vec::as`. That test tried to use `vec::as` to reinterpret a `sycl::vec<float, 3>` as `sycl::vec<int, 4>`, which is legal according to the spec because they both have the same storage size. (This is because a `vec` of length 3 has the same storage size as a `vec` of length 4.) However, the semantics are unclear because the content of the 4th element is undefined for a 3 element `vec`. It seems reasonable to simply state that such a use of `vec::as` is not allowed. I think our intent was to allow `vec::as` to reinterpret only the bits of defined elements in a `vec`, which is what this change does.
0x12CC
added a commit
to 0x12CC/SYCL-CTS
that referenced
this pull request
Jul 27, 2023
Disable `check_vector_as` comparisons for cases where the size of the elements differs after conversion. This matches the restriction described in KhronosGroup/SYCL-Docs#447. Signed-off-by: Michael Aziz <michael.aziz@intel.com>
keryell
approved these changes
Aug 10, 2023
TApplencourt
approved these changes
Aug 17, 2023
|
LGTM |
AerialMantis
approved these changes
Aug 17, 2023
|
not related to 455 |
keryell
added a commit
that referenced
this pull request
Sep 10, 2024
This change was prompted by a discussion in KhronosGroup/SYCL-CTS#756, relating to a test of `vec::as`. That test tried to use `vec::as` to reinterpret a `sycl::vec<float, 3>` as `sycl::vec<int, 4>`, which is legal according to the spec because they both have the same storage size. (This is because a `vec` of length 3 has the same storage size as a `vec` of length 4.) However, the semantics are unclear because the content of the 4th element is undefined for a 3 element `vec`. It seems reasonable to simply state that such a use of `vec::as` is not allowed. I think our intent was to allow `vec::as` to reinterpret only the bits of defined elements in a `vec`, which is what this change does.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change was prompted by a discussion in KhronosGroup/SYCL-CTS#756, relating to a test of
vec::as. That test tried to usevec::asto reinterpret asycl::vec<float, 3>assycl::vec<int, 4>, which is legal according to the spec because they both have the same storage size. (This is because avecof length 3 has the same storage size as avecof length 4.) However, the semantics are unclear because the content of the 4th element is undefined for a 3 elementvec.It seems reasonable to simply state that such a use of
vec::asis not allowed. I think our intent was to allowvec::asto reinterpret only the bits of defined elements in avec, which is what this change does.