| Issue |
169370
|
| Summary |
[X86][Clang] VectorExprEvaluator::VisitCallExpr / InterpretBuiltin - allow PD2PS fp truncation intrinsics to be used in constexpr
|
| Labels |
good first issue,
backend:X86,
clang:frontend,
clang:headers,
constexpr,
clang:bytecode
|
| Assignees |
|
| Reporter |
RKSimon
|
Similar to #160312 except initially we should only allow EXACT and FINITE conversions - any special case (DENORMAL/INF/NAN/etc.) or loss of precision shouldn't be allowed in constexpr for now.
```
__m128 _mm_cvtsd_ss (__m128 a, __m128d b)
__m128 _mm_mask_cvtsd_ss (__m128 src, __mmask8 k, __m128 a, __m128d b)
__m128 _mm_maskz_cvtsd_ss (__mmask8 k, __m128 a, __m128d b)
__m128 _mm_cvtpd_ps (__m128d a)
__m128 _mm_mask_cvtpd_ps (__m128 src, __mmask8 k, __m128d a)
__m128 _mm_maskz_cvtpd_ps (__mmask8 k, __m128d a)
__m128 _mm256_cvtpd_ps (__m256d a)
__m128 _mm256_mask_cvtpd_ps (__m128 src, __mmask8 k, __m256d a)
__m128 _mm256_maskz_cvtpd_ps (__mmask8 k, __m256d a)
__m256 _mm512_cvtpd_ps (__m512d a)
__m256 _mm512_mask_cvtpd_ps (__m256 src, __mmask8 k, __m512d a)
__m256 _mm512_maskz_cvtpd_ps (__mmask8 k, __m512d a)
__m512 _mm512_cvtpd_pslo (__m512d v2)
__m512 _mm512_mask_cvtpd_pslo (__m512 src, __mmask8 k, __m512d v2)
```
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
