FFTW: NEON SIMD is float-only (#26321)

Fix on Apple Mac M1 (aarch64):
```
configure: error: NEON requires single precision
```
This commit is contained in:
Axel Huebl 2021-09-28 19:40:45 -07:00 committed by GitHub
parent c578882bb2
commit c5410bd333
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -122,9 +122,9 @@ def configure(self, spec, prefix):
# Specific SIMD support.
# all precisions
simd_features = ['sse2', 'avx', 'avx2', 'avx512', 'avx-128-fma',
'kcvi', 'vsx', 'neon']
'kcvi', 'vsx']
# float only
float_simd_features = ['altivec', 'sse']
float_simd_features = ['altivec', 'sse', 'neon']
# Workaround PGI compiler bug when avx2 is enabled
if spec.satisfies('%pgi') and 'avx2' in simd_features: