Open mxswd opened 10 years ago
is there any difference between CFloat vs Float and CDouble vs Double for the purposes of FFI? AFAIK, the C* versions of float and double are just newtype wrappers and then the ghc ffi work treats them as Float and Double? (i could be totally wrong here, so please correct me if i'm wrong.)
Good catch #define CBLAS_INDEX size_t
seems to the the right defn, so it should be CSize right?
you're right, i'll check that shortly.
Yeah, leave the Floats and Doubles.
The types of the functions (i.e.
cblas_sdot_unsafe
) are Float and Double, when they are really CFloat and CDouble.Also
cblas_isamax_unsafe
/cblas_idamax_unsafe
is returning aCInt
, when really it returns aCUInt
.Also,
cblas_idamax_unsafe
's 2nd argument takes a Float, when it should take a Double?