Re:Lisper's first look at Haskell
by jedbrown on Friday July 23, 2010 @06:54AM
(#1540864)
Huh? Double precision SIMD? SSE2 instructions aren't slower than SSE.
[ Parent
| on HN ]
No, single precision, as in "the fastest way to multiply a double is to use floats instead". I've seen more than enough C programs unnecessarily using double that I wish floating point constants would just be float by default.
(Although it seems like clang does something special here, as I don't see many double-to-float conversions in its output.)
Then you aren't multiplying doubles, it would be a shame to get a completely wrong answer on an ill-conditioned problem because the implementation disregarded the spec, cf. "the fastest way to process UTF-8 is to truncate it to ASCII", "the fastest RSA encryption is ROT-13". Also note that floats are only faster with packed instructions (often not possible without changing APIs) or when memory bandwidth is an issue.