This is the mail archive of the java-discuss@sources.redhat.com mailing list for the Java project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Don't optimize by hand (Was Re: Lack of fabsf on Solaris (patch included))


Tom Tromey wrote:

> And we probably don't really implement strict fp
> semantics anyway (I haven't checked, but does fp multiply or whatever
> really call a function?).

Nope, and I'd argue that implementing strictfp is not important at all.
The 'strictfp' keyword is completely ignored by all the common, current
VMs on the x86.

I suspect that several of the natMath.cc functions could be replaced
with calls to __builtin functions, which presumably ensures that optimum
code is generated by the backend whatever the target. We also want the
compiler to inline them, but the implementations still need to exist in
natMath.cc in order to support reflection and the interpreter. It seems
cleaner to use gcc's built in functionality rather than relying on
another math library.

regards

  [ bryce ]



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]