This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Patches awaiting review

It's the first of the month, so I thought I'd post my list of patches
still in the review process:
x87 constants for doubles with -ffast-math

  RTH says he has no objections, but would like me to convince another
  reviewer of the benefits.  Using x87 constants we generate smaller
  and faster code, and improve the accuracy of routines such as Brad
  Lucier's FFT codes.  This applies not only to pi, but also in the
  (near) future to the other x87 constants, log2(10.0), log(2.0) and
  log10(2.0).  There should be not observable differences for code that
  writes the result to memory immediately, and any other arithmetic
  operation immediately suffers x87 extended precision anyway.
  Any takers?
Optimize strcmp(ptr,ptr) and friends
Promoted variable SUBREG optimizations

  This is another patch where I disagree with Richard Henderson's and
  Richard Kenner's paranoia, but I'm not sure if I have the necessary
  stamina and hit points following my recent bout.
Enable distributive FP optimizations with -ffast-math
Don't emit cmpstrsi and movstr* insns with -Os (take 2)
Generalize builtin_mathfn_code

  I'm a bit disappointed that this has taken so long to review.  Many
  thanks once again to Kaveh for holding off on his approved patch,
  in preference for this one.  No good deed goes unpunished.
PR optimization/12260: Avoid (const (neg (const (plus X Y))))
PR optimization/12289: expand_call patch (take 2)
Fix evaluation order assumptions in expr.c
Constant folding NEGATE_EXPR clean-ups

Very many thanks in advance,

Roger Sayle,                         E-mail:
OpenEye Scientific Software,         WWW:
Suite 1107, 3600 Cerrillos Road,     Tel: (+1) 505-473-7385
Santa Fe, New Mexico, 87507.         Fax: (+1) 505-473-0833

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