This is the mail archive of the gcc-patches@gcc.gnu.org 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]

Re: loop optimizations, mark 11a



  In message <19980602200518.A13795@dot.cygnus.com>you write:
  > 
  > --rwEMma7ioTxnRzrJ
  > Content-Type: text/plain; charset=us-ascii
  > 
  > Ignore mark 11, as it didn't fix the ia32 problem.  Which was not
  > in fact a bug in the old code, but rather an extended brain seizure
  > on my part.
  > 
  > Try this instead.
spec92 runs on the x86 now.  Performance is spotty (generally
better overall for integer, generally worse for FP).  I haven't
tried to bootstrap an x86 or evaluate the performance issues in
any detail (I will try a bootstrap this weekend if I can find a
few minutes).

spec92 wave5 is mis-compiled on the PA with your patches.  Haven't 
investigated why yet either.  Nor have I tried to bootstrap the
PA port yet.

jeff



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