cvs commit: fptools/ghc/compiler/nativeGen MachInstrs.hs PprMach.hs RegAllocInfo.hs fptools/ghc/compiler/cmm MachOp.hs

[Simon Marlow]

simonmar    2004/02/19 04:19:57 PST

  Modified files:        (Branch: backend-hacking-branch)
    ghc/compiler/nativeGen MachInstrs.hs PprMach.hs 
                           RegAllocInfo.hs 
    ghc/compiler/cmm     MachOp.hs 
  Log:
  On second thoughts, go back to 80-bit spills for floating point
  registers on x86.  There's no point in making things worse than
  before, and we don't really care too much about the overhead (floating
  point is already terribly slow on x86, both with -fvia-C and -fasm,
  unless you use -fexcess precision).
  
  Revision  Changes    Path
  1.1.2.21  +48 -0     fptools/ghc/compiler/nativeGen/Attic/MachInstrs.hs
  1.1.2.27  +1 -0      fptools/ghc/compiler/nativeGen/Attic/PprMach.hs
  1.1.2.18  +3 -3      fptools/ghc/compiler/nativeGen/Attic/RegAllocInfo.hs
  1.1.2.22  +5 -0      fptools/ghc/compiler/cmm/Attic/MachOp.hs