[Bug target/18154] Inefficient max/min code for PowerPC

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=18154

--- Comment #11 from Segher Boessenkool <segher at gcc dot gnu.org> ---
The signed version can be done in four insns:

1:      subfc   r5,r3,r4
        subfe   r6,r6,r6
        and     r7,r6,r5
        addc    r8,r7,r3

(superopt finds 16 versions, all similar).

The unsigned version can be done in six:

33:     subfc   r5,r3,r4
        srwi    r6,r4,31
        srwi    r7,r3,31
        subfe   r8,r6,r7
        and     r9,r8,r5
        addc    r10,r9,r3

(superopt finds 240 versions, many with one or two xoris ,,0x8000
which doesn't work for 64 bit, and many with srawi as well, which
can be more expensive than srwi; all remaining are similar).

For 32-bit min/max on a 64-bit cpu, we can use only "cheap", non-carry
instructions:

  extsw r3,r3
  extsw r4,r4
  subf r5,r4,r3
  srdi r6,r5,32
  and r7,r6,r5
  add r8,r7,r4

(and unsigned exts for unsigned).  Those extends often disappear into
surrounding insns, or because the ABI requires the regs to be extended
already, etc.