diff options
Diffstat (limited to 'gcc')
| -rw-r--r-- | gcc/ChangeLog | 6 | ||||
| -rw-r--r-- | gcc/config/s390/vector.md | 83 |
2 files changed, 85 insertions, 4 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 5e7f82d..9df1ff5 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,9 @@ +2017-09-26 Andreas Krebbel <krebbel@linux.vnet.ibm.com> + + * config/s390/vector.md ("vec_widen_umult_lo_<mode>") + ("vec_widen_umult_hi_<mode>", "vec_widen_smult_lo_<mode>") + ("vec_widen_smult_hi_<mode>"): New expander definitions. + 2017-09-26 Richard Earnshaw <rearnsha@arm.com> PR target/82175 diff --git a/gcc/config/s390/vector.md b/gcc/config/s390/vector.md index 3cf7989..29131cd 100644 --- a/gcc/config/s390/vector.md +++ b/gcc/config/s390/vector.md @@ -1065,10 +1065,85 @@ "vmlo<bhfgq>\t%v0,%v1,%v2" [(set_attr "op_type" "VRR")]) -; vec_widen_umult_hi -; vec_widen_umult_lo -; vec_widen_smult_hi -; vec_widen_smult_lo + +; Widening hi/lo multiplications + +; The S/390 instructions vml and vmh return the low or high parts of +; the double sized result elements in the corresponding elements of +; the target register. That's NOT what the vec_widen_umult_lo/hi +; patterns are expected to do. + +; We emulate the widening lo/hi multiplies with the even/odd versions +; followed by a vector merge + + +(define_expand "vec_widen_umult_lo_<mode>" + [(set (match_dup 3) + (unspec:<vec_double> [(match_operand:VI_QHS 1 "register_operand" "%v") + (match_operand:VI_QHS 2 "register_operand" "v")] + UNSPEC_VEC_UMULT_EVEN)) + (set (match_dup 4) + (unspec:<vec_double> [(match_dup 1) (match_dup 2)] + UNSPEC_VEC_UMULT_ODD)) + (set (match_operand:<vec_double> 0 "register_operand" "=v") + (unspec:<vec_double> [(match_dup 3) (match_dup 4)] + UNSPEC_VEC_MERGEL))] + "TARGET_VX" + { + operands[3] = gen_reg_rtx (<vec_double>mode); + operands[4] = gen_reg_rtx (<vec_double>mode); + }) + +(define_expand "vec_widen_umult_hi_<mode>" + [(set (match_dup 3) + (unspec:<vec_double> [(match_operand:VI_QHS 1 "register_operand" "%v") + (match_operand:VI_QHS 2 "register_operand" "v")] + UNSPEC_VEC_UMULT_EVEN)) + (set (match_dup 4) + (unspec:<vec_double> [(match_dup 1) (match_dup 2)] + UNSPEC_VEC_UMULT_ODD)) + (set (match_operand:<vec_double> 0 "register_operand" "=v") + (unspec:<vec_double> [(match_dup 3) (match_dup 4)] + UNSPEC_VEC_MERGEH))] + "TARGET_VX" + { + operands[3] = gen_reg_rtx (<vec_double>mode); + operands[4] = gen_reg_rtx (<vec_double>mode); + }) + +(define_expand "vec_widen_smult_lo_<mode>" + [(set (match_dup 3) + (unspec:<vec_double> [(match_operand:VI_QHS 1 "register_operand" "%v") + (match_operand:VI_QHS 2 "register_operand" "v")] + UNSPEC_VEC_SMULT_EVEN)) + (set (match_dup 4) + (unspec:<vec_double> [(match_dup 1) (match_dup 2)] + UNSPEC_VEC_SMULT_ODD)) + (set (match_operand:<vec_double> 0 "register_operand" "=v") + (unspec:<vec_double> [(match_dup 3) (match_dup 4)] + UNSPEC_VEC_MERGEL))] + "TARGET_VX" + { + operands[3] = gen_reg_rtx (<vec_double>mode); + operands[4] = gen_reg_rtx (<vec_double>mode); + }) + +(define_expand "vec_widen_smult_hi_<mode>" + [(set (match_dup 3) + (unspec:<vec_double> [(match_operand:VI_QHS 1 "register_operand" "%v") + (match_operand:VI_QHS 2 "register_operand" "v")] + UNSPEC_VEC_SMULT_EVEN)) + (set (match_dup 4) + (unspec:<vec_double> [(match_dup 1) (match_dup 2)] + UNSPEC_VEC_SMULT_ODD)) + (set (match_operand:<vec_double> 0 "register_operand" "=v") + (unspec:<vec_double> [(match_dup 3) (match_dup 4)] + UNSPEC_VEC_MERGEH))] + "TARGET_VX" + { + operands[3] = gen_reg_rtx (<vec_double>mode); + operands[4] = gen_reg_rtx (<vec_double>mode); + }) ; vec_widen_ushiftl_hi ; vec_widen_ushiftl_lo |