[Bug target/98167] [x86] Failure to optimize operation on indentically shuffled operands into a shuffle of the result of the operation

[crazylht at gmail dot com via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98167

--- Comment #3 from Hongtao.liu <crazylht at gmail dot com> ---
;; _3 = __builtin_ia32_shufps (b_2(D), b_2(D), 0);

(insn 7 6 8 (set (reg:V4SF 88)
        (reg/v:V4SF 86 [ b ])) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))

(insn 8 7 9 (set (reg:V4SF 89)
        (reg/v:V4SF 86 [ b ])) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))

(insn 9 8 10 (set (reg:V4SF 87)
        (vec_select:V4SF (vec_concat:V8SF (reg:V4SF 88)
                (reg:V4SF 89))
            (parallel [
                    (const_int 0 [0]) repeated x2
                    (const_int 4 [0x4]) repeated x2
                ]))) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))
;; _5 = __builtin_ia32_shufps (a_4(D), a_4(D), 0);

(insn 11 10 12 (set (reg:V4SF 91)
        (reg/v:V4SF 85 [ a ])) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))

(insn 12 11 13 (set (reg:V4SF 92)
        (reg/v:V4SF 85 [ a ])) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))

(insn 13 12 14 (set (reg:V4SF 90)
        (vec_select:V4SF (vec_concat:V8SF (reg:V4SF 91)
                (reg:V4SF 92))
            (parallel [
                    (const_int 0 [0]) repeated x2
                    (const_int 4 [0x4]) repeated x2
                ]))) "./gcc/include/xmmintrin.h":746:19 -1
     (nil))


Simplify upper to

(vec_duplicate:V4SF
  (vec_select:SF (reg:V4SF 86)
                 (parallel [(const_int 0)])

then add a combine splitter transform (mult:(vec_dup op1) (vec_dup op2)) to
(vec_dup (mult:op1 op2)?