[Bug tree-optimization/88713] Vectorized code slow vs. flang

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88713

--- Comment #40 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to rguent...@suse.de from comment #39)
> > > 
> > > Yes.  The lack of an expander for the rqsrt operation is probably
> > > more severe though (causing sqrt + approx recip to appear)
> > > 
> > 
> > Can we use UNSPEC_RSQRT14 here if UNSPEC_RSQRT28 isn't available?
> 
> I think we can but we lack an expander for this.  IIRC for the following
> existing expander the RTL is ignored and thus we could simply
> replace the TARGET_AVX512ER check with TARGET_AVX512F?
> 
> (define_expand "rsqrtv16sf2"
>   [(set (match_operand:V16SF 0 "register_operand")
>         (unspec:V16SF
>           [(match_operand:V16SF 1 "vector_operand")]
>           UNSPEC_RSQRT28))]
>   "TARGET_SSE_MATH && TARGET_AVX512ER"
> {
>   ix86_emit_swsqrtsf (operands[0], operands[1], V16SFmode, true);
>   DONE;
> })

Like this?

diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index 3af4adc63dd..c9b4750ccc4 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -1969,21 +1969,11 @@
    (set_attr "mode" "<ssescalarmode>")])

 (define_expand "rsqrt<mode>2"
-  [(set (match_operand:VF1_128_256 0 "register_operand")
-  (unspec:VF1_128_256
-    [(match_operand:VF1_128_256 1 "vector_operand")] UNSPEC_RSQRT))]
+  [(set (match_operand:VF_AVX512VL 0 "register_operand")
+  (unspec:VF_AVX512VL
+    [(match_operand:VF_AVX512VL 1 "vector_operand")]
+    UNSPEC_RSQRT))]
   "TARGET_SSE_MATH"
-{
-  ix86_emit_swsqrtsf (operands[0], operands[1], <MODE>mode, true);
-  DONE;
-})
-
-(define_expand "rsqrtv16sf2"
-  [(set (match_operand:V16SF 0 "register_operand")
-  (unspec:V16SF
-    [(match_operand:V16SF 1 "vector_operand")]
-    UNSPEC_RSQRT28))]
-  "TARGET_SSE_MATH && TARGET_AVX512ER"
 {
   ix86_emit_swsqrtsf (operands[0], operands[1], V16SFmode, true);
   DONE;