AVX512FP16: Add expander for sqrthf2.

gcc/ChangeLog:

	* config/i386/i386-features.c (i386-features.c): Handle
	E_HFmode.
	* config/i386/i386.md (sqrthf2): New expander.
	(*sqrthf2): New define_insn.
	* config/i386/sse.md
	(*<sse>_vmsqrt<mode>2<mask_scalar_name><round_scalar_name>):
	Extend to VFH_128.

gcc/testsuite/ChangeLog:

	* gcc.target/i386/avx512fp16-builtin-sqrt-1.c: New test.
	* gcc.target/i386/avx512fp16vl-builtin-sqrt-1.c: New test.

diff --git a/gcc/config/i386/i386-features.c b/gcc/config/i386/i386-features.c
index 14f816f..43bb676 100644 (file)
--- a/gcc/config/i386/i386-features.c
+++ b/gcc/config/i386/i386-features.c
@@ -2258,15 +2258,22 @@ remove_partial_avx_dependency (void)
 
          rtx zero;
          machine_mode dest_vecmode;
-         if (dest_mode == E_SFmode)
+         switch (dest_mode)
            {
+           case E_HFmode:
+             dest_vecmode = V8HFmode;
+             zero = gen_rtx_SUBREG (V8HFmode, v4sf_const0, 0);
+             break;
+           case E_SFmode:
              dest_vecmode = V4SFmode;
              zero = v4sf_const0;
-           }
-         else
-           {
+             break;
+           case E_DFmode:
              dest_vecmode = V2DFmode;
              zero = gen_rtx_SUBREG (V2DFmode, v4sf_const0, 0);
+             break;
+           default:
+             gcc_unreachable ();
            }
 
          /* Change source to vector mode.  */

diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 188f431..ae1a81c 100644 (file)
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -17041,6 +17041,19 @@
   DONE;
 })
 
+(define_insn "sqrthf2"
+  [(set (match_operand:HF 0 "register_operand" "=v,v")
+       (sqrt:HF
+         (match_operand:HF 1 "nonimmediate_operand" "v,m")))]
+  "TARGET_AVX512FP16"
+  "@
+   vsqrtsh\t{%d1, %0|%0, %d1}
+   vsqrtsh\t{%1, %d0|%d0, %1}"
+  [(set_attr "type" "sse")
+   (set_attr "prefix" "evex")
+   (set_attr "avx_partial_xmm_update" "false,true")
+   (set_attr "mode" "HF")])
+
 (define_insn "*sqrt<mode>2_sse"
   [(set (match_operand:MODEF 0 "register_operand" "=v,v,v")
        (sqrt:MODEF

diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index b08a9d3..e8aef0d 100644 (file)
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -2522,12 +2522,12 @@
    (set_attr "mode" "<ssescalarmode>")])
 
 (define_insn "*<sse>_vmsqrt<mode>2<mask_scalar_name><round_scalar_name>"
-  [(set (match_operand:VF_128 0 "register_operand" "=x,v")
-       (vec_merge:VF_128
-         (vec_duplicate:VF_128
+  [(set (match_operand:VFH_128 0 "register_operand" "=x,v")
+       (vec_merge:VFH_128
+         (vec_duplicate:VFH_128
            (sqrt:<ssescalarmode>
              (match_operand:<ssescalarmode> 1 "nonimmediate_operand" "xm,<round_scalar_constraint>")))
-         (match_operand:VF_128 2 "register_operand" "0,v")
+         (match_operand:VFH_128 2 "register_operand" "0,v")
          (const_int 1)))]
   "TARGET_SSE"
   "@

diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16-builtin-sqrt-1.c b/gcc/testsuite/gcc.target/i386/avx512fp16-builtin-sqrt-1.c
new file mode 100644 (file)
index 0000000..b4efc84
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx512fp16-builtin-sqrt-1.c
@@ -0,0 +1,18 @@
+/* { dg-do compile } */
+/* { dg-options "-Ofast -mavx512fp16 -mprefer-vector-width=512" } */
+
+_Float16
+f1 (_Float16 x)
+{
+  return __builtin_sqrtf16 (x);
+}
+
+void
+f2 (_Float16* __restrict psrc, _Float16* __restrict pdst)
+{
+  for (int i = 0; i != 32; i++)
+    pdst[i] = __builtin_sqrtf16 (psrc[i]);
+}
+
+/* { dg-final { scan-assembler-times "vsqrtsh\[^\n\r\]*xmm\[0-9\]" 1 } } */
+/* { dg-final { scan-assembler-times "vsqrtph\[^\n\r\]*zmm\[0-9\]" 1 } } */

diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16vl-builtin-sqrt-1.c b/gcc/testsuite/gcc.target/i386/avx512fp16vl-builtin-sqrt-1.c
new file mode 100644 (file)
index 0000000..08deb3e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/avx512fp16vl-builtin-sqrt-1.c
@@ -0,0 +1,19 @@
+/* { dg-do compile } */
+/* { dg-options "-Ofast -mavx512fp16 -mavx512vl" } */
+
+void
+f1 (_Float16* __restrict psrc, _Float16* __restrict pdst)
+{
+  for (int i = 0; i != 8; i++)
+    pdst[i] = __builtin_sqrtf16 (psrc[i]);
+}
+
+void
+f2 (_Float16* __restrict psrc, _Float16* __restrict pdst)
+{
+  for (int i = 0; i != 16; i++)
+    pdst[i] = __builtin_sqrtf16 (psrc[i]);
+}
+
+/* { dg-final { scan-assembler-times "vsqrtph\[^\n\r\]*xmm\[0-9\]" 1 } } */
+/* { dg-final { scan-assembler-times "vsqrtph\[^\n\r\]*ymm\[0-9\]" 1 } } */