math, cmd/internal/obj/ppc64: improve floor, ceil, trunc with asm

This adds the instructions frim, frip, and friz to the ppc64x
assembler for use in implementing the math.Floor, math.Ceil, and
math.Trunc functions to improve performance.

Fixes #17185

BenchmarkCeil-128                    21.4          6.99          -67.34%
BenchmarkFloor-128                   13.9          6.37          -54.17%
BenchmarkTrunc-128                   12.7          6.33          -50.16%

Change-Id: I96131bd4e8c9c8dbafb25bfeb544cf9d2dbb4282
Reviewed-on: https://go-review.googlesource.com/29654
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Michael Munday <munday@ca.ibm.com>

src/cmd/internal/obj/ppc64/a.out.go

@@ -496,6 +496,12 @@ const (
 	/* optional on 32-bit */
 	AFRES
 	AFRESCC
+	AFRIM
+	AFRIMCC
+	AFRIP
+	AFRIPCC
+	AFRIZ
+	AFRIZCC
 	AFRSQRTE
 	AFRSQRTECC
 	AFSEL

src/cmd/internal/obj/ppc64/anames.go

@@ -229,6 +229,12 @@ var Anames = []string{
 	"RFCI",
 	"FRES",
 	"FRESCC",
+	"FRIM",
+	"FRIMCC",
+	"FRIP",
+	"FRIPCC",
+	"FRIZ",
+	"FRIZCC",
 	"FRSQRTE",
 	"FRSQRTECC",
 	"FSEL",

src/cmd/internal/obj/ppc64/asm9.go

@@ -1371,6 +1371,12 @@ func buildop(ctxt *obj.Link) {
 			opset(AFCFIDUCC, r0)
 			opset(AFRES, r0)
 			opset(AFRESCC, r0)
+			opset(AFRIM, r0)
+			opset(AFRIMCC, r0)
+			opset(AFRIP, r0)
+			opset(AFRIPCC, r0)
+			opset(AFRIZ, r0)
+			opset(AFRIZCC, r0)
 			opset(AFRSQRTE, r0)
 			opset(AFRSQRTECC, r0)
 			opset(AFSQRT, r0)
@@ -3173,6 +3179,18 @@ func oprrr(ctxt *obj.Link, a obj.As) uint32 {
 		return OPVCC(59, 24, 0, 0)
 	case AFRESCC:
 		return OPVCC(59, 24, 0, 1)
+	case AFRIM:
+		return OPVCC(63, 488, 0, 0)
+	case AFRIMCC:
+		return OPVCC(63, 488, 0, 1)
+	case AFRIP:
+		return OPVCC(63, 456, 0, 0)
+	case AFRIPCC:
+		return OPVCC(63, 456, 0, 1)
+	case AFRIZ:
+		return OPVCC(63, 424, 0, 0)
+	case AFRIZCC:
+		return OPVCC(63, 424, 0, 1)
 	case AFRSP:
 		return OPVCC(63, 12, 0, 0)
 	case AFRSPCC:

src/math/floor_ppc64x.s

+// Copyright 2016 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build ppc64 ppc64le
+
+#include "textflag.h"
+
+TEXT ·Floor(SB),NOSPLIT,$0
+	FMOVD   x+0(FP), F0
+	FRIM	F0, F0
+	FMOVD   F0, ret+8(FP)
+	RET
+
+TEXT ·Ceil(SB),NOSPLIT,$0
+	FMOVD   x+0(FP), F0
+	FRIP    F0, F0
+	FMOVD	F0, ret+8(FP)
+	RET
+
+TEXT ·Trunc(SB),NOSPLIT,$0
+	FMOVD   x+0(FP), F0
+	FRIZ    F0, F0
+	FMOVD   F0, ret+8(FP)
+	RET

src/math/stubs_ppc64x.s

@@ -36,15 +36,6 @@ TEXT ·Expm1(SB),NOSPLIT,$0
 TEXT ·Exp(SB),NOSPLIT,$0
 	BR ·exp(SB)
 
-TEXT ·Floor(SB),NOSPLIT,$0
-	BR ·floor(SB)
-
-TEXT ·Ceil(SB),NOSPLIT,$0
-	BR ·ceil(SB)
-
-TEXT ·Trunc(SB),NOSPLIT,$0
-	BR ·trunc(SB)
-
 TEXT ·Frexp(SB),NOSPLIT,$0
 	BR ·frexp(SB)