Patchwork [rs6000] A few more vector builtins

login
register
mail settings
Submitter William J. Schmidt
Date Aug. 29, 2014, 6:28 p.m.
Message ID <1409336926.25590.84.camel@gnopaine>
Download mbox | patch
Permalink /patch/384351/
State New
Headers show

Comments

William J. Schmidt - Aug. 29, 2014, 6:28 p.m.
Hi,

This is the last in the current series of new vector built-ins.  This
group adds vec_ctf, vec_cts, and vec_ctu for vector double and vector
long long.  Additionally, it adds documentation for the built-ins added
in my last patch, since I forgot to add it then... /oops

Bootstrapped and tested on powerpc64le-unknown-linux-gnu with no
regressions.  Ok for trunk?

Thanks,
Bill


[gcc]

2014-08-29  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

	* config/rs6000/rs6000-builtin.def (XVCVSXDDP_SCALE):  New
	built-in definition.
	(XVCVUXDDP_SCALE): Likewise.
	(XVCVDPSXDS_SCALE): Likewise.
	(XVCVDPUXDS_SCALE): Likewise.
	* config/rs6000/rs6000-c.c (altivec_overloaded_builtins):  Add
	entries for VSX_BUILTIN_XVCVSXDDP_SCALE,
	VSX_BUILTIN_XVCVUXDDP_SCALE, VSX_BUILTIN_XVCVDPSXDS_SCALE, and
	VSX_BUILTIN_XVCVDPUXDS_SCALE.
	* config/rs6000/rs6000-protos.h (rs6000_scale_v2df): New
	prototype.
	* config/rs6000/rs6000.c (real.h): New include.
	(rs6000_scale_v2df): New function.
	* config/rs6000/vsx.md (UNSPEC_VSX_XVCVSXDDP): New unspec.
	(UNSPEC_VSX_XVCVUXDDP): Likewise.
	(UNSPEC_VSX_XVCVDPSXDS): Likewise.
	(UNSPEC_VSX_XVCVDPUXDS): Likewise.
	(vsx_xvcvsxddp_scale): New define_expand.
	(vsx_xvcvsxddp): New define_insn.
	(vsx_xvcvuxddp_scale): New define_expand.
	(vsx_xvcvuxddp): New define_insn.
	(vsx_xvcvdpsxds_scale): New define_expand.
	(vsx_xvcvdpsxds): New define_insn.
	(vsx_xvcvdpuxds_scale): New define_expand.
	(vsx_xvcvdpuxds): New define_insn.
	* doc/extend.texi (vec_ctf): Add new prototypes.
	(vec_cts): Likewise.
	(vec_ctu): Likewise.
	(vec_splat): Likewise.
	(vec_div): Likewise.
	(vec_mul): Likewise.

[gcc/testsuite]

2014-08-29  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>

	* gcc.target/powerpc/builtins-1.c: Add tests for vec_ctf,
	vec_cts, and vec_ctu.
	* gcc.target/powerpc/builtins-2.c: Likewise.
David Edelsohn - Aug. 30, 2014, 10:53 p.m.
On Fri, Aug 29, 2014 at 2:28 PM, Bill Schmidt
<wschmidt@linux.vnet.ibm.com> wrote:
> Hi,
>
> This is the last in the current series of new vector built-ins.  This
> group adds vec_ctf, vec_cts, and vec_ctu for vector double and vector
> long long.  Additionally, it adds documentation for the built-ins added
> in my last patch, since I forgot to add it then... /oops
>
> Bootstrapped and tested on powerpc64le-unknown-linux-gnu with no
> regressions.  Ok for trunk?
>
> Thanks,
> Bill
>
>
> [gcc]
>
> 2014-08-29  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
>
>         * config/rs6000/rs6000-builtin.def (XVCVSXDDP_SCALE):  New
>         built-in definition.
>         (XVCVUXDDP_SCALE): Likewise.
>         (XVCVDPSXDS_SCALE): Likewise.
>         (XVCVDPUXDS_SCALE): Likewise.
>         * config/rs6000/rs6000-c.c (altivec_overloaded_builtins):  Add
>         entries for VSX_BUILTIN_XVCVSXDDP_SCALE,
>         VSX_BUILTIN_XVCVUXDDP_SCALE, VSX_BUILTIN_XVCVDPSXDS_SCALE, and
>         VSX_BUILTIN_XVCVDPUXDS_SCALE.
>         * config/rs6000/rs6000-protos.h (rs6000_scale_v2df): New
>         prototype.
>         * config/rs6000/rs6000.c (real.h): New include.
>         (rs6000_scale_v2df): New function.
>         * config/rs6000/vsx.md (UNSPEC_VSX_XVCVSXDDP): New unspec.
>         (UNSPEC_VSX_XVCVUXDDP): Likewise.
>         (UNSPEC_VSX_XVCVDPSXDS): Likewise.
>         (UNSPEC_VSX_XVCVDPUXDS): Likewise.
>         (vsx_xvcvsxddp_scale): New define_expand.
>         (vsx_xvcvsxddp): New define_insn.
>         (vsx_xvcvuxddp_scale): New define_expand.
>         (vsx_xvcvuxddp): New define_insn.
>         (vsx_xvcvdpsxds_scale): New define_expand.
>         (vsx_xvcvdpsxds): New define_insn.
>         (vsx_xvcvdpuxds_scale): New define_expand.
>         (vsx_xvcvdpuxds): New define_insn.
>         * doc/extend.texi (vec_ctf): Add new prototypes.
>         (vec_cts): Likewise.
>         (vec_ctu): Likewise.
>         (vec_splat): Likewise.
>         (vec_div): Likewise.
>         (vec_mul): Likewise.
>
> [gcc/testsuite]
>
> 2014-08-29  Bill Schmidt  <wschmidt@linux.vnet.ibm.com>
>
>         * gcc.target/powerpc/builtins-1.c: Add tests for vec_ctf,
>         vec_cts, and vec_ctu.
>         * gcc.target/powerpc/builtins-2.c: Likewise.

Okay.

Thanks, David

Patch

Index: gcc/config/rs6000/rs6000-builtin.def
===================================================================
--- gcc/config/rs6000/rs6000-builtin.def	(revision 214735)
+++ gcc/config/rs6000/rs6000-builtin.def	(working copy)
@@ -1264,6 +1264,11 @@  BU_VSX_2 (DIV_V2DI,           "div_2di",        CO
 BU_VSX_2 (UDIV_V2DI,          "udiv_2di",       CONST,  vsx_udiv_v2di)
 BU_VSX_2 (MUL_V2DI,           "mul_2di",        CONST,  vsx_mul_v2di)
 
+BU_VSX_2 (XVCVSXDDP_SCALE,    "xvcvsxddp_scale", CONST, vsx_xvcvsxddp_scale)
+BU_VSX_2 (XVCVUXDDP_SCALE,    "xvcvuxddp_scale", CONST, vsx_xvcvuxddp_scale)
+BU_VSX_2 (XVCVDPSXDS_SCALE,   "xvcvdpsxds_scale", CONST, vsx_xvcvdpsxds_scale)
+BU_VSX_2 (XVCVDPUXDS_SCALE,   "xvcvdpuxds_scale", CONST, vsx_xvcvdpuxds_scale)
+
 /* VSX abs builtin functions.  */
 BU_VSX_A (XVABSDP,	      "xvabsdp",	CONST,	absv2df2)
 BU_VSX_A (XVNABSDP,	      "xvnabsdp",	CONST,	vsx_nabsv2df2)
Index: gcc/config/rs6000/rs6000-c.c
===================================================================
--- gcc/config/rs6000/rs6000-c.c	(revision 214735)
+++ gcc/config/rs6000/rs6000-c.c	(working copy)
@@ -1151,6 +1151,10 @@  const struct altivec_builtin_types altivec_overloa
     RS6000_BTI_V4SF, RS6000_BTI_unsigned_V4SI, RS6000_BTI_INTSI, 0 },
   { ALTIVEC_BUILTIN_VEC_CTF, ALTIVEC_BUILTIN_VCFSX,
     RS6000_BTI_V4SF, RS6000_BTI_V4SI, RS6000_BTI_INTSI, 0 },
+  { ALTIVEC_BUILTIN_VEC_CTF, VSX_BUILTIN_XVCVSXDDP_SCALE,
+    RS6000_BTI_V2DF, RS6000_BTI_V2DI, RS6000_BTI_INTSI, 0},
+  { ALTIVEC_BUILTIN_VEC_CTF, VSX_BUILTIN_XVCVUXDDP_SCALE,
+    RS6000_BTI_V2DF, RS6000_BTI_unsigned_V2DI, RS6000_BTI_INTSI, 0},
   { ALTIVEC_BUILTIN_VEC_VCFSX, ALTIVEC_BUILTIN_VCFSX,
     RS6000_BTI_V4SF, RS6000_BTI_V4SI, RS6000_BTI_INTSI, 0 },
   { ALTIVEC_BUILTIN_VEC_VCFUX, ALTIVEC_BUILTIN_VCFUX,
@@ -1157,8 +1161,12 @@  const struct altivec_builtin_types altivec_overloa
     RS6000_BTI_V4SF, RS6000_BTI_unsigned_V4SI, RS6000_BTI_INTSI, 0 },
   { ALTIVEC_BUILTIN_VEC_CTS, ALTIVEC_BUILTIN_VCTSXS,
     RS6000_BTI_V4SI, RS6000_BTI_V4SF, RS6000_BTI_INTSI, 0 },
+  { ALTIVEC_BUILTIN_VEC_CTS, VSX_BUILTIN_XVCVDPSXDS_SCALE,
+    RS6000_BTI_V2DI, RS6000_BTI_V2DF, RS6000_BTI_INTSI, 0 },
   { ALTIVEC_BUILTIN_VEC_CTU, ALTIVEC_BUILTIN_VCTUXS,
     RS6000_BTI_unsigned_V4SI, RS6000_BTI_V4SF, RS6000_BTI_INTSI, 0 },
+  { ALTIVEC_BUILTIN_VEC_CTU, VSX_BUILTIN_XVCVDPUXDS_SCALE,
+    RS6000_BTI_unsigned_V2DI, RS6000_BTI_V2DF, RS6000_BTI_INTSI, 0 },
   { VSX_BUILTIN_VEC_DIV, VSX_BUILTIN_XVDIVSP,
     RS6000_BTI_V4SF, RS6000_BTI_V4SF, RS6000_BTI_V4SF, 0 },
   { VSX_BUILTIN_VEC_DIV, VSX_BUILTIN_XVDIVDP,
Index: gcc/config/rs6000/rs6000-protos.h
===================================================================
--- gcc/config/rs6000/rs6000-protos.h	(revision 214735)
+++ gcc/config/rs6000/rs6000-protos.h	(working copy)
@@ -65,6 +65,7 @@  extern void altivec_expand_stvx_be (rtx, rtx, enum
 extern void altivec_expand_stvex_be (rtx, rtx, enum machine_mode, unsigned);
 extern void rs6000_expand_extract_even (rtx, rtx, rtx);
 extern void rs6000_expand_interleave (rtx, rtx, rtx, bool);
+extern void rs6000_scale_v2df (rtx, rtx, int);
 extern void build_mask64_2_operands (rtx, rtx *);
 extern int expand_block_clear (rtx[]);
 extern int expand_block_move (rtx[]);
Index: gcc/config/rs6000/rs6000.c
===================================================================
--- gcc/config/rs6000/rs6000.c	(revision 214735)
+++ gcc/config/rs6000/rs6000.c	(working copy)
@@ -81,6 +81,7 @@ 
 #include "builtins.h"
 #include "context.h"
 #include "tree-pass.h"
+#include "real.h"
 #if TARGET_XCOFF
 #include "xcoffout.h"  /* get declarations of xcoff_*_section_name */
 #endif
@@ -31245,6 +31246,23 @@  rs6000_expand_interleave (rtx target, rtx op0, rtx
   rs6000_do_expand_vec_perm (target, op0, op1, vmode, nelt, perm);
 }
 
+/* Scale a V2DF vector SRC by two to the SCALE and place in TGT.  */
+void
+rs6000_scale_v2df (rtx tgt, rtx src, int scale)
+{
+  HOST_WIDE_INT hwi_scale (scale);
+  REAL_VALUE_TYPE r_pow;
+  rtvec v = rtvec_alloc (2);
+  rtx elt;
+  rtx scale_vec = gen_reg_rtx (V2DFmode);
+  (void)real_powi (&r_pow, DFmode, &dconst2, hwi_scale);
+  elt = CONST_DOUBLE_FROM_REAL_VALUE (r_pow, DFmode);
+  RTVEC_ELT (v, 0) = elt;
+  RTVEC_ELT (v, 1) = elt;
+  rs6000_expand_vector_init (scale_vec, gen_rtx_PARALLEL (V2DFmode, v));
+  emit_insn (gen_mulv2df3 (tgt, src, scale_vec));
+}
+
 /* Return an RTX representing where to find the function value of a
    function returning MODE.  */
 static rtx
Index: gcc/config/rs6000/vsx.md
===================================================================
--- gcc/config/rs6000/vsx.md	(revision 214735)
+++ gcc/config/rs6000/vsx.md	(working copy)
@@ -264,6 +264,10 @@ 
    UNSPEC_VSX_DIVSD
    UNSPEC_VSX_DIVUD
    UNSPEC_VSX_MULSD
+   UNSPEC_VSX_XVCVSXDDP
+   UNSPEC_VSX_XVCVUXDDP
+   UNSPEC_VSX_XVCVDPSXDS
+   UNSPEC_VSX_XVCVDPUXDS
   ])
 
 ;; VSX moves
@@ -1356,6 +1360,102 @@ 
   "xscvspdpn %x0,%x1"
   [(set_attr "type" "fp")])
 
+;; Convert and scale (used by vec_ctf, vec_cts, vec_ctu for double/long long)
+
+(define_expand "vsx_xvcvsxddp_scale"
+  [(match_operand:V2DF 0 "vsx_register_operand" "")
+   (match_operand:V2DI 1 "vsx_register_operand" "")
+   (match_operand:QI 2 "immediate_operand" "")]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+{
+  rtx op0 = operands[0];
+  rtx op1 = operands[1];
+  int scale = INTVAL(operands[2]);
+  emit_insn (gen_vsx_xvcvsxddp (op0, op1));
+  if (scale != 0)
+    rs6000_scale_v2df (op0, op0, -scale);
+  DONE;
+})
+
+(define_insn "vsx_xvcvsxddp"
+  [(set (match_operand:V2DF 0 "vsx_register_operand" "=wa")
+        (unspec:V2DF [(match_operand:V2DI 1 "vsx_register_operand" "wa")]
+                     UNSPEC_VSX_XVCVSXDDP))]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+  "xvcvsxddp %x0,%x1"
+  [(set_attr "type" "vecdouble")])
+
+(define_expand "vsx_xvcvuxddp_scale"
+  [(match_operand:V2DF 0 "vsx_register_operand" "")
+   (match_operand:V2DI 1 "vsx_register_operand" "")
+   (match_operand:QI 2 "immediate_operand" "")]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+{
+  rtx op0 = operands[0];
+  rtx op1 = operands[1];
+  int scale = INTVAL(operands[2]);
+  emit_insn (gen_vsx_xvcvuxddp (op0, op1));
+  if (scale != 0)
+    rs6000_scale_v2df (op0, op0, -scale);
+  DONE;
+})
+
+(define_insn "vsx_xvcvuxddp"
+  [(set (match_operand:V2DF 0 "vsx_register_operand" "=wa")
+        (unspec:V2DF [(match_operand:V2DI 1 "vsx_register_operand" "wa")]
+                     UNSPEC_VSX_XVCVUXDDP))]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+  "xvcvuxddp %x0,%x1"
+  [(set_attr "type" "vecdouble")])
+
+(define_expand "vsx_xvcvdpsxds_scale"
+  [(match_operand:V2DI 0 "vsx_register_operand" "")
+   (match_operand:V2DF 1 "vsx_register_operand" "")
+   (match_operand:QI 2 "immediate_operand" "")]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+{
+  rtx op0 = operands[0];
+  rtx op1 = operands[1];
+  rtx tmp = gen_reg_rtx (V2DFmode);
+  int scale = INTVAL(operands[2]);
+  if (scale != 0)
+    rs6000_scale_v2df (tmp, op1, scale);
+  emit_insn (gen_vsx_xvcvdpsxds (op0, tmp));
+  DONE;
+})
+
+(define_insn "vsx_xvcvdpsxds"
+  [(set (match_operand:V2DI 0 "vsx_register_operand" "=wa")
+        (unspec:V2DI [(match_operand:V2DF 1 "vsx_register_operand" "wa")]
+                     UNSPEC_VSX_XVCVDPSXDS))]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+  "xvcvdpsxds %x0,%x1"
+  [(set_attr "type" "vecdouble")])
+
+(define_expand "vsx_xvcvdpuxds_scale"
+  [(match_operand:V2DI 0 "vsx_register_operand" "")
+   (match_operand:V2DF 1 "vsx_register_operand" "")
+   (match_operand:QI 2 "immediate_operand" "")]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+{
+  rtx op0 = operands[0];
+  rtx op1 = operands[1];
+  rtx tmp = gen_reg_rtx (V2DFmode);
+  int scale = INTVAL(operands[2]);
+  if (scale != 0)
+    rs6000_scale_v2df (tmp, op1, scale);
+  emit_insn (gen_vsx_xvcvdpuxds (op0, tmp));
+  DONE;
+})
+
+(define_insn "vsx_xvcvdpuxds"
+  [(set (match_operand:V2DI 0 "vsx_register_operand" "=wa")
+        (unspec:V2DI [(match_operand:V2DF 1 "vsx_register_operand" "wa")]
+                     UNSPEC_VSX_XVCVDPUXDS))]
+  "VECTOR_UNIT_VSX_P (V2DFmode)"
+  "xvcvdpuxds %x0,%x1"
+  [(set_attr "type" "vecdouble")])
+
 ;; Convert from 64-bit to 32-bit types
 ;; Note, favor the Altivec registers since the usual use of these instructions
 ;; is in vector converts and we need to use the Altivec vperm instruction.
Index: gcc/doc/extend.texi
===================================================================
--- gcc/doc/extend.texi	(revision 214735)
+++ gcc/doc/extend.texi	(working copy)
@@ -13882,6 +13882,8 @@  vector float vec_cpsgn (vector float, vector float
 
 vector float vec_ctf (vector unsigned int, const int);
 vector float vec_ctf (vector signed int, const int);
+vector double vec_ctf (vector unsigned long, const int);
+vector double vec_ctf (vector signed long, const int);
 
 vector float vec_vcfsx (vector signed int, const int);
 
@@ -13888,8 +13890,10 @@  vector float vec_vcfsx (vector signed int, const i
 vector float vec_vcfux (vector unsigned int, const int);
 
 vector signed int vec_cts (vector float, const int);
+vector signed long vec_cts (vector double, const int);
 
 vector unsigned int vec_ctu (vector float, const int);
+vector unsigned long vec_ctu (vector double, const int);
 
 void vec_dss (const int);
 
@@ -14725,6 +14729,8 @@  vector float vec_splat (vector float, const int);
 vector signed int vec_splat (vector signed int, const int);
 vector unsigned int vec_splat (vector unsigned int, const int);
 vector bool int vec_splat (vector bool int, const int);
+vector signed long vec_splat (vector signed long, const int);
+vector unsigned long vec_splat (vector unsigned long, const int);
 
 vector signed char vec_splats (signed char);
 vector unsigned char vec_splats (unsigned char);
@@ -15461,6 +15467,8 @@  vector bool long vec_cmplt (vector double, vector
 vector double vec_cpsgn (vector double, vector double);
 vector float vec_div (vector float, vector float);
 vector double vec_div (vector double, vector double);
+vector long vec_div (vector long, vector long);
+vector unsigned long vec_div (vector unsigned long, vector unsigned long);
 vector double vec_floor (vector double);
 vector double vec_ld (int, const vector double *);
 vector double vec_ld (int, const double *);
@@ -15487,6 +15495,8 @@  vector float vec_msub (vector float, vector float,
 vector double vec_msub (vector double, vector double, vector double);
 vector float vec_mul (vector float, vector float);
 vector double vec_mul (vector double, vector double);
+vector long vec_mul (vector long, vector long);
+vector unsigned long vec_mul (vector unsigned long, vector unsigned long);
 vector float vec_nearbyint (vector float);
 vector double vec_nearbyint (vector double);
 vector float vec_nmadd (vector float, vector float, vector float);
Index: gcc/testsuite/gcc.target/powerpc/builtins-1.c
===================================================================
--- gcc/testsuite/gcc.target/powerpc/builtins-1.c	(revision 214735)
+++ gcc/testsuite/gcc.target/powerpc/builtins-1.c	(working copy)
@@ -157,5 +157,10 @@  int main ()
   vector long long l8 = vec_mul (l3, l4);
   vector unsigned long long u6 = vec_mul (u3, u4);
 
+  vector double dh = vec_ctf (la, -2);
+  vector double di = vec_ctf (ua, 2);
+  vector long long l9 = vec_cts (dh, -2);
+  vector unsigned long long u7 = vec_ctu (di, 2);
+
   return 0;
 }
Index: gcc/testsuite/gcc.target/powerpc/builtins-2.c
===================================================================
--- gcc/testsuite/gcc.target/powerpc/builtins-2.c	(revision 214735)
+++ gcc/testsuite/gcc.target/powerpc/builtins-2.c	(working copy)
@@ -34,5 +34,14 @@  int main ()
       || ue[0] != 27L || ue[1] != 27L || uf[0] != 14L || uf[1] != 14L)
     abort ();
 
+  vector double da = vec_ctf (sa, -2);
+  vector double db = vec_ctf (ua, 2);
+  vector long long sg = vec_cts (da, -2);
+  vector unsigned long long ug = vec_ctu (db, 2);
+
+  if (da[0] != 108.0 || da[1] != -56.0 || db[0] != 6.75 || db[1] != 3.5
+      || sg[0] != 27L || sg[1] != -14L || ug[0] != 27L || ug[1] != 14L)
+    abort ();
+
   return 0;
 }