* [PATCH v3, rs6000] gcc mainline, add builtin support for vec_doublee, vec_doubleo, vec_doublel builtins
@ 2017-06-01 21:55 Carl E. Love
2017-06-01 22:42 ` Segher Boessenkool
0 siblings, 1 reply; 2+ messages in thread
From: Carl E. Love @ 2017-06-01 21:55 UTC (permalink / raw)
To: gcc-patches, David Edelsohn, Segher Boessenkool; +Cc: Bill Schmidt, cel
GCC Maintainers:
This is version 3 of the patch to add support for the various
vec_doublee, vec_doubleo, vec_doublel, vec_doubleh builtin-ins. I have
addressed the formatting comments from Segher on version 2. I have run
the patch through contrib/check_GNU_style.sh to check for obvious
formatting errors.
The patch has been tested on powerpc64le-unknown-linux-gnu (Power 8 LE)
with no regressions.
Is the patch OK for gcc mainline?
Carl Love
------------------------------------------------
Add vec_doublee, vec_doubleo, vec_doublel, vec_doubleh builtin-ins
gcc/ChangeLog:
2017-06-01 Carl Love <cel@us.ibm.com>
* config/rs6000/rs6000-c: Add support for built-in functions
vector double vec_doublee (vector signed int);
vector double vec_doublee (vector unsigned int);
vector double vec_doublee (vector float);
vector double vec_doubleh (vector signed int);
vector double vec_doubleh (vector unsigned int);
vector double vec_doubleh (vector float);
vector double vec_doublel (vector signed int);
vector double vec_doublel (vector unsigned int);
vector double vec_doublel (vector float);
vector double vec_doubleo (vector signed int);
vector double vec_doubleo (vector unsigned int);
vector double vec_doubleo (vector float);.
* config/rs6000/rs6000-builtin.def: Add definitions for DOUBLEE,
DOUBLEO, DOUBLEH, DOUBLEL, UNS_DOUBLEO, UNS_DOUBLEE, UNS_DOUBLEH,
UNS_DOUBLEL.
* config/rs6000/altivec.md: Add code generator for doublee<mode>2,
unsdoubleev4si2, doubleo<mode>2, unsdoubleov4si2, doubleh<mode>2,
unsdoublehv4si2, doublel<mode>2, unsdoublelv4si2, add mode attribute
VS_sxwsp.
* config/rs6000/altivec.h: Add define for vec_doublee, vec_doubleo,
vec_doublel, vec_doubleh.
* doc/extend.texi: Update the built-in documentation file for the
new built-in functions.
2017-06-01 Carl Love <cel@us.ibm.com>
gcc/testsuite/ChangeLog:
* gcc.target/powerpc/builtins-3-runnable.c: New file of runnable tests
for the new built-ins.
Signed-off-by: Carl Love <carll@us.ibm.com>
---
gcc/config/rs6000/altivec.h | 4 +
gcc/config/rs6000/altivec.md | 337 +++++++++++++++++++++
gcc/config/rs6000/rs6000-builtin.def | 21 ++
gcc/config/rs6000/rs6000-c.c | 29 ++
gcc/doc/extend.texi | 16 +
.../gcc.target/powerpc/builtins-3-runnable.c | 83 +++++
6 files changed, 490 insertions(+)
create mode 100644 gcc/testsuite/gcc.target/powerpc/builtins-3-runnable.c
diff --git a/gcc/config/rs6000/altivec.h b/gcc/config/rs6000/altivec.h
index c92bcce..20050eb 100644
--- a/gcc/config/rs6000/altivec.h
+++ b/gcc/config/rs6000/altivec.h
@@ -128,6 +128,10 @@
#define vec_ctu __builtin_vec_ctu
#define vec_cpsgn __builtin_vec_copysign
#define vec_double __builtin_vec_double
+#define vec_doublee __builtin_vec_doublee
+#define vec_doubleo __builtin_vec_doubleo
+#define vec_doublel __builtin_vec_doublel
+#define vec_doubleh __builtin_vec_doubleh
#define vec_expte __builtin_vec_expte
#define vec_floor __builtin_vec_floor
#define vec_loge __builtin_vec_loge
diff --git a/gcc/config/rs6000/altivec.md b/gcc/config/rs6000/altivec.md
index 649f181..af1fae3 100644
--- a/gcc/config/rs6000/altivec.md
+++ b/gcc/config/rs6000/altivec.md
@@ -203,6 +203,10 @@
(KF "FLOAT128_VECTOR_P (KFmode)")
(TF "FLOAT128_VECTOR_P (TFmode)")])
+;; Map the Vector convert single precision to double precision for integer
+;; versus floating point
+(define_mode_attr VS_sxwsp [(V4SI "sxw") (V4SF "sp")])
+
;; Specific iterator for parity which does not have a byte/half-word form, but
;; does have a quad word form
(define_mode_iterator VParity [V4SI
@@ -2739,6 +2743,339 @@
"stvewx %1,%y0"
[(set_attr "type" "vecstore")])
+;; Generate doublee
+;; signed int/float to double convert words 0 and 2
+(define_expand "doublee<mode>2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:VSX_W 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ machine_mode op_mode = GET_MODE (operands[1]);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ Input words 0 and 2 are where they need to be. */
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], operands[1]));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ take (operand[1] operand[1]) and shift left one word
+ 3 2 1 0 3 2 1 0 => 2 1 0 3
+ Input words 2 and 0 are now where they need to be for the
+ conversion. */
+ rtx rtx_tmp;
+ rtx rtx_val = GEN_INT (1);
+
+ rtx_tmp = gen_reg_rtx (op_mode);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate unsdoublee
+;; unsigned int to double convert words 0 and 2
+(define_expand "unsdoubleev4si2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:V4SI 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ Input words 0 and 2 are where they need to be. */
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], operands[1]));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ take (operand[1] operand[1]) and shift left one word
+ 3 2 1 0 3 2 1 0 => 2 1 0 3
+ Input words 2 and 0 are now where they need to be for the
+ conversion. */r
+ rtx rtx_tmp;
+ rtx rtx_val = GEN_INT (1);
+
+ rtx_tmp = gen_reg_rtx (V4SImode);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate doubleov
+;; signed int/float to double convert words 1 and 3
+(define_expand "doubleo<mode>2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:VSX_W 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ machine_mode op_mode = GET_MODE (operands[1]);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ take (operand[1] operand[1]) and shift left one word
+ 0 1 2 3 0 1 2 3 => 1 2 3 0
+ Input words 1 and 3 are now where they need to be for the
+ conversion. */
+ rtx rtx_tmp;
+ rtx rtx_val = GEN_INT (1);
+
+ rtx_tmp = gen_reg_rtx (op_mode);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ Input words 3 and 1 are where they need to be. */
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], operands[1]));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate unsdoubleov
+;; unsigned int to double convert words 1 and 3
+(define_expand "unsdoubleov4si2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:V4SI 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ take (operand[1] operand[1]) and shift left one word
+ 0 1 2 3 0 1 2 3 => 1 2 3 0
+ Input words 1 and 3 are now where they need to be for the
+ conversion. */
+ rtx rtx_tmp;
+ rtx rtx_val = GEN_INT (1);
+
+ rtx_tmp = gen_reg_rtx (V4SImode);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Want to convert the words 1 and 3.
+ Little endian word numbering for operand is 3 2 1 0.
+ Input words 3 and 1 are where they need to be. */
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], operands[1]));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate doublehv
+;; signed int/float to double convert words 0 and 1
+(define_expand "doubleh<mode>2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:VSX_W 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ rtx rtx_tmp;
+ rtx rtx_val;
+
+ machine_mode op_mode = GET_MODE (operands[1]);
+ rtx_tmp = gen_reg_rtx (op_mode);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ Shift operand left one word, rtx_tmp word order is now 1 2 3 0.
+ take (rts_tmp operand[1]) and shift left three words
+ 1 2 3 0 0 1 2 3 => 0 0 1 2
+ Input words 0 and 1 are now where they need to be for the
+ conversion. */
+ rtx_val = GEN_INT (1);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, rtx_tmp,
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ Shift operand left three words, rtx_tmp word order is now 0 3 2 1.
+ take (operand[1] rts_tmp) and shift left two words
+ 3 2 1 0 0 3 2 1 => 1 0 0 3
+ Input words 0 and 1 are now where they need to be for the
+ conversion. */
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (2);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ rtx_tmp, rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate unsdoublehv
+;; unsigned int to double convert words 0 and 1
+(define_expand "unsdoublehv4si2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:V4SI 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ rtx rtx_tmp = gen_reg_rtx (V4SImode);
+ rtx rtx_val = GEN_INT (12);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for words in operand is 0 1 2 3.
+ Shift operand left one word, rtx_tmp word order is now 1 2 3 0.
+ take (rts_tmp operand[1]) and shift left three words
+ 1 2 3 0 0 1 2 3 => 0 0 1 2
+ Input words 0 and 1 are now where they need to be for the
+ conversion. */
+ rtx_val = GEN_INT (1);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, rtx_tmp,
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ Shift operand left three words, rtx_tmp word order is now 0 3 2 1.
+ take (operand[1] rts_tmp) and shift left two words
+ 3 2 1 0 0 3 2 1 => 1 0 0 3
+ Input words 1 and 0 are now where they need to be for the
+ conversion. */
+ rtx_val = GEN_INT (3);
+
+ rtx_tmp = gen_reg_rtx (V4SImode);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (2);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ rtx_tmp, rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate doublelv
+;; signed int/float to double convert words 2 and 3
+(define_expand "doublel<mode>2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:VSX_W 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ rtx rtx_tmp;
+ rtx rtx_val = GEN_INT (3);
+
+ machine_mode op_mode = GET_MODE (operands[1]);
+ rtx_tmp = gen_reg_rtx (op_mode);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for operand is 0 1 2 3.
+ Shift operand left three words, rtx_tmp word order is now 3 0 1 2.
+ take (operand[1] rtx_tmp) and shift left two words
+ 0 1 2 3 3 0 1 2 => 2 3 3 0
+ now use convert instruction to convert word 2 and 3 in the
+ input vector. */
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (8);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ rtx_tmp, rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ Shift operand left one word, rtx_tmp word order is now 2 1 0 3.
+ take (rtx_tmp operand[1]) and shift left three words
+ 2 1 0 3 3 2 1 0 => 3 3 2 1
+ now use convert instruction to convert word 3 and 2 in the
+ input vector. */
+ rtx_val = GEN_INT (1);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_<mode> (rtx_tmp, rtx_tmp,
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
+;; Generate unsdoublelv
+;; unsigned int to double convert convert 2 and 3
+(define_expand "unsdoublelv4si2"
+ [(set (match_operand:V2DF 0 "register_operand" "=v")
+ (match_operand:V4SI 1 "register_operand" "v"))]
+ "TARGET_VSX"
+{
+ rtx rtx_tmp = gen_reg_rtx (V4SImode);
+ rtx rtx_val = GEN_INT (12);
+
+ if (VECTOR_ELT_ORDER_BIG)
+ {
+ /* Big endian word numbering for operand is 0 1 2 3.
+ Shift operand left three words, rtx_tmp word order is now 3 0 1 2.
+ take (operand[1] rtx_tmp) and shift left two words
+ 0 1 2 3 3 0 1 2 => 2 3 3 0
+ now use convert instruction to convert word 2 and 3 in the
+ input vector. */
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ operands[1], rtx_val));
+
+ rtx_val = GEN_INT (2);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, operands[1],
+ rtx_tmp, rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ else
+ {
+ /* Little endian word numbering for operand is 3 2 1 0.
+ Shift operand left one word, rtx_tmp word order is now 2 1 0 3.
+ take (rtx_tmp operand[1]) and shift left three words
+ 2 1 0 3 3 2 1 0 => 3 3 2 1
+ now use convert instruction to convert word 3 and 2 in the
+ input vector. */
+ rtx_val = GEN_INT (1);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp,
+ operands[1], operands[1], rtx_val));
+
+ rtx_val = GEN_INT (3);
+ emit_insn (gen_vsx_xxsldwi_v4si (rtx_tmp, rtx_tmp,
+ operands[1], rtx_val));
+ emit_insn (gen_vsx_xvcvuxwdp (operands[0], rtx_tmp));
+ }
+ DONE;
+}
+ [(set_attr "type" "veccomplex")])
+
;; Generate
;; xxlxor/vxor SCRATCH0,SCRATCH0,SCRATCH0
;; vsubu?m SCRATCH2,SCRATCH1,%1
diff --git a/gcc/config/rs6000/rs6000-builtin.def b/gcc/config/rs6000/rs6000-builtin.def
index ebe005a..de6b2cc 100644
--- a/gcc/config/rs6000/rs6000-builtin.def
+++ b/gcc/config/rs6000/rs6000-builtin.def
@@ -1719,6 +1719,19 @@ BU_VSX_1 (XSRDPIM, "xsrdpim", CONST, floordf2)
BU_VSX_1 (XSRDPIP, "xsrdpip", CONST, ceildf2)
BU_VSX_1 (XSRDPIZ, "xsrdpiz", CONST, btruncdf2)
+BU_VSX_1 (DOUBLEE_V4SI, "doublee_v4si", CONST, doubleev4si2)
+BU_VSX_1 (DOUBLEE_V4SF, "doublee_v4sf", CONST, doubleev4sf2)
+BU_VSX_1 (UNS_DOUBLEE_V4SI, "uns_doublee_v4si", CONST, unsdoubleev4si2)
+BU_VSX_1 (DOUBLEO_V4SI, "doubleo_v4si", CONST, doubleov4si2)
+BU_VSX_1 (DOUBLEO_V4SF, "doubleo_v4sf", CONST, doubleov4sf2)
+BU_VSX_1 (UNS_DOUBLEO_V4SI, "uns_doubleo_v4si", CONST, unsdoubleov4si2)
+BU_VSX_1 (DOUBLEH_V4SI, "doubleh_v4si", CONST, doublehv4si2)
+BU_VSX_1 (DOUBLEH_V4SF, "doubleh_v4sf", CONST, doublehv4sf2)
+BU_VSX_1 (UNS_DOUBLEH_V4SI, "uns_doubleh_v4si", CONST, unsdoublehv4si2)
+BU_VSX_1 (DOUBLEL_V4SI, "doublel_v4si", CONST, doublelv4si2)
+BU_VSX_1 (DOUBLEL_V4SF, "doublel_v4sf", CONST, doublelv4sf2)
+BU_VSX_1 (UNS_DOUBLEL_V4SI, "uns_doublel_v4si", CONST, unsdoublelv4si2)
+
/* VSX predicate functions. */
BU_VSX_P (XVCMPEQSP_P, "xvcmpeqsp_p", CONST, vector_eq_v4sf_p)
BU_VSX_P (XVCMPGESP_P, "xvcmpgesp_p", CONST, vector_ge_v4sf_p)
@@ -1809,6 +1822,14 @@ BU_VSX_OVERLOAD_2 (XXSPLTW, "xxspltw")
/* 1 argument VSX overloaded builtin functions. */
BU_VSX_OVERLOAD_1 (DOUBLE, "double")
+BU_VSX_OVERLOAD_1 (DOUBLEE, "doublee")
+BU_VSX_OVERLOAD_1 (UNS_DOUBLEE, "uns_doublee")
+BU_VSX_OVERLOAD_1 (DOUBLEO, "doubleo")
+BU_VSX_OVERLOAD_1 (UNS_DOUBLEO, "uns_doubleo")
+BU_VSX_OVERLOAD_1 (DOUBLEH, "doubleh")
+BU_VSX_OVERLOAD_1 (UNS_DOUBLEH, "uns_doubleh")
+BU_VSX_OVERLOAD_1 (DOUBLEL, "doublel")
+BU_VSX_OVERLOAD_1 (UNS_DOUBLEL, "uns_doublel")
/* VSX builtins that are handled as special cases. */
BU_VSX_OVERLOAD_X (LD, "ld")
diff --git a/gcc/config/rs6000/rs6000-c.c b/gcc/config/rs6000/rs6000-c.c
index 8039814..b602dee 100644
--- a/gcc/config/rs6000/rs6000-c.c
+++ b/gcc/config/rs6000/rs6000-c.c
@@ -1513,6 +1513,35 @@ const struct altivec_builtin_types altivec_overloaded_builtins[] = {
RS6000_BTI_V2DF, RS6000_BTI_V2DI, 0, 0 },
{ VSX_BUILTIN_VEC_DOUBLE, VSX_BUILTIN_XVCVUXDDP,
RS6000_BTI_V2DF, RS6000_BTI_unsigned_V2DI, 0, 0 },
+
+ { VSX_BUILTIN_VEC_DOUBLEE, VSX_BUILTIN_DOUBLEE_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEE, VSX_BUILTIN_UNS_DOUBLEE_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_unsigned_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEE, VSX_BUILTIN_DOUBLEE_V4SF,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SF, 0, 0 },
+
+ { VSX_BUILTIN_VEC_DOUBLEO, VSX_BUILTIN_DOUBLEO_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEO, VSX_BUILTIN_UNS_DOUBLEO_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_unsigned_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEO, VSX_BUILTIN_DOUBLEO_V4SF,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SF, 0, 0 },
+
+ { VSX_BUILTIN_VEC_DOUBLEH, VSX_BUILTIN_DOUBLEH_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEH, VSX_BUILTIN_UNS_DOUBLEH_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_unsigned_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEH, VSX_BUILTIN_DOUBLEH_V4SF,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SF, 0, 0 },
+
+ { VSX_BUILTIN_VEC_DOUBLEL, VSX_BUILTIN_DOUBLEL_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEL, VSX_BUILTIN_UNS_DOUBLEL_V4SI,
+ RS6000_BTI_V2DF, RS6000_BTI_unsigned_V4SI, 0, 0 },
+ { VSX_BUILTIN_VEC_DOUBLEL, VSX_BUILTIN_DOUBLEL_V4SF,
+ RS6000_BTI_V2DF, RS6000_BTI_V4SF, 0, 0 },
+
{ ALTIVEC_BUILTIN_VEC_LD, ALTIVEC_BUILTIN_LVX_V2DF,
RS6000_BTI_V2DF, RS6000_BTI_INTSI, ~RS6000_BTI_V2DF, 0 },
{ ALTIVEC_BUILTIN_VEC_LD, ALTIVEC_BUILTIN_LVX_V2DI,
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index 3511d25..a895ed4 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -15872,6 +15872,22 @@ 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);
+vector double vec_doublee (vector float);
+vector double vec_doublee (vector signed int);
+vector double vec_doublee (vector unsigned int);
+
+vector double vec_doubleo (vector float);
+vector double vec_doubleo (vector signed int);
+vector double vec_doubleo (vector unsigned int);
+
+vector double vec_doubleh (vector float);
+vector double vec_doubleh (vector signed int);
+vector double vec_doubleh (vector unsigned int);
+
+vector double vec_doublel (vector float);
+vector double vec_doublel (vector signed int);
+vector double vec_doublel (vector unsigned int);
+
void vec_dss (const int);
void vec_dssall (void);
diff --git a/gcc/testsuite/gcc.target/powerpc/builtins-3-runnable.c b/gcc/testsuite/gcc.target/powerpc/builtins-3-runnable.c
new file mode 100644
index 0000000..17a7786
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/builtins-3-runnable.c
@@ -0,0 +1,83 @@
+/* { dg-do run { target { powerpc*-*-linux* } } } */
+/* { dg-require-effective-target vsx_hw } */
+/* { dg-options "-O2 -mvsx -mcpu=power8" } */
+
+#include <altivec.h> // vector
+
+void abort (void);
+
+void test_result_dp(vector double vec_result, vector double vec_expected)
+{
+ if (vec_result[0] != vec_expected[0])
+ abort();
+
+ if (vec_result[1] != vec_expected[1])
+ abort();
+}
+
+int main()
+{
+ int i;
+ vector unsigned int vec_unint;
+ vector signed int vec_int;
+ vector float vec_flt, vec_flt_result, vec_flt_expected;
+ vector double vec_dble0, vec_dble1, vec_dble_result, vec_dble_expected;
+
+ vec_int = (vector signed int){ -1, 3, -5, 1234567 };
+ vec_unint = (vector unsigned int){ 9, 11, 15, 2468013579 };
+ vec_flt = (vector float){ -21., 3.5, -53., 78. };
+ vec_dble0 = (vector double){ 34.0, 97.0 };
+ vec_dble1 = (vector double){ 214.0, -5.5 };
+
+ /* conversion of words 0 and 2 */
+ vec_dble_expected = (vector double){-1.000000, -5.000000};
+ vec_dble_result = vec_doublee (vec_int);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){9.000000, 15.000000};
+ vec_dble_result = vec_doublee (vec_unint);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){-21.000000, -53.000000};
+ vec_dble_result = vec_doublee (vec_flt);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+
+ /* conversion of words 1 and 3 */
+ vec_dble_expected = (vector double){3.000000, 1234567.000000};
+ vec_dble_result = vec_doubleo (vec_int);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){11.000000, 2468013579.000000};
+ vec_dble_result = vec_doubleo (vec_unint);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){3.500000, 78.000000};
+ vec_dble_result = vec_doubleo (vec_flt);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+
+ /* conversion of words 0 and 1 */
+ vec_dble_expected = (vector double){-5.000000, 1234567.000000};
+ vec_dble_result = vec_doublel (vec_int);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){15.000000, 2468013579.000000};
+ vec_dble_result = vec_doublel (vec_unint);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){-53.000000, 78.000000};
+ vec_dble_result = vec_doublel (vec_flt);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+
+ /* conversion of words 2 and 3 */
+ vec_dble_expected = (vector double){-1.000000, 3.000000};
+ vec_dble_result = vec_doubleh (vec_int);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+ vec_dble_expected = (vector double){9.000000, 11.000000};
+ vec_dble_result = vec_doubleh (vec_unint);
+ test_result_dp(vec_dble_result, vec_dble_expected);
+
+}
--
1.9.1
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH v3, rs6000] gcc mainline, add builtin support for vec_doublee, vec_doubleo, vec_doublel builtins
2017-06-01 21:55 [PATCH v3, rs6000] gcc mainline, add builtin support for vec_doublee, vec_doubleo, vec_doublel builtins Carl E. Love
@ 2017-06-01 22:42 ` Segher Boessenkool
0 siblings, 0 replies; 2+ messages in thread
From: Segher Boessenkool @ 2017-06-01 22:42 UTC (permalink / raw)
To: Carl E. Love; +Cc: gcc-patches, David Edelsohn, Bill Schmidt
Hi Carl,
On Thu, Jun 01, 2017 at 02:55:45PM -0700, Carl E. Love wrote:
> Add vec_doublee, vec_doubleo, vec_doublel, vec_doubleh builtin-ins
>
> gcc/ChangeLog:
>
> 2017-06-01 Carl Love <cel@us.ibm.com>
>
> * config/rs6000/rs6000-c: Add support for built-in functions
> vector double vec_doublee (vector signed int);
> vector double vec_doublee (vector unsigned int);
> vector double vec_doublee (vector float);
> vector double vec_doubleh (vector signed int);
> vector double vec_doubleh (vector unsigned int);
> vector double vec_doubleh (vector float);
> vector double vec_doublel (vector signed int);
> vector double vec_doublel (vector unsigned int);
> vector double vec_doublel (vector float);
> vector double vec_doubleo (vector signed int);
> vector double vec_doubleo (vector unsigned int);
> vector double vec_doubleo (vector float);.
> * config/rs6000/rs6000-builtin.def: Add definitions for DOUBLEE,
> DOUBLEO, DOUBLEH, DOUBLEL, UNS_DOUBLEO, UNS_DOUBLEE, UNS_DOUBLEH,
> UNS_DOUBLEL.
> * config/rs6000/altivec.md: Add code generator for doublee<mode>2,
> unsdoubleev4si2, doubleo<mode>2, unsdoubleov4si2, doubleh<mode>2,
> unsdoublehv4si2, doublel<mode>2, unsdoublelv4si2, add mode attribute
> VS_sxwsp.
> * config/rs6000/altivec.h: Add define for vec_doublee, vec_doubleo,
> vec_doublel, vec_doubleh.
> * doc/extend.texi: Update the built-in documentation file for the
> new built-in functions.
>
> 2017-06-01 Carl Love <cel@us.ibm.com>
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/powerpc/builtins-3-runnable.c: New file of runnable tests
> for the new built-ins.
>
> Signed-off-by: Carl Love <carll@us.ibm.com>
We don't do signoffs in GCC, fwiw.
> +(define_expand "doublee<mode>2"
> + [(set (match_operand:V2DF 0 "register_operand" "=v")
> + (match_operand:VSX_W 1 "register_operand" "v"))]
> + "TARGET_VSX"
> +{
> + machine_mode op_mode = GET_MODE (operands[1]);
You indent with three spaces here, instead of two.
> +
> + if (VECTOR_ELT_ORDER_BIG)
> + {
Here you do two, okay.
> + /* Big endian word numbering for words in operand is 0 1 2 3.
> + Input words 0 and 2 are where they need to be. */
> + emit_insn (gen_vsx_xvcv<VS_sxwsp>dp (operands[0], operands[1]));
> + }
> + else
> + {
> + /* Little endian word numbering for operand is 3 2 1 0.
But here you do three again.
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/powerpc/builtins-3-runnable.c
> @@ -0,0 +1,83 @@
> +/* { dg-do run { target { powerpc*-*-linux* } } } */
> +/* { dg-require-effective-target vsx_hw } */
> +/* { dg-options "-O2 -mvsx -mcpu=power8" } */
This will then also need something like
/* { dg-skip-if "do not override -mcpu" { powerpc*-*-* } { "-mcpu=*" } { "-mcpu=power8" } } */
Okay for trunk with that addition, and the final whitespace gotchas fixed.
Thanks,
Segher
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2017-06-01 22:42 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-06-01 21:55 [PATCH v3, rs6000] gcc mainline, add builtin support for vec_doublee, vec_doubleo, vec_doublel builtins Carl E. Love
2017-06-01 22:42 ` Segher Boessenkool
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).