@@ -686,4 +686,11 @@ __altivec_scalar_pred(vec_any_nle,
to #define vec_step to __builtin_vec_step. */
#define vec_step(x) __builtin_vec_step (* (__typeof__ (x) *) 0)
+#ifdef _ARCH_PWR_FUTURE
+/* May modify these macro definitions if future capabilities overload
+ with support for different vector argument and result types. */
+#define vec_clzm(a, b) __builtin_altivec_vclzdm (a, b)
+#define vec_ctzm(a, b) __builtin_altivec_vctzdm (a, b)
+#endif
+
#endif /* _ALTIVEC_H */
@@ -160,6 +160,8 @@ (define_c_enum "unspec"
UNSPEC_BCD_OVERFLOW
UNSPEC_VRLMI
UNSPEC_VRLNM
+ UNSPEC_VCLZDM
+ UNSPEC_VCTZDM
])
(define_c_enum "unspecv"
@@ -4096,6 +4098,25 @@ (define_insn "*bcd<bcd_add_sub>_test2"
"bcd<bcd_add_sub>. %0,%1,%2,%3"
[(set_attr "type" "vecsimple")])
+(define_insn "vclzdm"
+ [(set (match_operand:V2DI 0 "altivec_register_operand" "=v")
+ (unspec:V2DI [(match_operand:V2DI 1 "altivec_register_operand" "v")
+ (match_operand:V2DI 2 "altivec_register_operand" "v")]
+ UNSPEC_VCLZDM))]
+ "TARGET_FUTURE"
+ "vclzdm %0,%1,%2"
+ [(set_attr "type" "vecsimple")])
+
+(define_insn "vctzdm"
+ [(set (match_operand:V2DI 0 "altivec_register_operand" "=v")
+ (unspec:V2DI [(match_operand:V2DI 1 "altivec_register_operand" "v")
+ (match_operand:V2DI 2 "altivec_register_operand" "v")]
+ UNSPEC_VCTZDM))]
+ "TARGET_FUTURE"
+ "vctzdm %0,%1,%2"
+ [(set_attr "type" "vecsimple")])
+
+
(define_expand "bcd<bcd_add_sub>_<code>"
[(parallel [(set (reg:CCFP CR6_REGNO)
(compare:CCFP
@@ -933,6 +933,42 @@
| RS6000_BTC_BINARY), \
CODE_FOR_ ## ICODE) /* ICODE */
+/* For vector builtins for instructions which may be added at some point in
+ the future that are encoded as altivec instructions, use
+ __builtin_altivec_ as the builtin name. */
+
+#define BU_FUTURE_V_0(ENUM, NAME, ATTR, ICODE) \
+ RS6000_BUILTIN_0 (FUTUREV_BUILTIN_ ## ENUM, /* ENUM */ \
+ "__builtin_altivec_" NAME, /* NAME */ \
+ RS6000_BTM_FUTURE, /* MASK */ \
+ (RS6000_BTC_ ## ATTR /* ATTR */ \
+ | RS6000_BTC_SPECIAL), \
+ CODE_FOR_ ## ICODE) /* ICODE */
+
+#define BU_FUTURE_V_1(ENUM, NAME, ATTR, ICODE) \
+ RS6000_BUILTIN_1 (FUTURE_BUILTIN_ ## ENUM, /* ENUM */ \
+ "__builtin_altivec_" NAME, /* NAME */ \
+ RS6000_BTM_FUTURE, /* MASK */ \
+ (RS6000_BTC_ ## ATTR /* ATTR */ \
+ | RS6000_BTC_UNARY), \
+ CODE_FOR_ ## ICODE) /* ICODE */
+
+#define BU_FUTURE_V_2(ENUM, NAME, ATTR, ICODE) \
+ RS6000_BUILTIN_2 (FUTURE_BUILTIN_ ## ENUM, /* ENUM */ \
+ "__builtin_altivec_" NAME, /* NAME */ \
+ RS6000_BTM_FUTURE, /* MASK */ \
+ (RS6000_BTC_ ## ATTR /* ATTR */ \
+ | RS6000_BTC_BINARY), \
+ CODE_FOR_ ## ICODE) /* ICODE */
+
+#define BU_FUTURE_V_3(ENUM, NAME, ATTR, ICODE) \
+ RS6000_BUILTIN_3 (FUTURE_BUILTIN_ ## ENUM, /* ENUM */ \
+ "__builtin_altivec_" NAME, /* NAME */ \
+ RS6000_BTM_FUTURE, /* MASK */ \
+ (RS6000_BTC_ ## ATTR /* ATTR */ \
+ | RS6000_BTC_TERNARY), \
+ CODE_FOR_ ## ICODE) /* ICODE */
+
#endif
@@ -2479,6 +2515,10 @@ BU_P9_OVERLOAD_2 (CMPRB, "byte_in_range")
BU_P9_OVERLOAD_2 (CMPRB2, "byte_in_either_range")
BU_P9_OVERLOAD_2 (CMPEQB, "byte_in_set")
+/* Future architecture vector built-ins. */
+BU_FUTURE_V_2 (VCLZDM, "vclzdm", CONST, vclzdm)
+BU_FUTURE_V_2 (VCTZDM, "vctzdm", CONST, vctzdm)
+
/* 1 argument crypto functions. */
BU_CRYPTO_1 (VSBOX, "vsbox", CONST, crypto_vsbox_v2di)
BU_CRYPTO_1 (VSBOX_BE, "vsbox_be", CONST, crypto_vsbox_v16qi)
@@ -448,6 +448,8 @@ rs6000_target_modify_macros (bool define_p, HOST_WIDE_INT flags,
rs6000_define_or_undefine_macro (define_p, "_ARCH_PWR8");
if ((flags & OPTION_MASK_MODULO) != 0)
rs6000_define_or_undefine_macro (define_p, "_ARCH_PWR9");
+ if ((flags & OPTION_MASK_FUTURE) != 0)
+ rs6000_define_or_undefine_macro (define_p, "_ARCH_PWR_FUTURE");
if ((flags & OPTION_MASK_SOFT_FLOAT) != 0)
rs6000_define_or_undefine_macro (define_p, "_SOFT_FLOAT");
if ((flags & OPTION_MASK_RECIP_PRECISION) != 0)
@@ -12926,6 +12926,8 @@ builtin_function_type (machine_mode mode_ret, machine_mode mode_arg0,
case P8V_BUILTIN_ORC_V4SI_UNS:
case P8V_BUILTIN_ORC_V2DI_UNS:
case P8V_BUILTIN_ORC_V1TI_UNS:
+ case FUTURE_BUILTIN_VCLZDM:
+ case FUTURE_BUILTIN_VCTZDM:
h.uns_p[0] = 1;
h.uns_p[1] = 1;
h.uns_p[2] = 1;
@@ -3336,7 +3336,8 @@ rs6000_builtin_mask_calculate (void)
&& TARGET_HARD_FLOAT
&& !TARGET_IEEEQUAD) ? RS6000_BTM_LDBL128 : 0)
| ((TARGET_FLOAT128_TYPE) ? RS6000_BTM_FLOAT128 : 0)
- | ((TARGET_FLOAT128_HW) ? RS6000_BTM_FLOAT128_HW : 0));
+ | ((TARGET_FLOAT128_HW) ? RS6000_BTM_FLOAT128_HW : 0)
+ | ((TARGET_FUTURE) ? RS6000_BTM_FUTURE : 0));
}
/* Implement TARGET_MD_ASM_ADJUST. All asm statements are considered
@@ -2304,6 +2304,8 @@ extern int frame_pointer_needed;
#define RS6000_BTM_POWERPC64 MASK_POWERPC64 /* 64-bit registers. */
#define RS6000_BTM_FLOAT128 MASK_FLOAT128_KEYWORD /* IEEE 128-bit float. */
#define RS6000_BTM_FLOAT128_HW MASK_FLOAT128_HW /* IEEE 128-bit float h/w. */
+#define RS6000_BTM_FUTURE MASK_FUTURE
+
#define RS6000_BTM_COMMON (RS6000_BTM_ALTIVEC \
| RS6000_BTM_VSX \
@@ -17645,6 +17645,7 @@ briefly described below.
* PowerPC AltiVec Built-in Functions Available on ISA 2.06::
* PowerPC AltiVec Built-in Functions Available on ISA 2.07::
* PowerPC AltiVec Built-in Functions Available on ISA 3.0::
+* PowerPC AltiVec Built-in Functions Available for a Future Architecture::
@end menu
@node PowerPC AltiVec Built-in Functions on ISA 2.05
@@ -20687,6 +20688,32 @@ void vec_xst (vector char, int, char *);
void vec_xst (vector unsigned char, int, vector unsigned char *);
void vec_xst (vector unsigned char, int, unsigned char *);
@end smallexample
+
+@node PowerPC AltiVec Built-in Functions Available for a Future Architecture
+@subsubsection PowerPC AltiVec Built-in Functions Available for a Future Architecture
+
+The following additional built-in functions are also available for the
+PowerPC family of processors, starting with a hypothetical CPU
+which may or may not be available in the future
+(@option{-mcpu=future}) or later:
+
+@smallexample
+@exdent vector unsigned long long int
+@exdent vec_clzm (vector unsigned long long int, vector unsigned long long int)
+@end smallexample
+Perform a vector count leading zeros under bit mask operation, as if
+implemented by the Future @code{vclzdm} instruction.
+@findex vec_clzm
+
+@smallexample
+@exdent vector unsigned long long int
+@exdent vec_ctzm (vector unsigned long long int, vector unsigned long long int)
+@end smallexample
+Perform a vector count trailing zeros under bit mask operation, as if
+implemented by the Future @code{vctzdm} instruction.
+@findex vec_ctzm
+
+
@node PowerPC Hardware Transactional Memory Built-in Functions
@subsection PowerPC Hardware Transactional Memory Built-in Functions
GCC provides two interfaces for accessing the Hardware Transactional
new file mode 100644
@@ -0,0 +1,54 @@
+/* { dg-do compile } */
+/* { dg-options "-mdejagnu-cpu=future" } */
+
+#include <altivec.h>
+
+extern void abort (void);
+
+vector unsigned long long int
+do_vec_clzm (vector unsigned long long int source,
+ vector unsigned long long int mask)
+{
+ return vec_clzm (source, mask);
+}
+
+int main (int argc, char *argv [])
+{
+ vector unsigned long long int source_a = { 0xa5f07e3cull, 0x7e3ca5f0ull };
+ vector unsigned long long int source_b = { 0x3ca5f07eull, 0x5a0fe7c3ull };
+
+ vector unsigned long long int mask_a = { 0xffff0000ull, 0x0000ffffull };
+ vector unsigned long long int mask_b = { 0x0f0f0f0full, 0xf0f0f0f0ull };
+
+ /* See cntlzdm-0.c for derivation of expected results.
+
+ result_aa [0] is compute (source [0], mask [0];
+ result_aa [1] is compute (source [1], mask [1].
+
+ result_ab [0] is compute (source [0], mask [2];
+ result_ab [1] is compute (source [1], mask [3].
+
+ result_ba [0] is compute (source [2], mask [0];
+ result_ba [1] is compute (source [3], mask [1].
+
+ result_bb [0] is compute (source [2], mask [2];
+ result_bb [1] is compute (source [3], mask [3]. */
+
+ vector unsigned long long int result_aa = { 0, 0 };
+ vector unsigned long long int result_ab = { 1, 1 };
+ vector unsigned long long int result_ba = { 2, 0 };
+ vector unsigned long long int result_bb = { 0, 1 };
+
+ if (!vec_all_eq (do_vec_clzm (source_a, mask_a), result_aa))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_a, mask_b), result_ab))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_b, mask_a), result_ba))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_b, mask_b), result_bb))
+ abort ();
+
+ return 0;
+}
+
+/* { dg-final { scan-assembler {\mvclzdm\M} } } */
new file mode 100644
@@ -0,0 +1,54 @@
+/* { dg-do run } */
+/* { dg-require-effective-target powerpc_future_hw } */
+/* { dg-options "-mdejagnu-cpu=future" } */
+
+#include <altivec.h>
+
+extern void abort (void);
+
+vector unsigned long long int
+do_vec_clzm (vector unsigned long long int source,
+ vector unsigned long long int mask)
+{
+ return vec_clzm (source, mask);
+}
+
+int main (int argc, char *argv [])
+{
+ vector unsigned long long int source_a = { 0xa5f07e3cull, 0x7e3ca5f0ull };
+ vector unsigned long long int source_b = { 0x3ca5f07eull, 0x5a0fe7c3ull };
+
+ vector unsigned long long int mask_a = { 0xffff0000ull, 0x0000ffffull };
+ vector unsigned long long int mask_b = { 0x0f0f0f0full, 0xf0f0f0f0ull };
+
+ /* See cntlzdm-0.c for derivation of expected results.
+
+ result_aa [0] is compute (source [0], mask [0];
+ result_aa [1] is compute (source [1], mask [1].
+
+ result_ab [0] is compute (source [0], mask [2];
+ result_ab [1] is compute (source [1], mask [3].
+
+ result_ba [0] is compute (source [2], mask [0];
+ result_ba [1] is compute (source [3], mask [1].
+
+ result_bb [0] is compute (source [2], mask [2];
+ result_bb [1] is compute (source [3], mask [3]. */
+
+ vector unsigned long long int result_aa = { 0, 0 };
+ vector unsigned long long int result_ab = { 1, 1 };
+ vector unsigned long long int result_ba = { 2, 0 };
+ vector unsigned long long int result_bb = { 0, 1 };
+
+ if (!vec_all_eq (do_vec_clzm (source_a, mask_a), result_aa))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_a, mask_b), result_ab))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_b, mask_a), result_ba))
+ abort ();
+ if (!vec_all_eq (do_vec_clzm (source_b, mask_b), result_bb))
+ abort ();
+
+ return 0;
+}
+
new file mode 100644
@@ -0,0 +1,54 @@
+/* { dg-do compile } */
+/* { dg-options "-mdejagnu-cpu=future" } */
+
+#include <altivec.h>
+
+extern void abort (void);
+
+vector unsigned long long int
+do_vec_ctzm (vector unsigned long long int source,
+ vector unsigned long long int mask)
+{
+ return vec_ctzm (source, mask);
+}
+
+int main (int argc, char *argv [])
+{
+ vector unsigned long long int source_a = { 0xa5f07e3cull, 0x7e3ca5f0ull };
+ vector unsigned long long int source_b = { 0x3ca5f07eull, 0x5a0fe7c3ull };
+
+ vector unsigned long long int mask_a = { 0xffff0000ull, 0x0000ffffull };
+ vector unsigned long long int mask_b = { 0x0f0f0f0full, 0xf0f0f0f0ull };
+
+ /* See cnttzdm-0.c for derivation of expected results.
+
+ result_aa [0] is compute (source [0], mask [0];
+ result_aa [1] is compute (source [1], mask [1].
+
+ result_ab [0] is compute (source [0], mask [2];
+ result_ab [1] is compute (source [1], mask [3].
+
+ result_ba [0] is compute (source [2], mask [0];
+ result_ba [1] is compute (source [3], mask [1].
+
+ result_bb [0] is compute (source [2], mask [2];
+ result_bb [1] is compute (source [3], mask [3]. */
+
+ vector unsigned long long int result_aa = { 4, 4 };
+ vector unsigned long long int result_ab = { 2, 0 };
+ vector unsigned long long int result_ba = { 0, 0 };
+ vector unsigned long long int result_bb = { 1, 2 };
+
+ if (!vec_all_eq (do_vec_ctzm (source_a, mask_a), result_aa))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_a, mask_b), result_ab))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_b, mask_a),result_ba))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_b, mask_b), result_bb))
+ abort ();
+
+ return 0;
+}
+
+/* { dg-final { scan-assembler {\mvctzdm\M} } } */
new file mode 100644
@@ -0,0 +1,53 @@
+/* { dg-do compile } */
+/* { dg-require-effective-target powerpc_future_hw } */
+/* { dg-options "-mdejagnu-cpu=future" } */
+
+#include <altivec.h>
+
+extern void abort (void);
+
+vector unsigned long long int
+do_vec_ctzm (vector unsigned long long int source,
+ vector unsigned long long int mask)
+{
+ return vec_ctzm (source, mask);
+}
+
+int main (int argc, char *argv [])
+{
+ vector unsigned long long int source_a = { 0xa5f07e3cull, 0x7e3ca5f0ull };
+ vector unsigned long long int source_b = { 0x3ca5f07eull, 0x5a0fe7c3ull };
+
+ vector unsigned long long int mask_a = { 0xffff0000ull, 0x0000ffffull };
+ vector unsigned long long int mask_b = { 0x0f0f0f0full, 0xf0f0f0f0ull };
+
+ /* See cnttzdm-0.c for derivation of expected results.
+
+ result_aa [0] is compute (source [0], mask [0];
+ result_aa [1] is compute (source [1], mask [1].
+
+ result_ab [0] is compute (source [0], mask [2];
+ result_ab [1] is compute (source [1], mask [3].
+
+ result_ba [0] is compute (source [2], mask [0];
+ result_ba [1] is compute (source [3], mask [1].
+
+ result_bb [0] is compute (source [2], mask [2];
+ result_bb [1] is compute (source [3], mask [3]. */
+
+ vector unsigned long long int result_aa = { 4, 4 };
+ vector unsigned long long int result_ab = { 2, 0 };
+ vector unsigned long long int result_ba = { 0, 0 };
+ vector unsigned long long int result_bb = { 1, 2 };
+
+ if (!vec_all_eq (do_vec_ctzm (source_a, mask_a), result_aa))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_a, mask_b), result_ab))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_b, mask_a),result_ba))
+ abort ();
+ if (!vec_all_eq (do_vec_ctzm (source_b, mask_b), result_bb))
+ abort ();
+
+ return 0;
+}
From: Kelvin Nilsen <kelvin@gcc.gnu.org> Add support for new vclzdm and vctzdm vector instructions that count leading and trailing zeros under control of a mask. Bootstrapped and tested on powerpc64le-unknown-linux-gnu with no regressions. Is this okay for master? Thanks, Bill [gcc] 2020-05-07 Kelvin Nilsen <kelvin@gcc.gnu.org> Bill Schmidt <wschmidt@linux.ibm.com> * config/rs6000/altivec.h (vec_clzm): New macro. (vec_ctzm): Likewise. * config/rs6000/altivec.md (UNSPEC_VCLZDM): New constant. (UNSPEC_VCTZDM): Likewise. (vclzdm): New insn. (vctzdm): Likewise. * config/rs6000/rs6000-builtin.def (BU_FUTURE_V_0): New macro. (BU_FUTURE_V_1): Likewise. (BU_FUTURE_V_2): Likewise. (BU_FUTURE_V_3): Likewise. (__builtin_altivec_vclzdm): New builtin definition. (__builtin_altivec_vctzdm): Likewise. * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Cause _ARCH_PWR_FUTURE macro to be defined if OPTION_MASK_FUTURE flag is set. * config/rs6000/rs6000-call.c (builtin_function_type): Set return value and parameter types to be unsigned for VCLZDM and VCTZDM. * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add support for TARGET_FUTURE flag. * config/rs6000/rs6000.h (RS6000_BTM_FUTURE): New macro constant. * doc/extend.texi (PowerPC Altivec Built-in Functions Available for a Future Architecture): New subsubsection. [gcc/testsuite] 2020-05-07 Kelvin Nilsen <kelvin@gcc.gnu.org> * gcc.target/powerpc/vec-clzm-0.c: New test. * gcc.target/powerpc/vec-clzm-1.c: New test. * gcc.target/powerpc/vec-ctzm-0.c: New test. * gcc.target/powerpc/vec-ctzm-1.c: New test. --- gcc/config/rs6000/altivec.h | 7 +++ gcc/config/rs6000/altivec.md | 21 ++++++++ gcc/config/rs6000/rs6000-builtin.def | 40 ++++++++++++++ gcc/config/rs6000/rs6000-c.c | 2 + gcc/config/rs6000/rs6000-call.c | 2 + gcc/config/rs6000/rs6000.c | 3 +- gcc/config/rs6000/rs6000.h | 2 + gcc/doc/extend.texi | 27 ++++++++++ gcc/testsuite/gcc.target/powerpc/vec-clzm-0.c | 54 +++++++++++++++++++ gcc/testsuite/gcc.target/powerpc/vec-clzm-1.c | 54 +++++++++++++++++++ gcc/testsuite/gcc.target/powerpc/vec-ctzm-0.c | 54 +++++++++++++++++++ gcc/testsuite/gcc.target/powerpc/vec-ctzm-1.c | 53 ++++++++++++++++++ 12 files changed, 318 insertions(+), 1 deletion(-) create mode 100644 gcc/testsuite/gcc.target/powerpc/vec-clzm-0.c create mode 100644 gcc/testsuite/gcc.target/powerpc/vec-clzm-1.c create mode 100644 gcc/testsuite/gcc.target/powerpc/vec-ctzm-0.c create mode 100644 gcc/testsuite/gcc.target/powerpc/vec-ctzm-1.c