@@ -373,6 +373,24 @@ arm_binop_none_unone_unone_qualifiers[SIMD_MAX_BUILTIN_ARGS]
#define BINOP_NONE_UNONE_UNONE_QUALIFIERS \
(arm_binop_none_unone_unone_qualifiers)
+static enum arm_type_qualifiers
+arm_binop_unone_unone_imm_qualifiers[SIMD_MAX_BUILTIN_ARGS]
+ = { qualifier_unsigned, qualifier_unsigned, qualifier_immediate };
+#define BINOP_UNONE_UNONE_IMM_QUALIFIERS \
+ (arm_binop_unone_unone_imm_qualifiers)
+
+static enum arm_type_qualifiers
+arm_binop_unone_unone_unone_qualifiers[SIMD_MAX_BUILTIN_ARGS]
+ = { qualifier_unsigned, qualifier_unsigned, qualifier_unsigned };
+#define BINOP_UNONE_UNONE_UNONE_QUALIFIERS \
+ (arm_binop_unone_unone_unone_qualifiers)
+
+static enum arm_type_qualifiers
+arm_binop_unone_none_imm_qualifiers[SIMD_MAX_BUILTIN_ARGS]
+ = { qualifier_unsigned, qualifier_none, qualifier_immediate };
+#define BINOP_UNONE_NONE_IMM_QUALIFIERS \
+ (arm_binop_unone_none_imm_qualifiers)
+
/* End of Qualifier for MVE builtins. */
/* void ([T element type] *, T, immediate). */
@@ -207,6 +207,24 @@ typedef struct { uint8x16_t val[4]; } uint8x16x4_t;
#define vcvtq_n_f32_u32(__a, __imm6) __arm_vcvtq_n_f32_u32(__a, __imm6)
#define vcreateq_f16(__a, __b) __arm_vcreateq_f16(__a, __b)
#define vcreateq_f32(__a, __b) __arm_vcreateq_f32(__a, __b)
+#define vcvtq_n_s16_f16(__a, __imm6) __arm_vcvtq_n_s16_f16(__a, __imm6)
+#define vcvtq_n_s32_f32(__a, __imm6) __arm_vcvtq_n_s32_f32(__a, __imm6)
+#define vcvtq_n_u16_f16(__a, __imm6) __arm_vcvtq_n_u16_f16(__a, __imm6)
+#define vcvtq_n_u32_f32(__a, __imm6) __arm_vcvtq_n_u32_f32(__a, __imm6)
+#define vcreateq_u8(__a, __b) __arm_vcreateq_u8(__a, __b)
+#define vcreateq_u16(__a, __b) __arm_vcreateq_u16(__a, __b)
+#define vcreateq_u32(__a, __b) __arm_vcreateq_u32(__a, __b)
+#define vcreateq_u64(__a, __b) __arm_vcreateq_u64(__a, __b)
+#define vcreateq_s8(__a, __b) __arm_vcreateq_s8(__a, __b)
+#define vcreateq_s16(__a, __b) __arm_vcreateq_s16(__a, __b)
+#define vcreateq_s32(__a, __b) __arm_vcreateq_s32(__a, __b)
+#define vcreateq_s64(__a, __b) __arm_vcreateq_s64(__a, __b)
+#define vshrq_n_s8(__a, __imm) __arm_vshrq_n_s8(__a, __imm)
+#define vshrq_n_s16(__a, __imm) __arm_vshrq_n_s16(__a, __imm)
+#define vshrq_n_s32(__a, __imm) __arm_vshrq_n_s32(__a, __imm)
+#define vshrq_n_u8(__a, __imm) __arm_vshrq_n_u8(__a, __imm)
+#define vshrq_n_u16(__a, __imm) __arm_vshrq_n_u16(__a, __imm)
+#define vshrq_n_u32(__a, __imm) __arm_vshrq_n_u32(__a, __imm)
#endif
__extension__ extern __inline void
@@ -753,6 +771,104 @@ __arm_vpnot (mve_pred16_t __a)
return __builtin_mve_vpnothi (__a);
}
+__extension__ extern __inline uint8x16_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_u8 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_uv16qi (__a, __b);
+}
+
+__extension__ extern __inline uint16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_u16 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_uv8hi (__a, __b);
+}
+
+__extension__ extern __inline uint32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_u32 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_uv4si (__a, __b);
+}
+
+__extension__ extern __inline uint64x2_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_u64 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_uv2di (__a, __b);
+}
+
+__extension__ extern __inline int8x16_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_s8 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_sv16qi (__a, __b);
+}
+
+__extension__ extern __inline int16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_s16 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_sv8hi (__a, __b);
+}
+
+__extension__ extern __inline int32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_s32 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_sv4si (__a, __b);
+}
+
+__extension__ extern __inline int64x2_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcreateq_s64 (uint64_t __a, uint64_t __b)
+{
+ return __builtin_mve_vcreateq_sv2di (__a, __b);
+}
+
+__extension__ extern __inline int8x16_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_s8 (int8x16_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_sv16qi (__a, __imm);
+}
+
+__extension__ extern __inline int16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_s16 (int16x8_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_sv8hi (__a, __imm);
+}
+
+__extension__ extern __inline int32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_s32 (int32x4_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_sv4si (__a, __imm);
+}
+
+__extension__ extern __inline uint8x16_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_u8 (uint8x16_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_uv16qi (__a, __imm);
+}
+
+__extension__ extern __inline uint16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_u16 (uint16x8_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_uv8hi (__a, __imm);
+}
+
+__extension__ extern __inline uint32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vshrq_n_u32 (uint32x4_t __a, const int __imm)
+{
+ return __builtin_mve_vshrq_n_uv4si (__a, __imm);
+}
+
#if (__ARM_FEATURE_MVE & 2) /* MVE Floating point. */
__extension__ extern __inline void
@@ -1165,6 +1281,34 @@ __arm_vcreateq_f32 (uint64_t __a, uint64_t __b)
return __builtin_mve_vcreateq_fv4sf (__a, __b);
}
+__extension__ extern __inline int16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcvtq_n_s16_f16 (float16x8_t __a, const int __imm6)
+{
+ return __builtin_mve_vcvtq_n_from_f_sv8hi (__a, __imm6);
+}
+
+__extension__ extern __inline int32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcvtq_n_s32_f32 (float32x4_t __a, const int __imm6)
+{
+ return __builtin_mve_vcvtq_n_from_f_sv4si (__a, __imm6);
+}
+
+__extension__ extern __inline uint16x8_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcvtq_n_u16_f16 (float16x8_t __a, const int __imm6)
+{
+ return __builtin_mve_vcvtq_n_from_f_uv8hi (__a, __imm6);
+}
+
+__extension__ extern __inline uint32x4_t
+__attribute__ ((__always_inline__, __gnu_inline__, __artificial__))
+__arm_vcvtq_n_u32_f32 (float32x4_t __a, const int __imm6)
+{
+ return __builtin_mve_vcvtq_n_from_f_uv4si (__a, __imm6);
+}
+
#endif
enum {
@@ -1598,6 +1742,16 @@ extern void *__ARM_undef;
int (*)[__ARM_mve_type_int16x8_t]: __arm_vqnegq_s16 (__ARM_mve_coerce(__p0, int16x8_t)), \
int (*)[__ARM_mve_type_int32x4_t]: __arm_vqnegq_s32 (__ARM_mve_coerce(__p0, int32x4_t)));})
+#define vshrq(p0,p1) __arm_vshrq(p0,p1)
+#define __arm_vshrq(p0,p1) ({ __typeof(p0) __p0 = (p0); \
+ _Generic( (int (*)[__ARM_mve_typeid(__p0)])0, \
+ int (*)[__ARM_mve_type_int8x16_t]: __arm_vshrq_n_s8 (__ARM_mve_coerce(__p0, int8x16_t), p1), \
+ int (*)[__ARM_mve_type_int16x8_t]: __arm_vshrq_n_s16 (__ARM_mve_coerce(__p0, int16x8_t), p1), \
+ int (*)[__ARM_mve_type_int32x4_t]: __arm_vshrq_n_s32 (__ARM_mve_coerce(__p0, int32x4_t), p1), \
+ int (*)[__ARM_mve_type_uint8x16_t]: __arm_vshrq_n_u8 (__ARM_mve_coerce(__p0, uint8x16_t), p1), \
+ int (*)[__ARM_mve_type_uint16x8_t]: __arm_vshrq_n_u16 (__ARM_mve_coerce(__p0, uint16x8_t), p1), \
+ int (*)[__ARM_mve_type_uint32x4_t]: __arm_vshrq_n_u32 (__ARM_mve_coerce(__p0, uint32x4_t), p1));})
+
#endif /* MVE Floating point. */
#ifdef __cplusplus
@@ -6,7 +6,7 @@
GCC is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published
- by the Free Software Foundation; either version 3, or (at your
+ by the Free Software Foundation; either version 3, or (at your
option) any later version.
GCC is distributed in the hope that it will be useful, but WITHOUT
@@ -81,3 +81,9 @@ VAR2 (BINOP_NONE_NONE_NONE, vbrsrq_n_f, v8hf, v4sf)
VAR2 (BINOP_NONE_NONE_IMM, vcvtq_n_to_f_s, v8hf, v4sf)
VAR2 (BINOP_NONE_UNONE_IMM, vcvtq_n_to_f_u, v8hf, v4sf)
VAR2 (BINOP_NONE_UNONE_UNONE, vcreateq_f, v8hf, v4sf)
+VAR2 (BINOP_UNONE_NONE_IMM, vcvtq_n_from_f_u, v8hi, v4si)
+VAR2 (BINOP_NONE_NONE_IMM, vcvtq_n_from_f_s, v8hi, v4si)
+VAR4 (BINOP_UNONE_UNONE_UNONE, vcreateq_u, v16qi, v8hi, v4si, v2di)
+VAR4 (BINOP_NONE_UNONE_UNONE, vcreateq_s, v16qi, v8hi, v4si, v2di)
+VAR3 (BINOP_UNONE_UNONE_IMM, vshrq_n_u, v16qi, v8hi, v4si)
+VAR3 (BINOP_NONE_NONE_IMM, vshrq_n_s, v16qi, v8hi, v4si)
@@ -35,7 +35,7 @@
;; in ARM/Thumb-2 state: Da, Db, Dc, Dd, Dn, DN, Dm, Dl, DL, Do, Dv, Dy, Di,
;; Dt, Dp, Dz, Tu
;; in Thumb-1 state: Pa, Pb, Pc, Pd, Pe
-;; in Thumb-2 state: Ha, Pj, PJ, Ps, Pt, Pu, Pv, Pw, Px, Py, Pz, Rd
+;; in Thumb-2 state: Ha, Pj, PJ, Ps, Pt, Pu, Pv, Pw, Px, Py, Pz, Rd, Rf, Rb
;; in all states: Pf, Pg, UM, U1
;; The following memory constraints have been used:
@@ -56,6 +56,16 @@
(and (match_code "const_int")
(match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 16")))
+(define_constraint "Rb"
+ "@internal In Thumb-2 state a constant in range 1 to 8"
+ (and (match_code "const_int")
+ (match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 8")))
+
+(define_constraint "Rf"
+ "@internal In Thumb-2 state a constant in range 1 to 32"
+ (and (match_code "const_int")
+ (match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 32")))
+
(define_register_constraint "t" "TARGET_32BIT ? VFP_LO_REGS : NO_REGS"
"The VFP registers @code{s0}-@code{s31}.")
@@ -22,6 +22,7 @@
(define_mode_iterator MVE_types [V16QI V8HI V4SI V2DI TI V8HF V4SF V2DF])
(define_mode_iterator MVE_VLD_ST [V16QI V8HI V4SI V8HF V4SF])
(define_mode_iterator MVE_0 [V8HF V4SF])
+(define_mode_iterator MVE_1 [V16QI V8HI V4SI V2DI])
(define_mode_iterator MVE_3 [V16QI V8HI])
(define_mode_iterator MVE_2 [V16QI V8HI V4SI])
(define_mode_iterator MVE_5 [V8HI V4SI])
@@ -38,7 +39,8 @@
VCVTPQ_U VCVTNQ_S VCVTNQ_U VCVTMQ_S VCVTMQ_U
VADDLVQ_U VCTP8Q VCTP16Q VCTP32Q VCTP64Q VPNOT
VCREATEQ_F VCVTQ_N_TO_F_S VCVTQ_N_TO_F_U VBRSRQ_N_F
- VSUBQ_N_F])
+ VSUBQ_N_F VCREATEQ_U VCREATEQ_S VSHRQ_N_S VSHRQ_N_U
+ VCVTQ_N_FROM_F_S VCVTQ_N_FROM_F_U])
(define_mode_attr MVE_CNVT [(V8HI "V8HF") (V4SI "V4SF")
(V8HF "V8HI") (V4SF "V4SI")])
@@ -55,10 +57,16 @@
(VCVTNQ_U "u") (VCVTMQ_S "s") (VCVTMQ_U "u")
(VCLZQ_U "u") (VCLZQ_S "s") (VREV32Q_U "u")
(VREV32Q_S "s") (VADDLVQ_U "u") (VADDLVQ_S "s")
- (VCVTQ_N_TO_F_S "s") (VCVTQ_N_TO_F_U "u")])
+ (VCVTQ_N_TO_F_S "s") (VCVTQ_N_TO_F_U "u")
+ (VCREATEQ_U "u") (VCREATEQ_S "s") (VSHRQ_N_S "s")
+ (VSHRQ_N_U "u") (VCVTQ_N_FROM_F_S "s")
+ (VCVTQ_N_FROM_F_U "u")])
(define_int_attr mode1 [(VCTP8Q "8") (VCTP16Q "16") (VCTP32Q "32")
(VCTP64Q "64")])
+(define_mode_attr MVE_pred2 [(V16QI "mve_imm_8") (V8HI "mve_imm_16")
+ (V4SI "mve_imm_32")])
+(define_mode_attr MVE_constraint2 [(V16QI "Rb") (V8HI "Rd") (V4SI "Rf")])
(define_int_iterator VCVTQ_TO_F [VCVTQ_TO_F_S VCVTQ_TO_F_U])
(define_int_iterator VMVNQ_N [VMVNQ_N_U VMVNQ_N_S])
@@ -79,6 +87,9 @@
(define_int_iterator VADDLVQ [VADDLVQ_U VADDLVQ_S])
(define_int_iterator VCTPQ [VCTP8Q VCTP16Q VCTP32Q VCTP64Q])
(define_int_iterator VCVTQ_N_TO_F [VCVTQ_N_TO_F_S VCVTQ_N_TO_F_U])
+(define_int_iterator VCREATEQ [VCREATEQ_U VCREATEQ_S])
+(define_int_iterator VSHRQ_N [VSHRQ_N_S VSHRQ_N_U])
+(define_int_iterator VCVTQ_N_FROM_F [VCVTQ_N_FROM_F_S VCVTQ_N_FROM_F_U])
(define_insn "*mve_mov<mode>"
[(set (match_operand:MVE_types 0 "s_register_operand" "=w,w,r,w,w,r,w")
@@ -743,3 +754,48 @@
"vmov %q0[2], %q0[0], %Q2, %Q1\;vmov %q0[3], %q0[1], %R2, %R1"
[(set_attr "type" "mve_move")
(set_attr "length""8")])
+
+;;
+;; [vcreateq_u, vcreateq_s])
+;;
+(define_insn "mve_vcreateq_<supf><mode>"
+ [
+ (set (match_operand:MVE_1 0 "s_register_operand" "=w")
+ (unspec:MVE_1 [(match_operand:DI 1 "s_register_operand" "r")
+ (match_operand:DI 2 "s_register_operand" "r")]
+ VCREATEQ))
+ ]
+ "TARGET_HAVE_MVE"
+ "vmov %q0[2], %q0[0], %Q2, %Q1\;vmov %q0[3], %q0[1], %R2, %R1"
+ [(set_attr "type" "mve_move")
+ (set_attr "length""8")])
+
+;;
+;; [vshrq_n_s, vshrq_n_u])
+;;
+(define_insn "mve_vshrq_n_<supf><mode>"
+ [
+ (set (match_operand:MVE_2 0 "s_register_operand" "=w")
+ (unspec:MVE_2 [(match_operand:MVE_2 1 "s_register_operand" "w")
+ (match_operand:SI 2 "<MVE_pred2>" "<MVE_constraint2>")]
+ VSHRQ_N))
+ ]
+ "TARGET_HAVE_MVE"
+ "vshr.<supf>%#<V_sz_elem>\t%q0, %q1, %2"
+ [(set_attr "type" "mve_move")
+])
+
+;;
+;; [vcvtq_n_from_f_s, vcvtq_n_from_f_u])
+;;
+(define_insn "mve_vcvtq_n_from_f_<supf><mode>"
+ [
+ (set (match_operand:MVE_5 0 "s_register_operand" "=w")
+ (unspec:MVE_5 [(match_operand:<MVE_CNVT> 1 "s_register_operand" "w")
+ (match_operand:SI 2 "mve_imm_16" "Rd")]
+ VCVTQ_N_FROM_F))
+ ]
+ "TARGET_HAVE_MVE && TARGET_HAVE_MVE_FLOAT"
+ "vcvt.<supf>%#<V_sz_elem>.f%#<V_sz_elem>\t%q0, %q1, %2"
+ [(set_attr "type" "mve_move")
+])
@@ -35,6 +35,14 @@
(define_predicate "mve_imm_16"
(match_test "satisfies_constraint_Rd (op)"))
+;; True for immediates in the range of 1 to 8 for MVE.
+(define_predicate "mve_imm_8"
+ (match_test "satisfies_constraint_Rb (op)"))
+
+;; True for immediates in the range of 1 to 32 for MVE.
+(define_predicate "mve_imm_32"
+ (match_test "satisfies_constraint_Rf (op)"))
+
; Predicate for stack protector guard's address in
; stack_protect_combined_set_insn and stack_protect_combined_test_insn patterns
(define_predicate "guard_addr_operand"
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int16x8_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_s16 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int32x4_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_s32 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int64x2_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_s64 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int8x16_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_s8 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint16x8_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_u16 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint32x4_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_u32 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint64x2_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_u64 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint8x16_t
+foo (uint64_t a, uint64_t b)
+{
+ return vcreateq_u8 (a, b);
+}
+
+/* { dg-final { scan-assembler "vmov" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int16x8_t
+foo (float16x8_t a)
+{
+ return vcvtq_n_s16_f16 (a, 1);
+}
+
+/* { dg-final { scan-assembler "vcvt.s16.f16" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int32x4_t
+foo (float32x4_t a)
+{
+ return vcvtq_n_s32_f32 (a, 1);
+}
+
+/* { dg-final { scan-assembler "vcvt.s32.f32" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint16x8_t
+foo (float16x8_t a)
+{
+ return vcvtq_n_u16_f16 (a, 1);
+}
+
+/* { dg-final { scan-assembler "vcvt.u16.f16" } } */
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint32x4_t
+foo (float32x4_t a)
+{
+ return vcvtq_n_u32_f32 (a, 1);
+}
+
+/* { dg-final { scan-assembler "vcvt.u32.f32" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int16x8_t
+foo (int16x8_t a)
+{
+ return vshrq_n_s16 (a, 16);
+}
+
+/* { dg-final { scan-assembler "vshr.s16" } } */
+
+int16x8_t
+foo1 (int16x8_t a)
+{
+ return vshrq (a, 16);
+}
+
+/* { dg-final { scan-assembler "vshr.s16" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int32x4_t
+foo (int32x4_t a)
+{
+ return vshrq_n_s32 (a, 32);
+}
+
+/* { dg-final { scan-assembler "vshr.s32" } } */
+
+int32x4_t
+foo1 (int32x4_t a)
+{
+ return vshrq (a, 32);
+}
+
+/* { dg-final { scan-assembler "vshr.s32" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+int8x16_t
+foo (int8x16_t a)
+{
+ return vshrq_n_s8 (a, 8);
+}
+
+/* { dg-final { scan-assembler "vshr.s8" } } */
+
+int8x16_t
+foo1 (int8x16_t a)
+{
+ return vshrq (a, 8);
+}
+
+/* { dg-final { scan-assembler "vshr.s8" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint16x8_t
+foo (uint16x8_t a)
+{
+ return vshrq_n_u16 (a, 16);
+}
+
+/* { dg-final { scan-assembler "vshr.u16" } } */
+
+uint16x8_t
+foo1 (uint16x8_t a)
+{
+ return vshrq (a, 16);
+}
+
+/* { dg-final { scan-assembler "vshr.u16" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint32x4_t
+foo (uint32x4_t a)
+{
+ return vshrq_n_u32 (a, 32);
+}
+
+/* { dg-final { scan-assembler "vshr.u32" } } */
+
+uint32x4_t
+foo1 (uint32x4_t a)
+{
+ return vshrq (a, 32);
+}
+
+/* { dg-final { scan-assembler "vshr.u32" } } */
new file mode 100644
@@ -0,0 +1,21 @@
+/* { dg-do compile } */
+/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */
+/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */
+
+#include "arm_mve.h"
+
+uint8x16_t
+foo (uint8x16_t a)
+{
+ return vshrq_n_u8 (a, 8);
+}
+
+/* { dg-final { scan-assembler "vshr.u8" } } */
+
+uint8x16_t
+foo1 (uint8x16_t a)
+{
+ return vshrq (a, 8);
+}
+
+/* { dg-final { scan-assembler "vshr.u8" } } */
Hello, This patch supports following MVE ACLE intrinsics with binary operands. vcvtq_n_s16_f16, vcvtq_n_s32_f32, vcvtq_n_u16_f16, vcvtq_n_u32_f32, vcreateq_u8, vcreateq_u16, vcreateq_u32, vcreateq_u64, vcreateq_s8, vcreateq_s16, vcreateq_s32, vcreateq_s64, vshrq_n_s8, vshrq_n_s16, vshrq_n_s32, vshrq_n_u8, vshrq_n_u16, vshrq_n_u32. Please refer to M-profile Vector Extension (MVE) intrinsics [1] for more details. [1] https://developer.arm.com/architectures/instruction-sets/simd-isas/helium/mve-intrinsics In this patch new constraints "Rb" and "Rf" are added, which checks the constant is with in the range of 1 to 8 and 1 to 32 respectively. Also a new predicates "mve_imm_8" and "mve_imm_32" are added, to check the the matching constraint Rb and Rf respectively. Regression tested on arm-none-eabi and found no regressions. Ok for trunk? Thanks, Srinath. gcc/ChangeLog: 2019-10-21 Andre Vieira <andre.simoesdiasvieira@arm.com> Mihail Ionescu <mihail.ionescu@arm.com> Srinath Parvathaneni <srinath.parvathaneni@arm.com> * config/arm/arm-builtins.c (BINOP_UNONE_UNONE_IMM_QUALIFIERS): Define qualifier for binary operands. (BINOP_UNONE_UNONE_UNONE_QUALIFIERS): Likewise. (BINOP_UNONE_NONE_IMM_QUALIFIERS): Likewise. * config/arm/arm_mve.h (vcvtq_n_s16_f16): Define macro. (vcvtq_n_s32_f32): Likewise. (vcvtq_n_u16_f16): Likewise. (vcvtq_n_u32_f32): Likewise. (vcreateq_u8): Likewise. (vcreateq_u16): Likewise. (vcreateq_u32): Likewise. (vcreateq_u64): Likewise. (vcreateq_s8): Likewise. (vcreateq_s16): Likewise. (vcreateq_s32): Likewise. (vcreateq_s64): Likewise. (vshrq_n_s8): Likewise. (vshrq_n_s16): Likewise. (vshrq_n_s32): Likewise. (vshrq_n_u8): Likewise. (vshrq_n_u16): Likewise. (vshrq_n_u32): Likewise. (__arm_vcreateq_u8): Define intrinsic. (__arm_vcreateq_u16): Likewise. (__arm_vcreateq_u32): Likewise. (__arm_vcreateq_u64): Likewise. (__arm_vcreateq_s8): Likewise. (__arm_vcreateq_s16): Likewise. (__arm_vcreateq_s32): Likewise. (__arm_vcreateq_s64): Likewise. (__arm_vshrq_n_s8): Likewise. (__arm_vshrq_n_s16): Likewise. (__arm_vshrq_n_s32): Likewise. (__arm_vshrq_n_u8): Likewise. (__arm_vshrq_n_u16): Likewise. (__arm_vshrq_n_u32): Likewise. (__arm_vcvtq_n_s16_f16): Likewise. (__arm_vcvtq_n_s32_f32): Likewise. (__arm_vcvtq_n_u16_f16): Likewise. (__arm_vcvtq_n_u32_f32): Likewise. (vshrq_n): Define polymorphic variant. * config/arm/arm_mve_builtins.def (BINOP_UNONE_UNONE_IMM_QUALIFIERS): Use it. (BINOP_UNONE_UNONE_UNONE_QUALIFIERS): Likewise. (BINOP_UNONE_NONE_IMM_QUALIFIERS): Likewise. * config/arm/constraints.md (Rb): Define constraint to check constant is in the range of 1 to 8. (Rf): Define constraint to check constant is in the range of 1 to 32. * config/arm/mve.md (mve_vcreateq_<supf><mode>): Define RTL pattern. (mve_vshrq_n_<supf><mode>): Likewise. (mve_vcvtq_n_from_f_<supf><mode>): Likewise. * config/arm/predicates.md (mve_imm_8): Define predicate to check the matching constraint Rb. (mve_imm_32): Define predicate to check the matching constraint Rf. gcc/testsuite/ChangeLog: 2019-10-21 Andre Vieira <andre.simoesdiasvieira@arm.com> Mihail Ionescu <mihail.ionescu@arm.com> Srinath Parvathaneni <srinath.parvathaneni@arm.com> * gcc.target/arm/mve/intrinsics/vcreateq_s16.c: New test. * gcc.target/arm/mve/intrinsics/vcreateq_s32.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_s64.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_s8.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_u16.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_u32.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_u64.c: Likewise. * gcc.target/arm/mve/intrinsics/vcreateq_u8.c: Likewise. * gcc.target/arm/mve/intrinsics/vcvtq_n_s16_f16.c: Likewise. * gcc.target/arm/mve/intrinsics/vcvtq_n_s32_f32.c: Likewise. * gcc.target/arm/mve/intrinsics/vcvtq_n_u16_f16.c: Likewise. * gcc.target/arm/mve/intrinsics/vcvtq_n_u32_f32.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_s16.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_s32.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_s8.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_u16.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_u32.c: Likewise. * gcc.target/arm/mve/intrinsics/vshrq_n_u8.c: Likewise. ############### Attachment also inlined for ease of reply ############### diff --git a/gcc/config/arm/arm-builtins.c b/gcc/config/arm/arm-builtins.c index c2dad057d1365914477c64d559aa1fd1c32bbf19..ec56dbcdd66662bb1de696142186955d75570772 100644 --- a/gcc/config/arm/arm-builtins.c +++ b/gcc/config/arm/arm-builtins.c @@ -373,6 +373,24 @@ arm_binop_none_unone_unone_qualifiers[SIMD_MAX_BUILTIN_ARGS] #define BINOP_NONE_UNONE_UNONE_QUALIFIERS \ (arm_binop_none_unone_unone_qualifiers) +static enum arm_type_qualifiers +arm_binop_unone_unone_imm_qualifiers[SIMD_MAX_BUILTIN_ARGS] + = { qualifier_unsigned, qualifier_unsigned, qualifier_immediate }; +#define BINOP_UNONE_UNONE_IMM_QUALIFIERS \ + (arm_binop_unone_unone_imm_qualifiers) + +static enum arm_type_qualifiers +arm_binop_unone_unone_unone_qualifiers[SIMD_MAX_BUILTIN_ARGS] + = { qualifier_unsigned, qualifier_unsigned, qualifier_unsigned }; +#define BINOP_UNONE_UNONE_UNONE_QUALIFIERS \ + (arm_binop_unone_unone_unone_qualifiers) + +static enum arm_type_qualifiers +arm_binop_unone_none_imm_qualifiers[SIMD_MAX_BUILTIN_ARGS] + = { qualifier_unsigned, qualifier_none, qualifier_immediate }; +#define BINOP_UNONE_NONE_IMM_QUALIFIERS \ + (arm_binop_unone_none_imm_qualifiers) + /* End of Qualifier for MVE builtins. */ /* void ([T element type] *, T, immediate). */ diff --git a/gcc/config/arm/arm_mve.h b/gcc/config/arm/arm_mve.h index 15b7ada025fe57b682f3873f0f275c43a30d1273..0318a2cd8c3cdea430a32506e75a5aceb4d3a475 100644 --- a/gcc/config/arm/arm_mve.h +++ b/gcc/config/arm/arm_mve.h @@ -207,6 +207,24 @@ typedef struct { uint8x16_t val[4]; } uint8x16x4_t; #define vcvtq_n_f32_u32(__a, __imm6) __arm_vcvtq_n_f32_u32(__a, __imm6) #define vcreateq_f16(__a, __b) __arm_vcreateq_f16(__a, __b) #define vcreateq_f32(__a, __b) __arm_vcreateq_f32(__a, __b) +#define vcvtq_n_s16_f16(__a, __imm6) __arm_vcvtq_n_s16_f16(__a, __imm6) +#define vcvtq_n_s32_f32(__a, __imm6) __arm_vcvtq_n_s32_f32(__a, __imm6) +#define vcvtq_n_u16_f16(__a, __imm6) __arm_vcvtq_n_u16_f16(__a, __imm6) +#define vcvtq_n_u32_f32(__a, __imm6) __arm_vcvtq_n_u32_f32(__a, __imm6) +#define vcreateq_u8(__a, __b) __arm_vcreateq_u8(__a, __b) +#define vcreateq_u16(__a, __b) __arm_vcreateq_u16(__a, __b) +#define vcreateq_u32(__a, __b) __arm_vcreateq_u32(__a, __b) +#define vcreateq_u64(__a, __b) __arm_vcreateq_u64(__a, __b) +#define vcreateq_s8(__a, __b) __arm_vcreateq_s8(__a, __b) +#define vcreateq_s16(__a, __b) __arm_vcreateq_s16(__a, __b) +#define vcreateq_s32(__a, __b) __arm_vcreateq_s32(__a, __b) +#define vcreateq_s64(__a, __b) __arm_vcreateq_s64(__a, __b) +#define vshrq_n_s8(__a, __imm) __arm_vshrq_n_s8(__a, __imm) +#define vshrq_n_s16(__a, __imm) __arm_vshrq_n_s16(__a, __imm) +#define vshrq_n_s32(__a, __imm) __arm_vshrq_n_s32(__a, __imm) +#define vshrq_n_u8(__a, __imm) __arm_vshrq_n_u8(__a, __imm) +#define vshrq_n_u16(__a, __imm) __arm_vshrq_n_u16(__a, __imm) +#define vshrq_n_u32(__a, __imm) __arm_vshrq_n_u32(__a, __imm) #endif __extension__ extern __inline void @@ -753,6 +771,104 @@ __arm_vpnot (mve_pred16_t __a) return __builtin_mve_vpnothi (__a); } +__extension__ extern __inline uint8x16_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_u8 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_uv16qi (__a, __b); +} + +__extension__ extern __inline uint16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_u16 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_uv8hi (__a, __b); +} + +__extension__ extern __inline uint32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_u32 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_uv4si (__a, __b); +} + +__extension__ extern __inline uint64x2_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_u64 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_uv2di (__a, __b); +} + +__extension__ extern __inline int8x16_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_s8 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_sv16qi (__a, __b); +} + +__extension__ extern __inline int16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_s16 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_sv8hi (__a, __b); +} + +__extension__ extern __inline int32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_s32 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_sv4si (__a, __b); +} + +__extension__ extern __inline int64x2_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcreateq_s64 (uint64_t __a, uint64_t __b) +{ + return __builtin_mve_vcreateq_sv2di (__a, __b); +} + +__extension__ extern __inline int8x16_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_s8 (int8x16_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_sv16qi (__a, __imm); +} + +__extension__ extern __inline int16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_s16 (int16x8_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_sv8hi (__a, __imm); +} + +__extension__ extern __inline int32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_s32 (int32x4_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_sv4si (__a, __imm); +} + +__extension__ extern __inline uint8x16_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_u8 (uint8x16_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_uv16qi (__a, __imm); +} + +__extension__ extern __inline uint16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_u16 (uint16x8_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_uv8hi (__a, __imm); +} + +__extension__ extern __inline uint32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vshrq_n_u32 (uint32x4_t __a, const int __imm) +{ + return __builtin_mve_vshrq_n_uv4si (__a, __imm); +} + #if (__ARM_FEATURE_MVE & 2) /* MVE Floating point. */ __extension__ extern __inline void @@ -1165,6 +1281,34 @@ __arm_vcreateq_f32 (uint64_t __a, uint64_t __b) return __builtin_mve_vcreateq_fv4sf (__a, __b); } +__extension__ extern __inline int16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcvtq_n_s16_f16 (float16x8_t __a, const int __imm6) +{ + return __builtin_mve_vcvtq_n_from_f_sv8hi (__a, __imm6); +} + +__extension__ extern __inline int32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcvtq_n_s32_f32 (float32x4_t __a, const int __imm6) +{ + return __builtin_mve_vcvtq_n_from_f_sv4si (__a, __imm6); +} + +__extension__ extern __inline uint16x8_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcvtq_n_u16_f16 (float16x8_t __a, const int __imm6) +{ + return __builtin_mve_vcvtq_n_from_f_uv8hi (__a, __imm6); +} + +__extension__ extern __inline uint32x4_t +__attribute__ ((__always_inline__, __gnu_inline__, __artificial__)) +__arm_vcvtq_n_u32_f32 (float32x4_t __a, const int __imm6) +{ + return __builtin_mve_vcvtq_n_from_f_uv4si (__a, __imm6); +} + #endif enum { @@ -1598,6 +1742,16 @@ extern void *__ARM_undef; int (*)[__ARM_mve_type_int16x8_t]: __arm_vqnegq_s16 (__ARM_mve_coerce(__p0, int16x8_t)), \ int (*)[__ARM_mve_type_int32x4_t]: __arm_vqnegq_s32 (__ARM_mve_coerce(__p0, int32x4_t)));}) +#define vshrq(p0,p1) __arm_vshrq(p0,p1) +#define __arm_vshrq(p0,p1) ({ __typeof(p0) __p0 = (p0); \ + _Generic( (int (*)[__ARM_mve_typeid(__p0)])0, \ + int (*)[__ARM_mve_type_int8x16_t]: __arm_vshrq_n_s8 (__ARM_mve_coerce(__p0, int8x16_t), p1), \ + int (*)[__ARM_mve_type_int16x8_t]: __arm_vshrq_n_s16 (__ARM_mve_coerce(__p0, int16x8_t), p1), \ + int (*)[__ARM_mve_type_int32x4_t]: __arm_vshrq_n_s32 (__ARM_mve_coerce(__p0, int32x4_t), p1), \ + int (*)[__ARM_mve_type_uint8x16_t]: __arm_vshrq_n_u8 (__ARM_mve_coerce(__p0, uint8x16_t), p1), \ + int (*)[__ARM_mve_type_uint16x8_t]: __arm_vshrq_n_u16 (__ARM_mve_coerce(__p0, uint16x8_t), p1), \ + int (*)[__ARM_mve_type_uint32x4_t]: __arm_vshrq_n_u32 (__ARM_mve_coerce(__p0, uint32x4_t), p1));}) + #endif /* MVE Floating point. */ #ifdef __cplusplus diff --git a/gcc/config/arm/arm_mve_builtins.def b/gcc/config/arm/arm_mve_builtins.def index 8d1e4fac3d75e87fbe334e64e1073cb1fef0d96d..eae37ba831361401241ed91950169f06dee29c3e 100644 --- a/gcc/config/arm/arm_mve_builtins.def +++ b/gcc/config/arm/arm_mve_builtins.def @@ -6,7 +6,7 @@ GCC is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published - by the Free Software Foundation; either version 3, or (at your + by the Free Software Foundation; either version 3, or (at your option) any later version. GCC is distributed in the hope that it will be useful, but WITHOUT @@ -81,3 +81,9 @@ VAR2 (BINOP_NONE_NONE_NONE, vbrsrq_n_f, v8hf, v4sf) VAR2 (BINOP_NONE_NONE_IMM, vcvtq_n_to_f_s, v8hf, v4sf) VAR2 (BINOP_NONE_UNONE_IMM, vcvtq_n_to_f_u, v8hf, v4sf) VAR2 (BINOP_NONE_UNONE_UNONE, vcreateq_f, v8hf, v4sf) +VAR2 (BINOP_UNONE_NONE_IMM, vcvtq_n_from_f_u, v8hi, v4si) +VAR2 (BINOP_NONE_NONE_IMM, vcvtq_n_from_f_s, v8hi, v4si) +VAR4 (BINOP_UNONE_UNONE_UNONE, vcreateq_u, v16qi, v8hi, v4si, v2di) +VAR4 (BINOP_NONE_UNONE_UNONE, vcreateq_s, v16qi, v8hi, v4si, v2di) +VAR3 (BINOP_UNONE_UNONE_IMM, vshrq_n_u, v16qi, v8hi, v4si) +VAR3 (BINOP_NONE_NONE_IMM, vshrq_n_s, v16qi, v8hi, v4si) diff --git a/gcc/config/arm/constraints.md b/gcc/config/arm/constraints.md index 86cd82334533acbe58be4a0c547cf38737d2bb84..978674991d4bf0fa2082f38a630f2c9e845a7881 100644 --- a/gcc/config/arm/constraints.md +++ b/gcc/config/arm/constraints.md @@ -35,7 +35,7 @@ ;; in ARM/Thumb-2 state: Da, Db, Dc, Dd, Dn, DN, Dm, Dl, DL, Do, Dv, Dy, Di, ;; Dt, Dp, Dz, Tu ;; in Thumb-1 state: Pa, Pb, Pc, Pd, Pe -;; in Thumb-2 state: Ha, Pj, PJ, Ps, Pt, Pu, Pv, Pw, Px, Py, Pz, Rd +;; in Thumb-2 state: Ha, Pj, PJ, Ps, Pt, Pu, Pv, Pw, Px, Py, Pz, Rd, Rf, Rb ;; in all states: Pf, Pg, UM, U1 ;; The following memory constraints have been used: @@ -56,6 +56,16 @@ (and (match_code "const_int") (match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 16"))) +(define_constraint "Rb" + "@internal In Thumb-2 state a constant in range 1 to 8" + (and (match_code "const_int") + (match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 8"))) + +(define_constraint "Rf" + "@internal In Thumb-2 state a constant in range 1 to 32" + (and (match_code "const_int") + (match_test "TARGET_HAVE_MVE && ival >= 1 && ival <= 32"))) + (define_register_constraint "t" "TARGET_32BIT ? VFP_LO_REGS : NO_REGS" "The VFP registers @code{s0}-@code{s31}.") diff --git a/gcc/config/arm/mve.md b/gcc/config/arm/mve.md index 06701b024dabae85446bb60060a8b331e540cd6d..4ae7646cf20683ec147af11f398d7460bd65f7ba 100644 --- a/gcc/config/arm/mve.md +++ b/gcc/config/arm/mve.md @@ -22,6 +22,7 @@ (define_mode_iterator MVE_types [V16QI V8HI V4SI V2DI TI V8HF V4SF V2DF]) (define_mode_iterator MVE_VLD_ST [V16QI V8HI V4SI V8HF V4SF]) (define_mode_iterator MVE_0 [V8HF V4SF]) +(define_mode_iterator MVE_1 [V16QI V8HI V4SI V2DI]) (define_mode_iterator MVE_3 [V16QI V8HI]) (define_mode_iterator MVE_2 [V16QI V8HI V4SI]) (define_mode_iterator MVE_5 [V8HI V4SI]) @@ -38,7 +39,8 @@ VCVTPQ_U VCVTNQ_S VCVTNQ_U VCVTMQ_S VCVTMQ_U VADDLVQ_U VCTP8Q VCTP16Q VCTP32Q VCTP64Q VPNOT VCREATEQ_F VCVTQ_N_TO_F_S VCVTQ_N_TO_F_U VBRSRQ_N_F - VSUBQ_N_F]) + VSUBQ_N_F VCREATEQ_U VCREATEQ_S VSHRQ_N_S VSHRQ_N_U + VCVTQ_N_FROM_F_S VCVTQ_N_FROM_F_U]) (define_mode_attr MVE_CNVT [(V8HI "V8HF") (V4SI "V4SF") (V8HF "V8HI") (V4SF "V4SI")]) @@ -55,10 +57,16 @@ (VCVTNQ_U "u") (VCVTMQ_S "s") (VCVTMQ_U "u") (VCLZQ_U "u") (VCLZQ_S "s") (VREV32Q_U "u") (VREV32Q_S "s") (VADDLVQ_U "u") (VADDLVQ_S "s") - (VCVTQ_N_TO_F_S "s") (VCVTQ_N_TO_F_U "u")]) + (VCVTQ_N_TO_F_S "s") (VCVTQ_N_TO_F_U "u") + (VCREATEQ_U "u") (VCREATEQ_S "s") (VSHRQ_N_S "s") + (VSHRQ_N_U "u") (VCVTQ_N_FROM_F_S "s") + (VCVTQ_N_FROM_F_U "u")]) (define_int_attr mode1 [(VCTP8Q "8") (VCTP16Q "16") (VCTP32Q "32") (VCTP64Q "64")]) +(define_mode_attr MVE_pred2 [(V16QI "mve_imm_8") (V8HI "mve_imm_16") + (V4SI "mve_imm_32")]) +(define_mode_attr MVE_constraint2 [(V16QI "Rb") (V8HI "Rd") (V4SI "Rf")]) (define_int_iterator VCVTQ_TO_F [VCVTQ_TO_F_S VCVTQ_TO_F_U]) (define_int_iterator VMVNQ_N [VMVNQ_N_U VMVNQ_N_S]) @@ -79,6 +87,9 @@ (define_int_iterator VADDLVQ [VADDLVQ_U VADDLVQ_S]) (define_int_iterator VCTPQ [VCTP8Q VCTP16Q VCTP32Q VCTP64Q]) (define_int_iterator VCVTQ_N_TO_F [VCVTQ_N_TO_F_S VCVTQ_N_TO_F_U]) +(define_int_iterator VCREATEQ [VCREATEQ_U VCREATEQ_S]) +(define_int_iterator VSHRQ_N [VSHRQ_N_S VSHRQ_N_U]) +(define_int_iterator VCVTQ_N_FROM_F [VCVTQ_N_FROM_F_S VCVTQ_N_FROM_F_U]) (define_insn "*mve_mov<mode>" [(set (match_operand:MVE_types 0 "s_register_operand" "=w,w,r,w,w,r,w") @@ -743,3 +754,48 @@ "vmov %q0[2], %q0[0], %Q2, %Q1\;vmov %q0[3], %q0[1], %R2, %R1" [(set_attr "type" "mve_move") (set_attr "length""8")]) + +;; +;; [vcreateq_u, vcreateq_s]) +;; +(define_insn "mve_vcreateq_<supf><mode>" + [ + (set (match_operand:MVE_1 0 "s_register_operand" "=w") + (unspec:MVE_1 [(match_operand:DI 1 "s_register_operand" "r") + (match_operand:DI 2 "s_register_operand" "r")] + VCREATEQ)) + ] + "TARGET_HAVE_MVE" + "vmov %q0[2], %q0[0], %Q2, %Q1\;vmov %q0[3], %q0[1], %R2, %R1" + [(set_attr "type" "mve_move") + (set_attr "length""8")]) + +;; +;; [vshrq_n_s, vshrq_n_u]) +;; +(define_insn "mve_vshrq_n_<supf><mode>" + [ + (set (match_operand:MVE_2 0 "s_register_operand" "=w") + (unspec:MVE_2 [(match_operand:MVE_2 1 "s_register_operand" "w") + (match_operand:SI 2 "<MVE_pred2>" "<MVE_constraint2>")] + VSHRQ_N)) + ] + "TARGET_HAVE_MVE" + "vshr.<supf>%#<V_sz_elem>\t%q0, %q1, %2" + [(set_attr "type" "mve_move") +]) + +;; +;; [vcvtq_n_from_f_s, vcvtq_n_from_f_u]) +;; +(define_insn "mve_vcvtq_n_from_f_<supf><mode>" + [ + (set (match_operand:MVE_5 0 "s_register_operand" "=w") + (unspec:MVE_5 [(match_operand:<MVE_CNVT> 1 "s_register_operand" "w") + (match_operand:SI 2 "mve_imm_16" "Rd")] + VCVTQ_N_FROM_F)) + ] + "TARGET_HAVE_MVE && TARGET_HAVE_MVE_FLOAT" + "vcvt.<supf>%#<V_sz_elem>.f%#<V_sz_elem>\t%q0, %q1, %2" + [(set_attr "type" "mve_move") +]) diff --git a/gcc/config/arm/predicates.md b/gcc/config/arm/predicates.md index 80a44bb32cbce1ee3f850a44b70a0e4ceed548be..2e038705b3594ef8756bbb56c46d3285261e946e 100644 --- a/gcc/config/arm/predicates.md +++ b/gcc/config/arm/predicates.md @@ -35,6 +35,14 @@ (define_predicate "mve_imm_16" (match_test "satisfies_constraint_Rd (op)")) +;; True for immediates in the range of 1 to 8 for MVE. +(define_predicate "mve_imm_8" + (match_test "satisfies_constraint_Rb (op)")) + +;; True for immediates in the range of 1 to 32 for MVE. +(define_predicate "mve_imm_32" + (match_test "satisfies_constraint_Rf (op)")) + ; Predicate for stack protector guard's address in ; stack_protect_combined_set_insn and stack_protect_combined_test_insn patterns (define_predicate "guard_addr_operand" diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s16.c new file mode 100644 index 0000000000000000000000000000000000000000..fad5de2e5d2603abac244a6554ee2351c3b05ddb --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s16.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int16x8_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_s16 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s32.c new file mode 100644 index 0000000000000000000000000000000000000000..b6013be8481b8d8a800eca2a2e275720283586d9 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s32.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int32x4_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_s32 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s64.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s64.c new file mode 100644 index 0000000000000000000000000000000000000000..74372bef4eaa53772765029ce20e51ef3ecb80d6 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s64.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int64x2_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_s64 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s8.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s8.c new file mode 100644 index 0000000000000000000000000000000000000000..9884f0435780c7895c0cf1825d4d024f496e5a33 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_s8.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int8x16_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_s8 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u16.c new file mode 100644 index 0000000000000000000000000000000000000000..cc60f6cd4dd7370dae5245a6499383659705af7e --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u16.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint16x8_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_u16 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u32.c new file mode 100644 index 0000000000000000000000000000000000000000..43a49bc0a4a86c7a090151e7a5f4b6887496b520 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u32.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint32x4_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_u32 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u64.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u64.c new file mode 100644 index 0000000000000000000000000000000000000000..d798273dab689fc95746c0788d31c7d56244dd67 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u64.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint64x2_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_u64 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u8.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u8.c new file mode 100644 index 0000000000000000000000000000000000000000..7bb6a75ab8a2d52834798f39e8fff1e43ec0cfd9 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcreateq_u8.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint8x16_t +foo (uint64_t a, uint64_t b) +{ + return vcreateq_u8 (a, b); +} + +/* { dg-final { scan-assembler "vmov" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s16_f16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s16_f16.c new file mode 100644 index 0000000000000000000000000000000000000000..f40561848707ba059506fb532b052489823262cb --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s16_f16.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int16x8_t +foo (float16x8_t a) +{ + return vcvtq_n_s16_f16 (a, 1); +} + +/* { dg-final { scan-assembler "vcvt.s16.f16" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s32_f32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s32_f32.c new file mode 100644 index 0000000000000000000000000000000000000000..1bb98c254f5ba72af2416b952c5c96b315d148a9 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_s32_f32.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int32x4_t +foo (float32x4_t a) +{ + return vcvtq_n_s32_f32 (a, 1); +} + +/* { dg-final { scan-assembler "vcvt.s32.f32" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u16_f16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u16_f16.c new file mode 100644 index 0000000000000000000000000000000000000000..7bded9d70b55b06a872789118df65f2cced0d7a1 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u16_f16.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint16x8_t +foo (float16x8_t a) +{ + return vcvtq_n_u16_f16 (a, 1); +} + +/* { dg-final { scan-assembler "vcvt.u16.f16" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u32_f32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u32_f32.c new file mode 100644 index 0000000000000000000000000000000000000000..cbcf93b3e102d2e47c18792d641f0e8f56fbdbe1 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vcvtq_n_u32_f32.c @@ -0,0 +1,13 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve.fp -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint32x4_t +foo (float32x4_t a) +{ + return vcvtq_n_u32_f32 (a, 1); +} + +/* { dg-final { scan-assembler "vcvt.u32.f32" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s16.c new file mode 100644 index 0000000000000000000000000000000000000000..71950e35e9f1a9185f45858bb9f4e5def844644d --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s16.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int16x8_t +foo (int16x8_t a) +{ + return vshrq_n_s16 (a, 16); +} + +/* { dg-final { scan-assembler "vshr.s16" } } */ + +int16x8_t +foo1 (int16x8_t a) +{ + return vshrq (a, 16); +} + +/* { dg-final { scan-assembler "vshr.s16" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s32.c new file mode 100644 index 0000000000000000000000000000000000000000..b31d7a479e778f8ef9bc50c75b07983de7f93fe2 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s32.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int32x4_t +foo (int32x4_t a) +{ + return vshrq_n_s32 (a, 32); +} + +/* { dg-final { scan-assembler "vshr.s32" } } */ + +int32x4_t +foo1 (int32x4_t a) +{ + return vshrq (a, 32); +} + +/* { dg-final { scan-assembler "vshr.s32" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s8.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s8.c new file mode 100644 index 0000000000000000000000000000000000000000..461a268501f86ad9ead5dbe68b21c438d6257dbe --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_s8.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +int8x16_t +foo (int8x16_t a) +{ + return vshrq_n_s8 (a, 8); +} + +/* { dg-final { scan-assembler "vshr.s8" } } */ + +int8x16_t +foo1 (int8x16_t a) +{ + return vshrq (a, 8); +} + +/* { dg-final { scan-assembler "vshr.s8" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u16.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u16.c new file mode 100644 index 0000000000000000000000000000000000000000..97c6323b4c645fb78fb28e4ab8ef83855fb60037 --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u16.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint16x8_t +foo (uint16x8_t a) +{ + return vshrq_n_u16 (a, 16); +} + +/* { dg-final { scan-assembler "vshr.u16" } } */ + +uint16x8_t +foo1 (uint16x8_t a) +{ + return vshrq (a, 16); +} + +/* { dg-final { scan-assembler "vshr.u16" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u32.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u32.c new file mode 100644 index 0000000000000000000000000000000000000000..42006566f635b81268e01d3e18426bc294484c5b --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u32.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint32x4_t +foo (uint32x4_t a) +{ + return vshrq_n_u32 (a, 32); +} + +/* { dg-final { scan-assembler "vshr.u32" } } */ + +uint32x4_t +foo1 (uint32x4_t a) +{ + return vshrq (a, 32); +} + +/* { dg-final { scan-assembler "vshr.u32" } } */ diff --git a/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u8.c b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u8.c new file mode 100644 index 0000000000000000000000000000000000000000..25679297b6d5165fabfd77953f48b0af36867dad --- /dev/null +++ b/gcc/testsuite/gcc.target/arm/mve/intrinsics/vshrq_n_u8.c @@ -0,0 +1,21 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-march=armv8.1-m.main+mve -mfloat-abi=hard -O2" } */ +/* { dg-skip-if "Skip if not auto" {*-*-*} {"-mfpu=*"} {"-mfpu=auto"} } */ + +#include "arm_mve.h" + +uint8x16_t +foo (uint8x16_t a) +{ + return vshrq_n_u8 (a, 8); +} + +/* { dg-final { scan-assembler "vshr.u8" } } */ + +uint8x16_t +foo1 (uint8x16_t a) +{ + return vshrq (a, 8); +} + +/* { dg-final { scan-assembler "vshr.u8" } } */