aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--gcc/ChangeLog27
-rw-r--r--gcc/config/aarch64/aarch64-sve-builtins.cc6
-rw-r--r--gcc/config/aarch64/aarch64-sve.md107
-rw-r--r--gcc/config/aarch64/aarch64.c46
-rw-r--r--gcc/testsuite/ChangeLog15
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_1.c31
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_10.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_2.c31
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_3.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_4.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_5.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_6.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_7.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_8.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/load_extend_9.c21
-rw-r--r--gcc/testsuite/gcc.target/aarch64/sve/reduc_4.c2
16 files changed, 358 insertions, 75 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index cccc042..83931c5 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,5 +1,32 @@
2019-11-16 Richard Sandiford <richard.sandiford@arm.com>
+ * config/aarch64/aarch64-sve.md
+ (@aarch64_load_<ANY_EXTEND:optab><VNx8_WIDE:mode><VNx8_NARROW:mode>):
+ (@aarch64_load_<ANY_EXTEND:optab><VNx4_WIDE:mode><VNx4_NARROW:mode>)
+ (@aarch64_load_<ANY_EXTEND:optab><VNx2_WIDE:mode><VNx2_NARROW:mode>):
+ Combine into...
+ (@aarch64_load_<ANY_EXTEND:optab><SVE_HSDI:mode><SVE_PARTIAL_I:mode>):
+ ...this new pattern, handling extension to partial modes as well
+ as full modes. Describe the extension as a predicated rather than
+ unpredicated extension.
+ (@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx8_WIDE:mode><VNx8_NARROW:mode>)
+ (@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx4_WIDE:mode><VNx4_NARROW:mode>)
+ (@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx2_WIDE:mode><VNx2_NARROW:mode>):
+ Combine into...
+ (@aarch64_ld<fn>f1_<ANY_EXTEND:optab><SVE_HSDI:mode><SVE_PARTIAL_I:mode>):
+ ...this new pattern, handling extension to partial modes as well
+ as full modes. Describe the extension as a predicated rather than
+ unpredicated extension.
+ * config/aarch64/aarch64-sve-builtins.cc
+ (function_expander::use_contiguous_load_insn): Add an extra
+ predicate for extending loads.
+ * config/aarch64/aarch64.c (aarch64_extending_load_p): New function.
+ (aarch64_sve_adjust_stmt_cost): Likewise.
+ (aarch64_add_stmt_cost): Use aarch64_sve_adjust_stmt_cost to adjust
+ the cost of SVE vector stmts.
+
+2019-11-16 Richard Sandiford <richard.sandiford@arm.com>
+
* config/aarch64/iterators.md (SVE_HSDI): New mode iterator.
(narrower_mask): Handle VNx4HI, VNx2HI and VNx2SI.
* config/aarch64/aarch64-sve.md
diff --git a/gcc/config/aarch64/aarch64-sve-builtins.cc b/gcc/config/aarch64/aarch64-sve-builtins.cc
index 63d903d..27736b9 100644
--- a/gcc/config/aarch64/aarch64-sve-builtins.cc
+++ b/gcc/config/aarch64/aarch64-sve-builtins.cc
@@ -2790,7 +2790,9 @@ function_expander::use_vcond_mask_insn (insn_code icode,
}
/* Implement the call using instruction ICODE, which loads memory operand 1
- into register operand 0 under the control of predicate operand 2. */
+ into register operand 0 under the control of predicate operand 2.
+ Extending loads have a further predicate (operand 3) that nominally
+ controls the extension. */
rtx
function_expander::use_contiguous_load_insn (insn_code icode)
{
@@ -2799,6 +2801,8 @@ function_expander::use_contiguous_load_insn (insn_code icode)
add_output_operand (icode);
add_mem_operand (mem_mode, get_contiguous_base (mem_mode));
add_input_operand (icode, args[0]);
+ if (GET_MODE_UNIT_BITSIZE (mem_mode) < type_suffix (0).element_bits)
+ add_input_operand (icode, CONSTM1_RTX (VNx16BImode));
return generate_insn (icode);
}
diff --git a/gcc/config/aarch64/aarch64-sve.md b/gcc/config/aarch64/aarch64-sve.md
index 40aeb95..ce1bd58 100644
--- a/gcc/config/aarch64/aarch64-sve.md
+++ b/gcc/config/aarch64/aarch64-sve.md
@@ -1189,39 +1189,22 @@
;; -------------------------------------------------------------------------
;; Predicated load and extend, with 8 elements per 128-bit block.
-(define_insn "@aarch64_load_<ANY_EXTEND:optab><VNx8_WIDE:mode><VNx8_NARROW:mode>"
- [(set (match_operand:VNx8_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx8_WIDE
- (unspec:VNx8_NARROW
- [(match_operand:VNx8BI 2 "register_operand" "Upl")
- (match_operand:VNx8_NARROW 1 "memory_operand" "m")]
- UNSPEC_LD1_SVE)))]
- "TARGET_SVE"
- "ld1<ANY_EXTEND:s><VNx8_NARROW:Vesize>\t%0.<VNx8_WIDE:Vetype>, %2/z, %1"
-)
-
-;; Predicated load and extend, with 4 elements per 128-bit block.
-(define_insn "@aarch64_load_<ANY_EXTEND:optab><VNx4_WIDE:mode><VNx4_NARROW:mode>"
- [(set (match_operand:VNx4_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx4_WIDE
- (unspec:VNx4_NARROW
- [(match_operand:VNx4BI 2 "register_operand" "Upl")
- (match_operand:VNx4_NARROW 1 "memory_operand" "m")]
- UNSPEC_LD1_SVE)))]
- "TARGET_SVE"
- "ld1<ANY_EXTEND:s><VNx4_NARROW:Vesize>\t%0.<VNx4_WIDE:Vetype>, %2/z, %1"
-)
-
-;; Predicated load and extend, with 2 elements per 128-bit block.
-(define_insn "@aarch64_load_<ANY_EXTEND:optab><VNx2_WIDE:mode><VNx2_NARROW:mode>"
- [(set (match_operand:VNx2_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx2_WIDE
- (unspec:VNx2_NARROW
- [(match_operand:VNx2BI 2 "register_operand" "Upl")
- (match_operand:VNx2_NARROW 1 "memory_operand" "m")]
- UNSPEC_LD1_SVE)))]
- "TARGET_SVE"
- "ld1<ANY_EXTEND:s><VNx2_NARROW:Vesize>\t%0.<VNx2_WIDE:Vetype>, %2/z, %1"
+(define_insn_and_rewrite "@aarch64_load_<ANY_EXTEND:optab><SVE_HSDI:mode><SVE_PARTIAL_I:mode>"
+ [(set (match_operand:SVE_HSDI 0 "register_operand" "=w")
+ (unspec:SVE_HSDI
+ [(match_operand:<SVE_HSDI:VPRED> 3 "general_operand" "UplDnm")
+ (ANY_EXTEND:SVE_HSDI
+ (unspec:SVE_PARTIAL_I
+ [(match_operand:<SVE_PARTIAL_I:VPRED> 2 "register_operand" "Upl")
+ (match_operand:SVE_PARTIAL_I 1 "memory_operand" "m")]
+ UNSPEC_LD1_SVE))]
+ UNSPEC_PRED_X))]
+ "TARGET_SVE && (~<SVE_HSDI:narrower_mask> & <SVE_PARTIAL_I:self_mask>) == 0"
+ "ld1<ANY_EXTEND:s><SVE_PARTIAL_I:Vesize>\t%0.<SVE_HSDI:Vctype>, %2/z, %1"
+ "&& !CONSTANT_P (operands[3])"
+ {
+ operands[3] = CONSTM1_RTX (<SVE_HSDI:VPRED>mode);
+ }
)
;; -------------------------------------------------------------------------
@@ -1268,46 +1251,24 @@
;; - LDNF1W
;; -------------------------------------------------------------------------
-;; Predicated first-faulting or non-faulting load and extend, with 8 elements
-;; per 128-bit block.
-(define_insn "@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx8_WIDE:mode><VNx8_NARROW:mode>"
- [(set (match_operand:VNx8_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx8_WIDE
- (unspec:VNx8_NARROW
- [(match_operand:VNx8BI 2 "register_operand" "Upl")
- (match_operand:VNx8_NARROW 1 "aarch64_sve_ld<fn>f1_operand" "Ut<fn>")
- (reg:VNx16BI FFRT_REGNUM)]
- SVE_LDFF1_LDNF1)))]
- "TARGET_SVE"
- "ld<fn>f1<ANY_EXTEND:s><VNx8_NARROW:Vesize>\t%0.<VNx8_WIDE:Vetype>, %2/z, %1"
-)
-
-;; Predicated first-faulting or non-faulting load and extend, with 4 elements
-;; per 128-bit block.
-(define_insn "@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx4_WIDE:mode><VNx4_NARROW:mode>"
- [(set (match_operand:VNx4_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx4_WIDE
- (unspec:VNx4_NARROW
- [(match_operand:VNx4BI 2 "register_operand" "Upl")
- (match_operand:VNx4_NARROW 1 "aarch64_sve_ld<fn>f1_operand" "Ut<fn>")
- (reg:VNx16BI FFRT_REGNUM)]
- SVE_LDFF1_LDNF1)))]
- "TARGET_SVE"
- "ld<fn>f1<ANY_EXTEND:s><VNx4_NARROW:Vesize>\t%0.<VNx4_WIDE:Vetype>, %2/z, %1"
-)
-
-;; Predicated first-faulting or non-faulting load and extend, with 2 elements
-;; per 128-bit block.
-(define_insn "@aarch64_ld<fn>f1_<ANY_EXTEND:optab><VNx2_WIDE:mode><VNx2_NARROW:mode>"
- [(set (match_operand:VNx2_WIDE 0 "register_operand" "=w")
- (ANY_EXTEND:VNx2_WIDE
- (unspec:VNx2_NARROW
- [(match_operand:VNx2BI 2 "register_operand" "Upl")
- (match_operand:VNx2_NARROW 1 "aarch64_sve_ld<fn>f1_operand" "Ut<fn>")
- (reg:VNx16BI FFRT_REGNUM)]
- SVE_LDFF1_LDNF1)))]
- "TARGET_SVE"
- "ld<fn>f1<ANY_EXTEND:s><VNx2_NARROW:Vesize>\t%0.<VNx2_WIDE:Vetype>, %2/z, %1"
+;; Predicated first-faulting or non-faulting load and extend.
+(define_insn_and_rewrite "@aarch64_ld<fn>f1_<ANY_EXTEND:optab><SVE_HSDI:mode><SVE_PARTIAL_I:mode>"
+ [(set (match_operand:SVE_HSDI 0 "register_operand" "=w")
+ (unspec:SVE_HSDI
+ [(match_operand:<SVE_HSDI:VPRED> 3 "general_operand" "UplDnm")
+ (ANY_EXTEND:SVE_HSDI
+ (unspec:SVE_PARTIAL_I
+ [(match_operand:<SVE_PARTIAL_I:VPRED> 2 "register_operand" "Upl")
+ (match_operand:SVE_PARTIAL_I 1 "aarch64_sve_ld<fn>f1_operand" "Ut<fn>")
+ (reg:VNx16BI FFRT_REGNUM)]
+ SVE_LDFF1_LDNF1))]
+ UNSPEC_PRED_X))]
+ "TARGET_SVE && (~<SVE_HSDI:narrower_mask> & <SVE_PARTIAL_I:self_mask>) == 0"
+ "ld<fn>f1<ANY_EXTEND:s><SVE_PARTIAL_I:Vesize>\t%0.<SVE_HSDI:Vctype>, %2/z, %1"
+ "&& !CONSTANT_P (operands[3])"
+ {
+ operands[3] = CONSTM1_RTX (<SVE_HSDI:VPRED>mode);
+ }
)
;; -------------------------------------------------------------------------
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index d175e1f..305c6da 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -12879,6 +12879,49 @@ aarch64_builtin_vectorization_cost (enum vect_cost_for_stmt type_of_cost,
}
}
+/* Return true if STMT_INFO extends the result of a load. */
+static bool
+aarch64_extending_load_p (stmt_vec_info stmt_info)
+{
+ gassign *assign = dyn_cast <gassign *> (stmt_info->stmt);
+ if (!assign || !CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (assign)))
+ return false;
+
+ tree rhs = gimple_assign_rhs1 (stmt_info->stmt);
+ tree lhs_type = TREE_TYPE (gimple_assign_lhs (assign));
+ tree rhs_type = TREE_TYPE (rhs);
+ if (!INTEGRAL_TYPE_P (lhs_type)
+ || !INTEGRAL_TYPE_P (rhs_type)
+ || TYPE_PRECISION (lhs_type) <= TYPE_PRECISION (rhs_type))
+ return false;
+
+ stmt_vec_info def_stmt_info = stmt_info->vinfo->lookup_def (rhs);
+ return (def_stmt_info
+ && STMT_VINFO_DATA_REF (def_stmt_info)
+ && DR_IS_READ (STMT_VINFO_DATA_REF (def_stmt_info)));
+}
+
+/* STMT_COST is the cost calculated by aarch64_builtin_vectorization_cost
+ for STMT_INFO, which has cost kind KIND. Adjust the cost as necessary
+ for SVE targets. */
+static unsigned int
+aarch64_sve_adjust_stmt_cost (vect_cost_for_stmt kind, stmt_vec_info stmt_info,
+ unsigned int stmt_cost)
+{
+ /* Unlike vec_promote_demote, vector_stmt conversions do not change the
+ vector register size or number of units. Integer promotions of this
+ type therefore map to SXT[BHW] or UXT[BHW].
+
+ Most loads have extending forms that can do the sign or zero extension
+ on the fly. Optimistically assume that a load followed by an extension
+ will fold to this form during combine, and that the extension therefore
+ comes for free. */
+ if (kind == vector_stmt && aarch64_extending_load_p (stmt_info))
+ stmt_cost = 0;
+
+ return stmt_cost;
+}
+
/* Implement targetm.vectorize.add_stmt_cost. */
static unsigned
aarch64_add_stmt_cost (void *data, int count, enum vect_cost_for_stmt kind,
@@ -12894,6 +12937,9 @@ aarch64_add_stmt_cost (void *data, int count, enum vect_cost_for_stmt kind,
int stmt_cost =
aarch64_builtin_vectorization_cost (kind, vectype, misalign);
+ if (stmt_info && vectype && aarch64_sve_mode_p (TYPE_MODE (vectype)))
+ stmt_cost = aarch64_sve_adjust_stmt_cost (kind, stmt_info, stmt_cost);
+
/* Statements in an inner loop relative to the loop being
vectorized are weighted more heavily. The value here is
arbitrary and could potentially be improved with analysis. */
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 6a753a2..e13aa9f 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,5 +1,20 @@
2019-11-16 Richard Sandiford <richard.sandiford@arm.com>
+ * gcc.target/aarch64/sve/load_extend_1.c: New test.
+ * gcc.target/aarch64/sve/load_extend_2.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_3.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_4.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_5.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_6.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_7.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_8.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_9.c: Likewise.
+ * gcc.target/aarch64/sve/load_extend_10.c: Likewise.
+ * gcc.target/aarch64/sve/reduc_4.c: Add
+ --param aarch64-sve-compare-costs=0.
+
+2019-11-16 Richard Sandiford <richard.sandiford@arm.com>
+
* gcc.target/aarch64/sve/cost_model_1.c: Expect the loop to be
vectorized with bytes stored in 32-bit containers.
* gcc.target/aarch64/sve/extend_1.c: New test.
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_1.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_1.c
new file mode 100644
index 0000000..d881fc0
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_1.c
@@ -0,0 +1,31 @@
+/* { dg-options "-O2 -ftree-vectorize" } */
+
+#include <stdint.h>
+
+#define TEST_LOOP(TYPE1, TYPE2) \
+ void \
+ f_##TYPE1##_##TYPE2 (TYPE1 *restrict dst, TYPE1 *restrict src1, \
+ TYPE2 *restrict src2, int n) \
+ { \
+ for (int i = 0; i < n; ++i) \
+ dst[i] += src1[i] + src2[i]; \
+ }
+
+#define TEST_ALL(T) \
+ T (uint16_t, uint8_t) \
+ T (uint32_t, uint8_t) \
+ T (uint64_t, uint8_t) \
+ T (uint32_t, uint16_t) \
+ T (uint64_t, uint16_t) \
+ T (uint64_t, uint32_t)
+
+TEST_ALL (TEST_LOOP)
+
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.h,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tuxt.\t} } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_10.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_10.c
new file mode 100644
index 0000000..494c4af
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_10.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint32_t *src1, int16_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (int32_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1sh\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tsxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tsxtw\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_2.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_2.c
new file mode 100644
index 0000000..a21f2f9
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_2.c
@@ -0,0 +1,31 @@
+/* { dg-options "-O2 -ftree-vectorize" } */
+
+#include <stdint.h>
+
+#define TEST_LOOP(TYPE1, TYPE2) \
+ void \
+ f_##TYPE1##_##TYPE2 (TYPE1 *restrict dst, TYPE1 *restrict src1, \
+ TYPE2 *restrict src2, int n) \
+ { \
+ for (int i = 0; i < n; ++i) \
+ dst[i] += src1[i] + src2[i]; \
+ }
+
+#define TEST_ALL(T) \
+ T (int16_t, int8_t) \
+ T (int32_t, int8_t) \
+ T (int64_t, int8_t) \
+ T (int32_t, int16_t) \
+ T (int64_t, int16_t) \
+ T (int64_t, int32_t)
+
+TEST_ALL (TEST_LOOP)
+
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.h,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1sh\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1sh\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1sw\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-not {\tsxt.\t} } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_3.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_3.c
new file mode 100644
index 0000000..defd24d
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_3.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint32_t *dst, uint16_t *src1, uint8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (uint16_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.s,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.h, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tuxth\tz[0-9]+\.s,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_4.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_4.c
new file mode 100644
index 0000000..9fe31ef
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_4.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint16_t *src1, uint8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (uint16_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.h, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tuxth\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_5.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_5.c
new file mode 100644
index 0000000..9876293
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_5.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint32_t *src1, uint8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (uint32_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1b\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tuxtw\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_6.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_6.c
new file mode 100644
index 0000000..8aee86f
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_6.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint32_t *src1, uint16_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (uint32_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tuxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tuxtw\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_7.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_7.c
new file mode 100644
index 0000000..397f4f6
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_7.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint32_t *dst, uint16_t *src1, int8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (int16_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.s,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.s,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.h, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tsxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tsxth\tz[0-9]+\.s,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_8.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_8.c
new file mode 100644
index 0000000..cb81f79
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_8.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint16_t *src1, int8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (int16_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1h\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.h, z[0-9]+\.h, z[0-9]+\.h\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tsxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tsxth\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/load_extend_9.c b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_9.c
new file mode 100644
index 0000000..10a6b8a
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/load_extend_9.c
@@ -0,0 +1,21 @@
+/* { dg-options "-O2 -ftree-vectorize -msve-vector-bits=512" } */
+
+#include <stdint.h>
+
+void
+f1 (uint64_t *dst, uint32_t *src1, int8_t *src2)
+{
+ for (int i = 0; i < 7; ++i)
+ dst[i] += (int32_t) (src1[i] + src2[i]);
+}
+
+/* { dg-final { scan-assembler-times {\tld1sb\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1w\tz[0-9]+\.d,} 1 } } */
+/* { dg-final { scan-assembler-times {\tld1d\tz[0-9]+\.d,} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tadd\tz} 2 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.s, z[0-9]+\.s, z[0-9]+\.s\n} 1 } } */
+/* { dg-final { scan-assembler-times {\tadd\tz[0-9]+\.d, z[0-9]+\.d, z[0-9]+\.d\n} 1 } } */
+
+/* { dg-final { scan-assembler-times {\tsxt.\t} 1 } } */
+/* { dg-final { scan-assembler-times {\tsxtw\tz[0-9]+\.d,} 1 } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/reduc_4.c b/gcc/testsuite/gcc.target/aarch64/sve/reduc_4.c
index eb4b231..b0260c9 100644
--- a/gcc/testsuite/gcc.target/aarch64/sve/reduc_4.c
+++ b/gcc/testsuite/gcc.target/aarch64/sve/reduc_4.c
@@ -1,5 +1,5 @@
/* { dg-do compile } */
-/* { dg-options "-O2 -ftree-vectorize -ffast-math" } */
+/* { dg-options "-O2 -ftree-vectorize -ffast-math --param aarch64-sve-compare-costs=0" } */
double
f (double *restrict a, double *restrict b, int *lookup)
generated by cgit v1.1 at 2024-11-11 12:11:38 +0000