Updated in V2 > > Can you instead of mangling in float support use separate (match like > for the below cases? I tried, but reported duplicated defination since they share same pattern like
(cond (simple_comparison@6 @0 @1) (convert@4 @2) (convert@5 @3)) No idea how to split that. > > > @@ -11308,6 +11311,50 @@ and, > > && single_use (@4) > > && single_use (@5)))) > > > > +(match (cond_expr_convert_p @0 @2 @3 @6) > > + (cond (simple_comparison@6 @0 @1) (float@4 @2) (float@5 @3)) > > + (if (SCALAR_FLOAT_TYPE_P (type) && !flag_trapping_math > > + && TYPE_PRECISION (type) != TYPE_PRECISION (TREE_TYPE (@0)) > > so this fails to constrain the comparison types (above we check > INTEGRAL_TYPE_P), > if it happens to be a vector type using TYPE_PRECISION will ICE. > > I think the main intent of the vectorizer pattern is to match up the > _size_ of the > vector elements, so maybe re-formulate the constraint this way with > operand_equal_p (TYPE_SIZE (type), TYPE_SIZE (TREE_TYPE (@0))) > Changed. > This is also because precision on floats is not equal to the number of bits in > the mode. > > > + && TYPE_PRECISION (TREE_TYPE (@0)) > > + == TYPE_PRECISION (TREE_TYPE (@2)) > > + && INTEGRAL_TYPE_P (TREE_TYPE (@2)) > > + && TREE_TYPE (@2) == TREE_TYPE (@3) > > + && single_use (@4) > > + && single_use (@5)))) > > + > > +(match (cond_expr_convert_p @0 @2 @3 @6) > > + (cond (simple_comparison@6 @0 @1) (fix_trunc@4 @2) (fix_trunc@5 @3)) > > + (if (INTEGRAL_TYPE_P (type) && !flag_trapping_math > > + && TYPE_PRECISION (type) != TYPE_PRECISION (TREE_TYPE (@0)) > > + && TYPE_PRECISION (TREE_TYPE (@0)) > > + == TYPE_PRECISION (TREE_TYPE (@2)) > > + && SCALAR_FLOAT_TYPE_P (TREE_TYPE (@2)) > > + && TREE_TYPE (@2) == TREE_TYPE (@3) > > Please use types_match () instead of TREE_TYPE pointer compares. Changed. > > > + && single_use (@4) > > + && single_use (@5)))) > > + > > +(match (cond_expr_convert_p @0 @2 @3 @6) > > + (cond (simple_comparison@6 @0 @1) (REAL_CST@2) (convert@5 @3)) > > I think the same issue exists for INTEGER_CSTs. INTEGER_CSTs are already handled by vect_recog_over_widening_pattern. > > > + (if ((INTEGRAL_TYPE_P (type) > > + || (!flag_trapping_math && SCALAR_FLOAT_TYPE_P (type))) > > + && TYPE_PRECISION (type) != TYPE_PRECISION (TREE_TYPE (@0)) > > + && TYPE_PRECISION (TREE_TYPE (@0)) > > + == TYPE_PRECISION (TREE_TYPE (@3)) > > + && SCALAR_FLOAT_TYPE_P (TREE_TYPE (@3)) > > + && single_use (@5) > > + && const_unop (CONVERT_EXPR, TREE_TYPE (@3), @2)))) > > I'm not sure this is a good check? Say, for type == double and > typeof(@3) == float > the REAL_CST can have extra precision that you'd drop when rewriting this as > (double)(cnd ? (float)@2 : @3). You'd need to check the REAL_CST is exactly > representable in the type of @3. Same for a possible integer case. Same for If the REAL_CST is exactly reprensentable, const_unop can return corresponding tree. Otherwise NULL_TREE is returned. > handling fix_trunc/float when one case is a constant. > > Can you split the patch into two to separate out the handling of constants? Changed. Bootstrapped and regtested separately for the first patch and the second on x86_64-pc-linux-gnu{-m32,}. Ok for trunk? For floating point, !flag_trapping_math is needed for the pattern which transforms 2 conversions to 1 conversion, and may lose 1 potential trap. There shouldn't be any accuracy issue. gcc/ChangeLog: PR tree-optimization/103771 * match.pd (cond_expr_convert_p): Extend the match to handle scalar floating point type. * tree-vect-patterns.cc (vect_recog_cond_expr_convert_pattern): Handle floating point type. gcc/testsuite/ChangeLog: * gcc.target/i386/pr103771-4.c: New test. --- gcc/match.pd | 52 +++++++++++--- gcc/testsuite/gcc.target/i386/pr103771-4.c | 82 ++++++++++++++++++++++ gcc/tree-vect-patterns.cc | 8 ++- 3 files changed, 131 insertions(+), 11 deletions(-) create mode 100644 gcc/testsuite/gcc.target/i386/pr103771-4.c diff --git a/gcc/match.pd b/gcc/match.pd index ab496d923cc..789e3d33326 100644 --- a/gcc/match.pd +++ b/gcc/match.pd @@ -11294,26 +11294,58 @@ and, (match (ctz_table_index @1 @2 @3) (rshift (mult (bit_and:c (negate @1) @1) INTEGER_CST@2) INTEGER_CST@3)) +/* Floatint point/integer comparison and integer->integer + or floating point -> float point conversion. */ (match (cond_expr_convert_p @0 @2 @3 @6) (cond (simple_comparison@6 @0 @1) (convert@4 @2) (convert@5 @3)) - (if (INTEGRAL_TYPE_P (type) - && INTEGRAL_TYPE_P (TREE_TYPE (@2)) - && INTEGRAL_TYPE_P (TREE_TYPE (@0)) - && INTEGRAL_TYPE_P (TREE_TYPE (@3)) - && TYPE_PRECISION (type) != TYPE_PRECISION (TREE_TYPE (@0)) - && TYPE_PRECISION (TREE_TYPE (@0)) - == TYPE_PRECISION (TREE_TYPE (@2)) - && TYPE_PRECISION (TREE_TYPE (@0)) - == TYPE_PRECISION (TREE_TYPE (@3)) + (if ((INTEGRAL_TYPE_P (type) + || (!flag_trapping_math && SCALAR_FLOAT_TYPE_P (type))) + && ((INTEGRAL_TYPE_P (TREE_TYPE (@2)) + && INTEGRAL_TYPE_P (TREE_TYPE (@3))) + || (SCALAR_FLOAT_TYPE_P (TREE_TYPE (@2)) + && types_match (TREE_TYPE (@2), TREE_TYPE (@3)))) + && !operand_equal_p (TYPE_SIZE (type), TYPE_SIZE (TREE_TYPE (@0))) + && operand_equal_p (TYPE_SIZE (TREE_TYPE (@0)), + TYPE_SIZE (TREE_TYPE (@2))) + && operand_equal_p (TYPE_SIZE (TREE_TYPE (@0)), + TYPE_SIZE (TREE_TYPE (@3))) /* For vect_recog_cond_expr_convert_pattern, @2 and @3 can differ in signess when convert is truncation, but not ok for extension since it's sign_extend vs zero_extend. */ - && (TYPE_PRECISION (TREE_TYPE (@0)) > TYPE_PRECISION (type) + && (known_gt (tree_to_poly_uint64 (TYPE_SIZE (TREE_TYPE (@0))), + tree_to_poly_uint64 (TYPE_SIZE (type))) || (TYPE_UNSIGNED (TREE_TYPE (@2)) == TYPE_UNSIGNED (TREE_TYPE (@3)))) && single_use (@4) && single_use (@5)))) +/* Floating point or integer comparison and integer to floating point + conversion. */ +(match (cond_expr_convert_p @0 @2 @3 @6) + (cond (simple_comparison@6 @0 @1) (float@4 @2) (float@5 @3)) + (if (SCALAR_FLOAT_TYPE_P (type) && !flag_trapping_math + && INTEGRAL_TYPE_P (TREE_TYPE (@2)) + && types_match (TREE_TYPE (@2), TREE_TYPE (@3)) + && !operand_equal_p (TYPE_SIZE (type), TYPE_SIZE (TREE_TYPE (@0))) + && operand_equal_p (TYPE_SIZE (TREE_TYPE (@0)), + TYPE_SIZE (TREE_TYPE (@2))) + && single_use (@4) + && single_use (@5)))) + +/* Floating point or integer comparison and floating point to integer + conversion. */ +(match (cond_expr_convert_p @0 @2 @3 @6) + (cond (simple_comparison@6 @0 @1) (fix_trunc@4 @2) (fix_trunc@5 @3)) + (if (INTEGRAL_TYPE_P (type) && !flag_trapping_math + && SCALAR_FLOAT_TYPE_P (TREE_TYPE (@2)) + && types_match (TREE_TYPE (@2), TREE_TYPE (@3)) + && !operand_equal_p (TYPE_SIZE (type), + TYPE_SIZE (TREE_TYPE (@0))) + && operand_equal_p (TYPE_SIZE (TREE_TYPE (@0)), + TYPE_SIZE (TREE_TYPE (@2))) + && single_use (@4) + && single_use (@5)))) + (for bit_op (bit_and bit_ior bit_xor) (match (bitwise_induction_p @0 @2 @3) (bit_op:c diff --git a/gcc/testsuite/gcc.target/i386/pr103771-4.c b/gcc/testsuite/gcc.target/i386/pr103771-4.c new file mode 100644 index 00000000000..299337dce6d --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr103771-4.c @@ -0,0 +1,82 @@ +/* { dg-do compile } */ +/* { dg-options "-march=x86-64-v4 -Ofast -fdump-tree-vect-details" } */ +/* { dg-final { scan-assembler-not "kshift" { target { ! ia32 } } } } */ +/* { dg-final { scan-tree-dump-times "loop vectorized using 64 byte vectors" 6 "vect" { target { ! ia32 } } } } */ + +void +foo (float* a, float* b, int* c, int* d, long long* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + long long tmp = c[i]; + long long tmp2 = d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} + +void +foo1 (double* a, double* b, long long* c, long long* d, int* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + int tmp = (int)c[i]; + int tmp2 = (int)d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} + +void +foo2 (float* a, float* b, int* c, int* d, double* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + double tmp = c[i]; + double tmp2 = d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} + +void +foo3 (double* a, double* b, long long* c, long long* d, float* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + float tmp = c[i]; + float tmp2 = d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} + +void +foo4 (int* a, int* b, int* c, int* d, double* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + double tmp = c[i]; + double tmp2 = d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} + +void +foo5 (long long* a, long long* b, long long* c, long long* d, float* __restrict e, int n) +{ + for (int i = 0 ; i != n; i++) + { + float tmp = c[i]; + float tmp2 = d[i]; + if (a[i] < b[i]) + tmp = tmp2; + e[i] = tmp; + } +} diff --git a/gcc/tree-vect-patterns.cc b/gcc/tree-vect-patterns.cc index ca19add83c0..d8484766cf7 100644 --- a/gcc/tree-vect-patterns.cc +++ b/gcc/tree-vect-patterns.cc @@ -1098,6 +1098,7 @@ vect_recog_cond_expr_convert_pattern (vec_info *vinfo, tree lhs, match[4], temp, type, new_lhs, op2; gimple *cond_stmt; gimple *pattern_stmt; + enum tree_code code = NOP_EXPR; if (!last_stmt) return NULL; @@ -1111,6 +1112,11 @@ vect_recog_cond_expr_convert_pattern (vec_info *vinfo, vect_pattern_detected ("vect_recog_cond_expr_convert_pattern", last_stmt); + if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (lhs))) + code = INTEGRAL_TYPE_P (TREE_TYPE (match[1])) ? FLOAT_EXPR : CONVERT_EXPR; + else if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (match[1]))) + code = FIX_TRUNC_EXPR; + op2 = match[2]; type = TREE_TYPE (match[1]); if (TYPE_SIGN (type) != TYPE_SIGN (TREE_TYPE (match[2]))) @@ -1127,7 +1133,7 @@ vect_recog_cond_expr_convert_pattern (vec_info *vinfo, append_pattern_def_seq (vinfo, stmt_vinfo, cond_stmt, get_vectype_for_scalar_type (vinfo, type)); new_lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL); - pattern_stmt = gimple_build_assign (new_lhs, NOP_EXPR, temp); + pattern_stmt = gimple_build_assign (new_lhs, code, temp); *type_out = STMT_VINFO_VECTYPE (stmt_vinfo); if (dump_enabled_p ()) -- 2.34.1