1 /* Tuning model description for the Ampere1B core.
2 Copyright (C) 2023-2024 Free Software Foundation, Inc.
4 This file is part of GCC.
6 GCC is free software; you can redistribute it and/or modify it
7 under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3, or (at your option)
11 GCC is distributed in the hope that it will be useful, but
12 WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with GCC; see the file COPYING3. If not see
18 <http://www.gnu.org/licenses/>. */
20 #ifndef GCC_AARCH64_H_AMPERE1B
21 #define GCC_AARCH64_H_AMPERE1B
25 static const cpu_prefetch_tune ampere1b_prefetch_tune
=
28 64, /* l1_cache_size */
29 64, /* l1_cache_line_size */
30 2048, /* l2_cache_size */
31 true, /* prefetch_dynamic_strides */
32 -1, /* minimum_stride */
33 -1 /* default_opt_level */
36 static const advsimd_vec_cost ampere1b_advsimd_vector_cost
=
38 1, /* int_stmt_cost */
40 0, /* ld2_st2_permute_cost */
41 0, /* ld3_st3_permute_cost */
42 0, /* ld4_st4_permute_cost */
44 8, /* reduc_i8_cost */
45 6, /* reduc_i16_cost */
46 4, /* reduc_i32_cost */
47 2, /* reduc_i64_cost */
48 9, /* reduc_f16_cost */
49 6, /* reduc_f32_cost */
50 3, /* reduc_f64_cost */
51 5, /* store_elt_extra_cost */
52 5, /* vec_to_scalar_cost */
53 5, /* scalar_to_vec_cost */
54 4, /* align_load_cost */
55 4, /* unalign_load_cost */
56 1, /* unalign_store_cost */
60 /* Ampere-1B costs for vector insn classes. */
61 static const struct cpu_vector_cost ampere1b_vector_cost
=
63 1, /* scalar_int_stmt_cost */
64 3, /* scalar_fp_stmt_cost */
65 4, /* scalar_load_cost */
66 1, /* scalar_store_cost */
67 1, /* cond_taken_branch_cost */
68 1, /* cond_not_taken_branch_cost */
69 &ere1b_advsimd_vector_cost
, /* advsimd */
71 nullptr /* issue_info */
74 static const struct tune_params ampere1b_tunings
=
76 &ere1b_extra_costs
,
77 &generic_addrcost_table
,
78 &generic_regmove_cost
,
79 &ere1b_vector_cost
,
81 &generic_approx_modes
,
82 SVE_NOT_IMPLEMENTED
, /* sve_width */
91 (AARCH64_FUSE_ADRP_ADD
| AARCH64_FUSE_AES_AESMC
|
92 AARCH64_FUSE_MOV_MOVK
| AARCH64_FUSE_MOVK_MOVK
|
93 AARCH64_FUSE_ALU_BRANCH
/* adds, ands, bics, ccmp, ccmn */ |
94 AARCH64_FUSE_CMP_BRANCH
| AARCH64_FUSE_ALU_CBZ
|
95 AARCH64_FUSE_ADDSUB_2REG_CONST1
),
97 "32", /* function_align. */
98 "4", /* jump_align. */
99 "32:16", /* loop_align. */
100 2, /* int_reassoc_width. */
101 4, /* fp_reassoc_width. */
102 4, /* fma_reassoc_width. */
103 2, /* vec_reassoc_width. */
104 2, /* min_div_recip_mul_sf. */
105 2, /* min_div_recip_mul_df. */
106 0, /* max_case_values. */
107 tune_params::AUTOPREFETCHER_STRONG
, /* autoprefetcher_model. */
108 (AARCH64_EXTRA_TUNE_CHEAP_SHIFT_EXTEND
109 | AARCH64_EXTRA_TUNE_AVOID_CROSS_LOOP_FMA
110 | AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA
), /* tune_flags. */
111 &ere1b_prefetch_tune
,
112 AARCH64_LDP_STP_POLICY_ALIGNED
, /* ldp_policy_model. */
113 AARCH64_LDP_STP_POLICY_ALIGNED
/* stp_policy_model. */
116 #endif /* GCC_AARCH64_H_AMPERE1B */