From: Richard Earnshaw <Richard.Earnshaw@foss.arm.com>
To: Tamar Christina <tamar.christina@arm.com>, gcc-patches@gcc.gnu.org
Cc: nd@arm.com, Richard.Earnshaw@arm.com, Marcus.Shawcroft@arm.com,
Kyrylo.Tkachov@arm.com, richard.sandiford@arm.com
Subject: Re: [PATCH 4/6]AArch64: Add new generic-armv9-a CPU and make it the default for Armv9
Date: Thu, 16 Nov 2023 09:23:57 +0000 [thread overview]
Message-ID: <ff8b3f13-9b35-46b2-9722-9c2b2e0dcaac@foss.arm.com> (raw)
In-Reply-To: <ZVT6+jY9p1IiuycF@arm.com>
On 15/11/2023 17:08, Tamar Christina wrote:
> Hi All,
>
> This patch adds a new generic scheduling model "generic-armv9-a" and makes it
> the default for all Armv9 architectures.
>
> -mcpu=generic and -mtune=generic is kept around for those that really want the
> deprecated cost model.
>
> Bootstrapped Regtested on aarch64-none-linux-gnu and no issues.
>
> Ok for master?
>
> Thanks,
> Tamar
>
> gcc/ChangeLog:
>
> PR target/111370
> * config/aarch64/aarch64-arches.def (armv9-a, armv9.1-a, armv9.2-a,
> armv9.3-a): Update to generic-armv9-a.
> * config/aarch64/aarch64-cores.def (generic-armv9-a): New.
> * config/aarch64/aarch64-tune.md: Regenerate.
> * config/aarch64/aarch64.cc: Include generic_armv9_a.h.
> * config/aarch64/tuning_models/generic_armv9_a.h: New file.
OK, but see the comment on patch 3 about 'generic'.
R.
>
> --- inline copy of patch --
> diff --git a/gcc/config/aarch64/aarch64-arches.def b/gcc/config/aarch64/aarch64-arches.def
> index f89e4ea1f48acc2875c9a834d93d94c94163cddc..6b9a19c490ba0b35082077e877b19906138f039b 100644
> --- a/gcc/config/aarch64/aarch64-arches.def
> +++ b/gcc/config/aarch64/aarch64-arches.def
> @@ -40,9 +40,9 @@ AARCH64_ARCH("armv8.6-a", generic_armv8_a, V8_6A, 8, (V8_5A, I8MM, BF
> AARCH64_ARCH("armv8.7-a", generic_armv8_a, V8_7A, 8, (V8_6A, LS64))
> AARCH64_ARCH("armv8.8-a", generic_armv8_a, V8_8A, 8, (V8_7A, MOPS))
> AARCH64_ARCH("armv8-r", generic_armv8_a, V8R , 8, (V8_4A))
> -AARCH64_ARCH("armv9-a", generic, V9A , 9, (V8_5A, SVE2))
> -AARCH64_ARCH("armv9.1-a", generic, V9_1A, 9, (V8_6A, V9A))
> -AARCH64_ARCH("armv9.2-a", generic, V9_2A, 9, (V8_7A, V9_1A))
> -AARCH64_ARCH("armv9.3-a", generic, V9_3A, 9, (V8_8A, V9_2A))
> +AARCH64_ARCH("armv9-a", generic_armv9_a, V9A , 9, (V8_5A, SVE2))
> +AARCH64_ARCH("armv9.1-a", generic_armv9_a, V9_1A, 9, (V8_6A, V9A))
> +AARCH64_ARCH("armv9.2-a", generic_armv9_a, V9_2A, 9, (V8_7A, V9_1A))
> +AARCH64_ARCH("armv9.3-a", generic_armv9_a, V9_3A, 9, (V8_8A, V9_2A))
>
> #undef AARCH64_ARCH
> diff --git a/gcc/config/aarch64/aarch64-cores.def b/gcc/config/aarch64/aarch64-cores.def
> index 30f4dd04ed71823bc34c0c405d49963b6b2d1375..16752b77f4baf8d1aa8a5406826aa29e367120c5 100644
> --- a/gcc/config/aarch64/aarch64-cores.def
> +++ b/gcc/config/aarch64/aarch64-cores.def
> @@ -191,6 +191,7 @@ AARCH64_CORE("demeter", demeter, cortexa57, V9A, (I8MM, BF16, SVE2_BITPERM, RNG,
>
> /* Generic Architecture Processors. */
> AARCH64_CORE("generic", generic, cortexa53, V8A, (), generic, 0x0, 0x0, -1)
> -AARCH64_CORE("generic-armv8-a", generic_armv8_a, cortexa53, V8A, (), generic_armv8_a, 0x0, 0x0, -1)
> +AARCH64_CORE("generic-armv8-a", generic_armv8_a, cortexa53, V8A, (), generic_armv8_a, 0x0, 0x0, -1)
> +AARCH64_CORE("generic-armv9-a", generic_armv9_a, cortexa53, V9A, (), generic_armv9_a, 0x0, 0x0, -1)
>
> #undef AARCH64_CORE
> diff --git a/gcc/config/aarch64/aarch64-tune.md b/gcc/config/aarch64/aarch64-tune.md
> index 0a32056f255de455f47a0b7395dfef0af84c6b5e..61bb85211252970f0a0526929d6b88353bdd930f 100644
> --- a/gcc/config/aarch64/aarch64-tune.md
> +++ b/gcc/config/aarch64/aarch64-tune.md
> @@ -1,5 +1,5 @@
> ;; -*- buffer-read-only: t -*-
> ;; Generated automatically by gentune.sh from aarch64-cores.def
> (define_attr "tune"
> - "cortexa34,cortexa35,cortexa53,cortexa57,cortexa72,cortexa73,thunderx,thunderxt88p1,thunderxt88,octeontx,octeontxt81,octeontxt83,thunderxt81,thunderxt83,ampere1,ampere1a,emag,xgene1,falkor,qdf24xx,exynosm1,phecda,thunderx2t99p1,vulcan,thunderx2t99,cortexa55,cortexa75,cortexa76,cortexa76ae,cortexa77,cortexa78,cortexa78ae,cortexa78c,cortexa65,cortexa65ae,cortexx1,cortexx1c,neoversen1,ares,neoversee1,octeontx2,octeontx2t98,octeontx2t96,octeontx2t93,octeontx2f95,octeontx2f95n,octeontx2f95mm,a64fx,tsv110,thunderx3t110,neoversev1,zeus,neoverse512tvb,saphira,cortexa57cortexa53,cortexa72cortexa53,cortexa73cortexa35,cortexa73cortexa53,cortexa75cortexa55,cortexa76cortexa55,cortexr82,cortexa510,cortexa520,cortexa710,cortexa715,cortexa720,cortexx2,cortexx3,cortexx4,neoversen2,neoversev2,demeter,generic,generic_armv8_a"
> + "cortexa34,cortexa35,cortexa53,cortexa57,cortexa72,cortexa73,thunderx,thunderxt88p1,thunderxt88,octeontx,octeontxt81,octeontxt83,thunderxt81,thunderxt83,ampere1,ampere1a,emag,xgene1,falkor,qdf24xx,exynosm1,phecda,thunderx2t99p1,vulcan,thunderx2t99,cortexa55,cortexa75,cortexa76,cortexa76ae,cortexa77,cortexa78,cortexa78ae,cortexa78c,cortexa65,cortexa65ae,cortexx1,cortexx1c,neoversen1,ares,neoversee1,octeontx2,octeontx2t98,octeontx2t96,octeontx2t93,octeontx2f95,octeontx2f95n,octeontx2f95mm,a64fx,tsv110,thunderx3t110,neoversev1,zeus,neoverse512tvb,saphira,cortexa57cortexa53,cortexa72cortexa53,cortexa73cortexa35,cortexa73cortexa53,cortexa75cortexa55,cortexa76cortexa55,cortexr82,cortexa510,cortexa520,cortexa710,cortexa715,cortexa720,cortexx2,cortexx3,cortexx4,neoversen2,neoversev2,demeter,generic,generic_armv8_a,generic_armv9_a"
> (const (symbol_ref "((enum attr_tune) aarch64_tune)")))
> diff --git a/gcc/config/aarch64/aarch64.cc b/gcc/config/aarch64/aarch64.cc
> index 08635e0df9cfa02286f3950383a32f6f93d1b4e0..5bed5f84cef242ec01f8510c76a450f81a985521 100644
> --- a/gcc/config/aarch64/aarch64.cc
> +++ b/gcc/config/aarch64/aarch64.cc
> @@ -356,6 +356,7 @@ static const struct aarch64_flag_desc aarch64_tuning_flags[] =
> /* Tuning parameters. */
> #include "tuning_models/generic.h"
> #include "tuning_models/generic_armv8_a.h"
> +#include "tuning_models/generic_armv9_a.h"
> #include "tuning_models/cortexa35.h"
> #include "tuning_models/cortexa53.h"
> #include "tuning_models/cortexa57.h"
> diff --git a/gcc/config/aarch64/tuning_models/generic_armv9_a.h b/gcc/config/aarch64/tuning_models/generic_armv9_a.h
> new file mode 100644
> index 0000000000000000000000000000000000000000..c017468592a9dba74ddd432247aaf51a70bb34b5
> --- /dev/null
> +++ b/gcc/config/aarch64/tuning_models/generic_armv9_a.h
> @@ -0,0 +1,245 @@
> +/* Tuning model description for AArch64 architecture.
> + Copyright (C) 2009-2023 Free Software Foundation, Inc.
> +
> + This file is part of GCC.
> +
> + 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 option)
> + any later version.
> +
> + GCC is distributed in the hope that it will be useful, but
> + WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + General Public License for more details.
> +
> + You should have received a copy of the GNU General Public License
> + along with GCC; see the file COPYING3. If not see
> + <http://www.gnu.org/licenses/>. */
> +
> +#ifndef GCC_AARCH64_H_GENERIC_ARMV9_A
> +#define GCC_AARCH64_H_GENERIC_ARMV9_A
> +
> +#include "generic.h"
> +#include "generic_armv8_a.h"
> +
> +static const struct cpu_addrcost_table generic_armv9_a_addrcost_table =
> +{
> + {
> + 1, /* hi */
> + 0, /* si */
> + 0, /* di */
> + 1, /* ti */
> + },
> + 0, /* pre_modify */
> + 0, /* post_modify */
> + 2, /* post_modify_ld3_st3 */
> + 2, /* post_modify_ld4_st4 */
> + 0, /* register_offset */
> + 0, /* register_sextend */
> + 0, /* register_zextend */
> + 0 /* imm_offset */
> +};
> +
> +static const struct cpu_regmove_cost generic_armv9_a_regmove_cost =
> +{
> + 1, /* GP2GP */
> + /* Spilling to int<->fp instead of memory is recommended so set
> + realistic costs compared to memmov_cost. */
> + 3, /* GP2FP */
> + 2, /* FP2GP */
> + 2 /* FP2FP */
> +};
> +
> +static const advsimd_vec_cost generic_armv9_a_advsimd_vector_cost =
> +{
> + 2, /* int_stmt_cost */
> + 2, /* fp_stmt_cost */
> + 2, /* ld2_st2_permute_cost */
> + 2, /* ld3_st3_permute_cost */
> + 3, /* ld4_st4_permute_cost */
> + 3, /* permute_cost */
> + 4, /* reduc_i8_cost */
> + 4, /* reduc_i16_cost */
> + 2, /* reduc_i32_cost */
> + 2, /* reduc_i64_cost */
> + 6, /* reduc_f16_cost */
> + 4, /* reduc_f32_cost */
> + 2, /* reduc_f64_cost */
> + 2, /* store_elt_extra_cost */
> + /* This value is just inherited from the Cortex-A57 table. */
> + 8, /* vec_to_scalar_cost */
> + /* This depends very much on what the scalar value is and
> + where it comes from. E.g. some constants take two dependent
> + instructions or a load, while others might be moved from a GPR.
> + 4 seems to be a reasonable compromise in practice. */
> + 4, /* scalar_to_vec_cost */
> + 4, /* align_load_cost */
> + 4, /* unalign_load_cost */
> + /* Although stores have a latency of 2 and compete for the
> + vector pipes, in practice it's better not to model that. */
> + 1, /* unalign_store_cost */
> + 1 /* store_cost */
> +};
> +
> +static const sve_vec_cost generic_armv9_a_sve_vector_cost =
> +{
> + {
> + 2, /* int_stmt_cost */
> + 2, /* fp_stmt_cost */
> + 3, /* ld2_st2_permute_cost */
> + 4, /* ld3_st3_permute_cost */
> + 4, /* ld4_st4_permute_cost */
> + 3, /* permute_cost */
> + /* Theoretically, a reduction involving 15 scalar ADDs could
> + complete in ~5 cycles and would have a cost of 15. [SU]ADDV
> + completes in 11 cycles, so give it a cost of 15 + 6. */
> + 21, /* reduc_i8_cost */
> + /* Likewise for 7 scalar ADDs (~3 cycles) vs. 9: 7 + 6. */
> + 13, /* reduc_i16_cost */
> + /* Likewise for 3 scalar ADDs (~2 cycles) vs. 8: 3 + 6. */
> + 9, /* reduc_i32_cost */
> + /* Likewise for 1 scalar ADD (~1 cycles) vs. 2: 1 + 1. */
> + 2, /* reduc_i64_cost */
> + /* Theoretically, a reduction involving 7 scalar FADDs could
> + complete in ~8 cycles and would have a cost of 14. FADDV
> + completes in 6 cycles, so give it a cost of 14 - 2. */
> + 12, /* reduc_f16_cost */
> + /* Likewise for 3 scalar FADDs (~4 cycles) vs. 4: 6 - 0. */
> + 6, /* reduc_f32_cost */
> + /* Likewise for 1 scalar FADD (~2 cycles) vs. 2: 2 - 0. */
> + 2, /* reduc_f64_cost */
> + 2, /* store_elt_extra_cost */
> + /* This value is just inherited from the Cortex-A57 table. */
> + 8, /* vec_to_scalar_cost */
> + /* See the comment above the Advanced SIMD versions. */
> + 4, /* scalar_to_vec_cost */
> + 4, /* align_load_cost */
> + 4, /* unalign_load_cost */
> + /* Although stores have a latency of 2 and compete for the
> + vector pipes, in practice it's better not to model that. */
> + 1, /* unalign_store_cost */
> + 1 /* store_cost */
> + },
> + 3, /* clast_cost */
> + 10, /* fadda_f16_cost */
> + 6, /* fadda_f32_cost */
> + 4, /* fadda_f64_cost */
> + /* A strided Advanced SIMD x64 load would take two parallel FP loads
> + (8 cycles) plus an insertion (2 cycles). Assume a 64-bit SVE gather
> + is 1 cycle more. The Advanced SIMD version is costed as 2 scalar loads
> + (cost 8) and a vec_construct (cost 2). Add a full vector operation
> + (cost 2) to that, to avoid the difference being lost in rounding.
> +
> + There is no easy comparison between a strided Advanced SIMD x32 load
> + and an SVE 32-bit gather, but cost an SVE 32-bit gather as 1 vector
> + operation more than a 64-bit gather. */
> + 14, /* gather_load_x32_cost */
> + 12, /* gather_load_x64_cost */
> + 3 /* scatter_store_elt_cost */
> +};
> +
> +static const aarch64_scalar_vec_issue_info generic_armv9_a_scalar_issue_info =
> +{
> + 3, /* loads_stores_per_cycle */
> + 2, /* stores_per_cycle */
> + 4, /* general_ops_per_cycle */
> + 0, /* fp_simd_load_general_ops */
> + 1 /* fp_simd_store_general_ops */
> +};
> +
> +static const aarch64_advsimd_vec_issue_info generic_armv9_a_advsimd_issue_info =
> +{
> + {
> + 3, /* loads_stores_per_cycle */
> + 2, /* stores_per_cycle */
> + 2, /* general_ops_per_cycle */
> + 0, /* fp_simd_load_general_ops */
> + 1 /* fp_simd_store_general_ops */
> + },
> + 2, /* ld2_st2_general_ops */
> + 2, /* ld3_st3_general_ops */
> + 3 /* ld4_st4_general_ops */
> +};
> +
> +static const aarch64_sve_vec_issue_info generic_armv9_a_sve_issue_info =
> +{
> + {
> + {
> + 3, /* loads_per_cycle */
> + 2, /* stores_per_cycle */
> + 2, /* general_ops_per_cycle */
> + 0, /* fp_simd_load_general_ops */
> + 1 /* fp_simd_store_general_ops */
> + },
> + 2, /* ld2_st2_general_ops */
> + 3, /* ld3_st3_general_ops */
> + 3 /* ld4_st4_general_ops */
> + },
> + 2, /* pred_ops_per_cycle */
> + 2, /* while_pred_ops */
> + 2, /* int_cmp_pred_ops */
> + 1, /* fp_cmp_pred_ops */
> + 1, /* gather_scatter_pair_general_ops */
> + 1 /* gather_scatter_pair_pred_ops */
> +};
> +
> +static const aarch64_vec_issue_info generic_armv9_a_vec_issue_info =
> +{
> + &generic_armv9_a_scalar_issue_info,
> + &generic_armv9_a_advsimd_issue_info,
> + &generic_armv9_a_sve_issue_info
> +};
> +
> +/* Neoverse N2 costs for vector insn classes. */
> +static const struct cpu_vector_cost generic_armv9_a_vector_cost =
> +{
> + 1, /* scalar_int_stmt_cost */
> + 2, /* scalar_fp_stmt_cost */
> + 4, /* scalar_load_cost */
> + 1, /* scalar_store_cost */
> + 1, /* cond_taken_branch_cost */
> + 1, /* cond_not_taken_branch_cost */
> + &generic_armv9_a_advsimd_vector_cost, /* advsimd */
> + &generic_armv9_a_sve_vector_cost, /* sve */
> + &generic_armv9_a_vec_issue_info /* issue_info */
> +};
> +
> +static const struct tune_params generic_armv9_a_tunings =
> +{
> + &cortexa76_extra_costs,
> + &generic_armv9_a_addrcost_table,
> + &generic_armv9_a_regmove_cost,
> + &generic_armv9_a_vector_cost,
> + &generic_branch_cost,
> + &generic_approx_modes,
> + SVE_SCALABLE, /* sve_width */
> + { 4, /* load_int. */
> + 1, /* store_int. */
> + 6, /* load_fp. */
> + 2, /* store_fp. */
> + 6, /* load_pred. */
> + 1 /* store_pred. */
> + }, /* memmov_cost. */
> + 3, /* issue_rate */
> + (AARCH64_FUSE_AES_AESMC | AARCH64_FUSE_CMP_BRANCH), /* fusible_ops */
> + "32:16", /* function_align. */
> + "4", /* jump_align. */
> + "32:16", /* loop_align. */
> + 2, /* int_reassoc_width. */
> + 4, /* fp_reassoc_width. */
> + 1, /* fma_reassoc_width. */
> + 2, /* vec_reassoc_width. */
> + 2, /* min_div_recip_mul_sf. */
> + 2, /* min_div_recip_mul_df. */
> + 0, /* max_case_values. */
> + tune_params::AUTOPREFETCHER_WEAK, /* autoprefetcher_model. */
> + (AARCH64_EXTRA_TUNE_CHEAP_SHIFT_EXTEND
> + | AARCH64_EXTRA_TUNE_USE_NEW_VECTOR_COSTS
> + | AARCH64_EXTRA_TUNE_MATCHED_VECTOR_THROUGHPUT), /* tune_flags. */
> + &generic_prefetch_tune,
> + AARCH64_LDP_STP_POLICY_ALWAYS, /* ldp_policy_model. */
> + AARCH64_LDP_STP_POLICY_ALWAYS /* stp_policy_model. */
> +};
> +
> +#endif /* GCC_AARCH64_H_GENERIC_ARMV9_A. */
>
>
>
>
next prev parent reply other threads:[~2023-11-16 9:24 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-15 17:06 [PATCH 1/6]AArch64: Refactor costs models to different files Tamar Christina
2023-11-15 17:07 ` [PATCH 2/6]AArch64: Remove special handling of generic cpu Tamar Christina
2023-11-16 9:14 ` Richard Earnshaw
2023-11-15 17:07 ` [PATCH 3/6]AArch64: Add new generic-armv8-a CPU and make it the default Tamar Christina
2023-11-16 9:23 ` Richard Earnshaw
2023-11-15 17:08 ` [PATCH 4/6]AArch64: Add new generic-armv9-a CPU and make it the default for Armv9 Tamar Christina
2023-11-16 9:23 ` Richard Earnshaw [this message]
2023-11-15 17:08 ` [PATCH 6/6]AArch64: only emit mismatch error when features would be disabled Tamar Christina
2023-11-16 9:26 ` Richard Earnshaw
2023-11-16 9:33 ` Tamar Christina
2023-11-16 9:41 ` Richard Earnshaw
2023-11-16 9:50 ` Tamar Christina
2023-11-16 10:33 ` Richard Earnshaw
2023-11-16 9:13 ` [PATCH 1/6]AArch64: Refactor costs models to different files Richard Earnshaw
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ff8b3f13-9b35-46b2-9722-9c2b2e0dcaac@foss.arm.com \
--to=richard.earnshaw@foss.arm.com \
--cc=Kyrylo.Tkachov@arm.com \
--cc=Marcus.Shawcroft@arm.com \
--cc=Richard.Earnshaw@arm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=nd@arm.com \
--cc=richard.sandiford@arm.com \
--cc=tamar.christina@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).