From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1075) id B68013858D32; Sun, 29 Jan 2023 03:25:38 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B68013858D32 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1674962738; bh=/K45u1ZtQ7w9F+sV/d71F8vg9DYsK/NOOIXEH4d49YM=; h=From:To:Subject:Date:From; b=t1MKu6jhy4BiwbzNBf5vPGgKU4ANtNswq+m9K6/pJ+TCUrN5/SB4xDN52PaYPjl6Y 1N5T7lKWPW7kunHy7/QQatce2aBzZkRh4x2RdqGH5m47iA4fIjKeHobL1xZRfaneee tcUc26FQ16WbkOpXC53gjns2V+RBoAuZvn2lmVdQ= MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="utf-8" From: Jan Hubicka To: gcc-cvs@gcc.gnu.org Subject: [gcc r12-9087] Disable gather/scatter for zen4 X-Act-Checkin: gcc X-Git-Author: Jan Hubicka X-Git-Refname: refs/heads/releases/gcc-12 X-Git-Oldrev: 489c81db7d4f75894e9d34aa90fe7224cfafb53a X-Git-Newrev: 7790d4b2e5c6ed0d4957e3b7948e24023447fbfd Message-Id: <20230129032538.B68013858D32@sourceware.org> Date: Sun, 29 Jan 2023 03:25:38 +0000 (GMT) List-Id: https://gcc.gnu.org/g:7790d4b2e5c6ed0d4957e3b7948e24023447fbfd commit r12-9087-g7790d4b2e5c6ed0d4957e3b7948e24023447fbfd Author: Jan Hubicka Date: Mon Jan 16 15:40:45 2023 +0100 Disable gather/scatter for zen4 this patch adds more tunes for zen4: - new tunes for avx512 scater instructions. In micro benchmarks these seems consistent loss compared to open-coded coe - disable use of gather for zen4 While these are win for a micro benchmarks (based on TSVC), enabling gather is a loss for parest. So for now it seems safe to keep it off. - disable pass to avoid FMA chains for znver4 since fmadd was optimized and does not seem to cause regressions. * config/i386/i386.cc (ix86_vectorize_builtin_scatter): Guard scatter by TARGET_USE_SCATTER. * config/i386/i386.h (TARGET_USE_SCATTER_2PARTS, TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New macros. * config/i386/x86-tune.def (TARGET_USE_SCATTER_2PARTS, TARGET_USE_SCATTER_4PARTS, TARGET_USE_SCATTER): New tunes. (X86_TUNE_AVOID_256FMA_CHAINS, X86_TUNE_AVOID_512FMA_CHAINS): Disable for znver4. (X86_TUNE_USE_GATHER): Disable for zen4. (cherry picked from commit 967592488c64a86f37bef3dabebb56364f14acdd) Diff: --- gcc/config/i386/i386.cc | 7 +++++++ gcc/config/i386/i386.h | 6 ++++++ gcc/config/i386/x86-tune.def | 23 +++++++++++++++++++---- 3 files changed, 32 insertions(+), 4 deletions(-) diff --git a/gcc/config/i386/i386.cc b/gcc/config/i386/i386.cc index ad37f84fe06..962f8c82b48 100644 --- a/gcc/config/i386/i386.cc +++ b/gcc/config/i386/i386.cc @@ -18918,6 +18918,13 @@ ix86_vectorize_builtin_scatter (const_tree vectype, if (!TARGET_AVX512F) return NULL_TREE; + if (known_eq (TYPE_VECTOR_SUBPARTS (vectype), 2u) + ? !TARGET_USE_SCATTER_2PARTS + : (known_eq (TYPE_VECTOR_SUBPARTS (vectype), 4u) + ? !TARGET_USE_SCATTER_4PARTS + : !TARGET_USE_SCATTER)) + return NULL_TREE; + if ((TREE_CODE (index_type) != INTEGER_TYPE && !POINTER_TYPE_P (index_type)) || (TYPE_MODE (index_type) != SImode diff --git a/gcc/config/i386/i386.h b/gcc/config/i386/i386.h index 7a079072e19..fce0b3564a8 100644 --- a/gcc/config/i386/i386.h +++ b/gcc/config/i386/i386.h @@ -392,10 +392,16 @@ extern unsigned char ix86_tune_features[X86_TUNE_LAST]; ix86_tune_features[X86_TUNE_AVOID_4BYTE_PREFIXES] #define TARGET_USE_GATHER_2PARTS \ ix86_tune_features[X86_TUNE_USE_GATHER_2PARTS] +#define TARGET_USE_SCATTER_2PARTS \ + ix86_tune_features[X86_TUNE_USE_SCATTER_2PARTS] #define TARGET_USE_GATHER_4PARTS \ ix86_tune_features[X86_TUNE_USE_GATHER_4PARTS] +#define TARGET_USE_SCATTER_4PARTS \ + ix86_tune_features[X86_TUNE_USE_SCATTER_4PARTS] #define TARGET_USE_GATHER \ ix86_tune_features[X86_TUNE_USE_GATHER] +#define TARGET_USE_SCATTER \ + ix86_tune_features[X86_TUNE_USE_SCATTER] #define TARGET_FUSE_CMP_AND_BRANCH_32 \ ix86_tune_features[X86_TUNE_FUSE_CMP_AND_BRANCH_32] #define TARGET_FUSE_CMP_AND_BRANCH_64 \ diff --git a/gcc/config/i386/x86-tune.def b/gcc/config/i386/x86-tune.def index eb3ab800264..8c3c1b41e79 100644 --- a/gcc/config/i386/x86-tune.def +++ b/gcc/config/i386/x86-tune.def @@ -469,28 +469,43 @@ DEF_TUNE (X86_TUNE_AVOID_4BYTE_PREFIXES, "avoid_4byte_prefixes", DEF_TUNE (X86_TUNE_USE_GATHER_2PARTS, "use_gather_2parts", ~(m_ZNVER1 | m_ZNVER2 | m_ZNVER3 | m_ZNVER4 | m_ALDERLAKE | m_GENERIC)) +/* X86_TUNE_USE_SCATTER_2PARTS: Use scater instructions for vectors with 2 + elements. */ +DEF_TUNE (X86_TUNE_USE_SCATTER_2PARTS, "use_scatter_2parts", + ~(m_ZNVER4 | m_GENERIC)) + /* X86_TUNE_USE_GATHER_4PARTS: Use gather instructions for vectors with 4 elements. */ DEF_TUNE (X86_TUNE_USE_GATHER_4PARTS, "use_gather_4parts", ~(m_ZNVER1 | m_ZNVER2 | m_ZNVER3 | m_ZNVER4 | m_ALDERLAKE | m_GENERIC)) +/* X86_TUNE_USE_SCATTER_4PARTS: Use scater instructions for vectors with 4 + elements. */ +DEF_TUNE (X86_TUNE_USE_SCATTER_4PARTS, "use_scatter_4parts", + ~(m_ZNVER4 | m_GENERIC)) + /* X86_TUNE_USE_GATHER: Use gather instructions for vectors with 8 or more elements. */ DEF_TUNE (X86_TUNE_USE_GATHER, "use_gather", - ~(m_ZNVER1 | m_ZNVER2 | m_ALDERLAKE | m_GENERIC)) + ~(m_ZNVER1 | m_ZNVER2 | m_ZNVER4 | m_ALDERLAKE | m_GENERIC)) + +/* X86_TUNE_USE_SCATTER: Use scater instructions for vectors with 8 or more + elements. */ +DEF_TUNE (X86_TUNE_USE_SCATTER, "use_scatter", + ~(m_ZNVER4)) /* X86_TUNE_AVOID_128FMA_CHAINS: Avoid creating loops with tight 128bit or smaller FMA chain. */ -DEF_TUNE (X86_TUNE_AVOID_128FMA_CHAINS, "avoid_fma_chains", m_ZNVER) +DEF_TUNE (X86_TUNE_AVOID_128FMA_CHAINS, "avoid_fma_chains", m_ZNVER1 | m_ZNVER2 | m_ZNVER3) /* X86_TUNE_AVOID_256FMA_CHAINS: Avoid creating loops with tight 256bit or smaller FMA chain. */ -DEF_TUNE (X86_TUNE_AVOID_256FMA_CHAINS, "avoid_fma256_chains", m_ZNVER2 | m_ZNVER3 | m_ZNVER4 +DEF_TUNE (X86_TUNE_AVOID_256FMA_CHAINS, "avoid_fma256_chains", m_ZNVER2 | m_ZNVER3 | m_ALDERLAKE | m_SAPPHIRERAPIDS) /* X86_TUNE_AVOID_512FMA_CHAINS: Avoid creating loops with tight 512bit or smaller FMA chain. */ -DEF_TUNE (X86_TUNE_AVOID_512FMA_CHAINS, "avoid_fma512_chains", m_ZNVER4) +DEF_TUNE (X86_TUNE_AVOID_512FMA_CHAINS, "avoid_fma512_chains", m_NONE) /* X86_TUNE_V2DF_REDUCTION_PREFER_PHADDPD: Prefer haddpd for v2df vector reduction. */