From patchwork Tue Aug 8 14:07:08 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Joe Ramsay X-Patchwork-Id: 1818626 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@legolas.ozlabs.org Authentication-Results: legolas.ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=sourceware.org (client-ip=2620:52:3:1:0:246e:9693:128c; helo=server2.sourceware.org; envelope-from=libc-alpha-bounces+incoming=patchwork.ozlabs.org@sourceware.org; receiver=) Authentication-Results: legolas.ozlabs.org; dkim=pass (1024-bit key; secure) header.d=sourceware.org header.i=@sourceware.org header.a=rsa-sha256 header.s=default header.b=fjHOVIHA; dkim-atps=neutral Received: from server2.sourceware.org (server2.sourceware.org [IPv6:2620:52:3:1:0:246e:9693:128c]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (P-384) server-digest SHA384) (No client certificate requested) by legolas.ozlabs.org (Postfix) with ESMTPS id 4RKw6443Szz1yYC for ; Wed, 9 Aug 2023 00:09:00 +1000 (AEST) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id C98B338708B0 for ; Tue, 8 Aug 2023 14:08:57 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org C98B338708B0 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1691503737; bh=42GNK/N+OBxV2LnL9hFp3nUoobjjPg9CvSjWdyd7SXw=; h=To:CC:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=fjHOVIHAitqUVGBsp9PXC8v2mgjuKTPpsQkfF/eZ/pViDDtPL7lN+kgODqW8Rt0qa tLW7/YRX8VVozalbDzD5lu5oqbJStDZL8cGAsrybMuu4n8aMA1IC8zkmGxzJWnApwf ud+jK3kc/NAG8iBzEaZPwiA37ZFxpSpY3Z+A6YiA= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from EUR03-DBA-obe.outbound.protection.outlook.com (mail-dbaeur03on2050.outbound.protection.outlook.com [40.107.104.50]) by sourceware.org (Postfix) with ESMTPS id 58DE93858D20 for ; Tue, 8 Aug 2023 14:07:30 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 58DE93858D20 Received: from AM5PR1001CA0035.EURPRD10.PROD.OUTLOOK.COM (2603:10a6:206:2::48) by PAXPR08MB6685.eurprd08.prod.outlook.com (2603:10a6:102:133::18) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6652.26; Tue, 8 Aug 2023 14:07:25 +0000 Received: from AM7EUR03FT045.eop-EUR03.prod.protection.outlook.com (2603:10a6:206:2:cafe::96) by AM5PR1001CA0035.outlook.office365.com (2603:10a6:206:2::48) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6652.27 via Frontend Transport; Tue, 8 Aug 2023 14:07:25 +0000 X-MS-Exchange-Authentication-Results: spf=pass (sender IP is 63.35.35.123) smtp.mailfrom=arm.com; dkim=pass (signature was verified) header.d=armh.onmicrosoft.com;dmarc=pass action=none header.from=arm.com; Received-SPF: Pass (protection.outlook.com: domain of arm.com designates 63.35.35.123 as permitted sender) receiver=protection.outlook.com; client-ip=63.35.35.123; helo=64aa7808-outbound-1.mta.getcheckrecipient.com; pr=C Received: from 64aa7808-outbound-1.mta.getcheckrecipient.com (63.35.35.123) by AM7EUR03FT045.mail.protection.outlook.com (100.127.140.150) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6678.17 via Frontend Transport; Tue, 8 Aug 2023 14:07:25 +0000 Received: ("Tessian outbound 997ae1cc9f47:v145"); Tue, 08 Aug 2023 14:07:25 +0000 X-CheckRecipientChecked: true X-CR-MTA-CID: 6cc73ee0d74e831d X-CR-MTA-TID: 64aa7808 Received: from 340ca0f60534.1 by 64aa7808-outbound-1.mta.getcheckrecipient.com id 03B608B9-151C-47FE-93EC-B6C065AEE753.1; Tue, 08 Aug 2023 14:07:18 +0000 Received: from EUR05-AM6-obe.outbound.protection.outlook.com by 64aa7808-outbound-1.mta.getcheckrecipient.com with ESMTPS id 340ca0f60534.1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384); Tue, 08 Aug 2023 14:07:18 +0000 ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=aEw9KQrFQOtsLV0Y4lp8Vld2CMz5z3MoroMrQYG4/q2REwTXNiQA2rkSo2CF4fIfuYGiZV3zfDtyA78o7E2TIaJNJMsFOAyJyj3y9+9h1yNMhrT1xgsVFya8dxO4CwxhQUw6oOptNPJJXUxPrAcANInwNqBFt2ZOYuVBSZBawGdY6EeqNPbuSGIm6OkBw5au/FDDjaEj71t2mkl1lTbrvKW3gWrfg6V9mGV41b/EFyDvcS4moxYEnws94IbYFgMeVAtHiKIJbZnWneTMu6Jg6xWEUtcWxHot5D0+iiiwvMtjJBSAEy/OXMwozwbx0AEjW3XgbOqnAHduymBWmgoeRg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=42GNK/N+OBxV2LnL9hFp3nUoobjjPg9CvSjWdyd7SXw=; b=FHZqklUyZrIC3rKS5/+CZtdfsZ+nayygetbXkehlQOBMI0ZBtJ2CNcBzOKclG5WxzUW73+SoboWghLuH5KXWxLNcCcL6LKvJN21+sXpUD+Yyb7Blfr9uveiztaMIqjd337cn/OBukoIcJ1acLwSGNWL/E3OM0zejzq/CHzK55SB7gAcmDsVOcQtQPv5nTMBnJEkV6whruJvpCEmPxPsFwwvM5HlmT/DTL0k4loPuuWrsnEPCt8CqxnMzx9IYDtTV92Ue/KhGSqZO1fJlvdEk6mpjWboYW0AUe/hj01pLuX6KpoW7fFXUm0jYHl4WBp1KH/vLpWuSVcXq/V9L/Ck9SA== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass (sender ip is 40.67.248.234) smtp.rcpttodomain=sourceware.org smtp.mailfrom=arm.com; dmarc=pass (p=none sp=none pct=100) action=none header.from=arm.com; dkim=none (message not signed); arc=none Received: from AM0PR03CA0002.eurprd03.prod.outlook.com (2603:10a6:208:14::15) by AS8PR08MB9790.eurprd08.prod.outlook.com (2603:10a6:20b:615::21) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6652.26; Tue, 8 Aug 2023 14:07:14 +0000 Received: from AM7EUR03FT044.eop-EUR03.prod.protection.outlook.com (2603:10a6:208:14:cafe::69) by AM0PR03CA0002.outlook.office365.com (2603:10a6:208:14::15) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6652.27 via Frontend Transport; Tue, 8 Aug 2023 14:07:14 +0000 X-MS-Exchange-Authentication-Results: spf=pass (sender IP is 40.67.248.234) smtp.mailfrom=arm.com; dkim=none (message not signed) header.d=none;dmarc=pass action=none header.from=arm.com; Received-SPF: Pass (protection.outlook.com: domain of arm.com designates 40.67.248.234 as permitted sender) receiver=protection.outlook.com; client-ip=40.67.248.234; helo=nebula.arm.com; pr=C Received: from nebula.arm.com (40.67.248.234) by AM7EUR03FT044.mail.protection.outlook.com (100.127.140.169) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.20.6678.17 via Frontend Transport; Tue, 8 Aug 2023 14:07:14 +0000 Received: from AZ-NEU-EX04.Arm.com (10.251.24.32) by AZ-NEU-EX03.Arm.com (10.251.24.31) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2507.27; Tue, 8 Aug 2023 14:07:11 +0000 Received: from vcn-man-apps.manchester.arm.com (10.32.108.22) by mail.arm.com (10.251.24.32) with Microsoft SMTP Server id 15.1.2507.27 via Frontend Transport; Tue, 8 Aug 2023 14:07:11 +0000 To: CC: Joe Ramsay Subject: [PATCH 4/5 v2] aarch64: Add vector implementations of log10 routines Date: Tue, 8 Aug 2023 15:07:08 +0100 Message-ID: <20230808140709.47733-4-Joe.Ramsay@arm.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20230808140709.47733-1-Joe.Ramsay@arm.com> References: <20230808140709.47733-1-Joe.Ramsay@arm.com> MIME-Version: 1.0 X-EOPAttributedMessage: 1 X-MS-TrafficTypeDiagnostic: AM7EUR03FT044:EE_|AS8PR08MB9790:EE_|AM7EUR03FT045:EE_|PAXPR08MB6685:EE_ X-MS-Office365-Filtering-Correlation-Id: c989cb5e-8e37-4886-6e2f-08db9818cb31 x-checkrecipientrouted: true NoDisclaimer: true X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam-Untrusted: BCL:0; X-Microsoft-Antispam-Message-Info-Original: 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 X-Forefront-Antispam-Report-Untrusted: CIP:40.67.248.234; CTRY:IE; LANG:en; SCL:1; SRV:; IPV:NLI; SFV:NSPM; H:nebula.arm.com; PTR:InfoDomainNonexistent; CAT:NONE; SFS:(13230028)(4636009)(376002)(39860400002)(136003)(346002)(396003)(1800799003)(186006)(451199021)(82310400008)(46966006)(36840700001)(2616005)(47076005)(82740400003)(81166007)(478600001)(356005)(336012)(36860700001)(83380400001)(426003)(40480700001)(26005)(1076003)(41300700001)(70586007)(316002)(8676002)(8936002)(70206006)(4326008)(30864003)(2906002)(6916009)(7696005)(6666004)(5660300002)(86362001)(36756003)(2004002)(36900700001)(473944003); DIR:OUT; SFP:1101; X-MS-Exchange-Transport-CrossTenantHeadersStamped: AS8PR08MB9790 X-MS-Exchange-Transport-CrossTenantHeadersStripped: AM7EUR03FT045.eop-EUR03.prod.protection.outlook.com X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id-Prvs: e1f2f4eb-eca1-4728-bb13-08db9818c486 X-Microsoft-Antispam: BCL:0; X-Microsoft-Antispam-Message-Info: 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 X-Forefront-Antispam-Report: CIP:63.35.35.123; CTRY:IE; LANG:en; SCL:1; SRV:; IPV:CAL; SFV:NSPM; H:64aa7808-outbound-1.mta.getcheckrecipient.com; PTR:ec2-63-35-35-123.eu-west-1.compute.amazonaws.com; CAT:NONE; SFS:(13230028)(4636009)(396003)(376002)(39860400002)(346002)(136003)(82310400008)(451199021)(186006)(1800799003)(46966006)(36840700001)(40470700004)(86362001)(40480700001)(41300700001)(336012)(478600001)(40460700003)(1076003)(8676002)(8936002)(26005)(7696005)(6666004)(2616005)(426003)(5660300002)(47076005)(36756003)(2906002)(30864003)(83380400001)(36860700001)(316002)(70206006)(6916009)(70586007)(82740400003)(81166007)(4326008)(2004002)(473944003); DIR:OUT; SFP:1101; X-OriginatorOrg: arm.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 08 Aug 2023 14:07:25.7927 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: c989cb5e-8e37-4886-6e2f-08db9818cb31 X-MS-Exchange-CrossTenant-Id: f34e5979-57d9-4aaa-ad4d-b122a662184d X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=f34e5979-57d9-4aaa-ad4d-b122a662184d; Ip=[63.35.35.123]; Helo=[64aa7808-outbound-1.mta.getcheckrecipient.com] X-MS-Exchange-CrossTenant-AuthSource: AM7EUR03FT045.eop-EUR03.prod.protection.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: PAXPR08MB6685 X-Spam-Status: No, score=-12.9 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, GIT_PATCH_0, KAM_DMARC_NONE, KAM_SHORT, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H2, SPF_HELO_NONE, SPF_NONE, TXREP, UNPARSEABLE_RELAY autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Joe Ramsay via Libc-alpha From: Joe Ramsay Reply-To: Joe Ramsay Errors-To: libc-alpha-bounces+incoming=patchwork.ozlabs.org@sourceware.org Sender: "Libc-alpha" A table is also added, which is shared between AdvSIMD and SVE log10. --- Changes to v1: Fix check-abi Thanks, Joe sysdeps/aarch64/fpu/Makefile | 4 +- sysdeps/aarch64/fpu/Versions | 4 + sysdeps/aarch64/fpu/bits/math-vector.h | 4 + sysdeps/aarch64/fpu/log10_advsimd.c | 117 ++++++++++++ sysdeps/aarch64/fpu/log10_sve.c | 77 ++++++++ sysdeps/aarch64/fpu/log10f_advsimd.c | 79 ++++++++ sysdeps/aarch64/fpu/log10f_sve.c | 94 ++++++++++ .../fpu/test-double-advsimd-wrappers.c | 1 + .../aarch64/fpu/test-double-sve-wrappers.c | 1 + .../aarch64/fpu/test-float-advsimd-wrappers.c | 1 + sysdeps/aarch64/fpu/test-float-sve-wrappers.c | 1 + sysdeps/aarch64/fpu/v_log10_data.c | 175 ++++++++++++++++++ sysdeps/aarch64/fpu/vecmath_config.h | 9 + sysdeps/aarch64/libm-test-ulps | 8 + .../unix/sysv/linux/aarch64/libmvec.abilist | 4 + 15 files changed, 578 insertions(+), 1 deletion(-) create mode 100644 sysdeps/aarch64/fpu/log10_advsimd.c create mode 100644 sysdeps/aarch64/fpu/log10_sve.c create mode 100644 sysdeps/aarch64/fpu/log10f_advsimd.c create mode 100644 sysdeps/aarch64/fpu/log10f_sve.c create mode 100644 sysdeps/aarch64/fpu/v_log10_data.c diff --git a/sysdeps/aarch64/fpu/Makefile b/sysdeps/aarch64/fpu/Makefile index c3f204ff0d..0a047a150d 100644 --- a/sysdeps/aarch64/fpu/Makefile +++ b/sysdeps/aarch64/fpu/Makefile @@ -2,6 +2,7 @@ libmvec-supported-funcs = cos \ exp \ exp2 \ log \ + log10 \ log2 \ sin \ tan @@ -18,7 +19,8 @@ libmvec-support = $(addsuffix f_advsimd,$(float-advsimd-funcs)) \ $(addsuffix _sve,$(double-sve-funcs)) \ v_log_data \ v_exp_data \ - v_log2_data + v_log2_data \ + v_log10_data endif sve-cflags = -march=armv8-a+sve diff --git a/sysdeps/aarch64/fpu/Versions b/sysdeps/aarch64/fpu/Versions index ffe62a6f65..358efed5ee 100644 --- a/sysdeps/aarch64/fpu/Versions +++ b/sysdeps/aarch64/fpu/Versions @@ -22,6 +22,10 @@ libmvec { _ZGVnN2v_exp2; _ZGVsMxv_exp2f; _ZGVsMxv_exp2; + _ZGVnN4v_log10f; + _ZGVnN2v_log10; + _ZGVsMxv_log10f; + _ZGVsMxv_log10; _ZGVnN4v_log2f; _ZGVnN2v_log2; _ZGVsMxv_log2f; diff --git a/sysdeps/aarch64/fpu/bits/math-vector.h b/sysdeps/aarch64/fpu/bits/math-vector.h index 92f214b194..59f2efa6d7 100644 --- a/sysdeps/aarch64/fpu/bits/math-vector.h +++ b/sysdeps/aarch64/fpu/bits/math-vector.h @@ -53,6 +53,7 @@ __vpcs __f32x4_t _ZGVnN4v_cosf (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_expf (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_exp2f (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_logf (__f32x4_t); +__vpcs __f32x4_t _ZGVnN4v_log10f (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_log2f (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_sinf (__f32x4_t); __vpcs __f32x4_t _ZGVnN4v_tanf (__f32x4_t); @@ -61,6 +62,7 @@ __vpcs __f64x2_t _ZGVnN2v_cos (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_exp (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_exp2 (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_log (__f64x2_t); +__vpcs __f64x2_t _ZGVnN2v_log10 (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_log2 (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_sin (__f64x2_t); __vpcs __f64x2_t _ZGVnN2v_tan (__f64x2_t); @@ -74,6 +76,7 @@ __sv_f32_t _ZGVsMxv_cosf (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_expf (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_exp2f (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_logf (__sv_f32_t, __sv_bool_t); +__sv_f32_t _ZGVsMxv_log10f (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_log2f (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_sinf (__sv_f32_t, __sv_bool_t); __sv_f32_t _ZGVsMxv_tanf (__sv_f32_t, __sv_bool_t); @@ -82,6 +85,7 @@ __sv_f64_t _ZGVsMxv_cos (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_exp (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_exp2 (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_log (__sv_f64_t, __sv_bool_t); +__sv_f64_t _ZGVsMxv_log10 (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_log2 (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_sin (__sv_f64_t, __sv_bool_t); __sv_f64_t _ZGVsMxv_tan (__sv_f64_t, __sv_bool_t); diff --git a/sysdeps/aarch64/fpu/log10_advsimd.c b/sysdeps/aarch64/fpu/log10_advsimd.c new file mode 100644 index 0000000000..f518b55b60 --- /dev/null +++ b/sysdeps/aarch64/fpu/log10_advsimd.c @@ -0,0 +1,117 @@ +/* Double-precision vector (AdvSIMD) log10 function + + Copyright (C) 2023 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include "v_math.h" +#include "poly_advsimd_f64.h" + +#define N (1 << V_LOG10_TABLE_BITS) + +static const struct data +{ + float64x2_t poly[5]; + float64x2_t invln10, log10_2, ln2; + uint64x2_t min_norm, special_bound, sign_exp_mask; +} data = { + /* Computed from log coefficients divided by log(10) then rounded to double + precision. */ + .poly = { V2 (-0x1.bcb7b1526e506p-3), V2 (0x1.287a7636be1d1p-3), + V2 (-0x1.bcb7b158af938p-4), V2 (0x1.63c78734e6d07p-4), + V2 (-0x1.287461742fee4p-4) }, + .ln2 = V2 (0x1.62e42fefa39efp-1), + .invln10 = V2 (0x1.bcb7b1526e50ep-2), + .log10_2 = V2 (0x1.34413509f79ffp-2), + .min_norm = V2 (0x0010000000000000), /* asuint64(0x1p-1022). */ + .special_bound = V2 (0x7fe0000000000000), /* asuint64(inf) - min_norm. */ + .sign_exp_mask = V2 (0xfff0000000000000), +}; + +#define Off v_u64 (0x3fe6900900000000) +#define IndexMask v_u64 (N - 1) + +#define T(s, i) __v_log10_data.s[i] + +struct entry +{ + float64x2_t invc; + float64x2_t log10c; +}; + +static inline struct entry +lookup (uint64x2_t i) +{ + struct entry e; + e.invc[0] = T (invc, i[0]); + e.log10c[0] = T (log10c, i[0]); + e.invc[1] = T (invc, i[1]); + e.log10c[1] = T (log10c, i[1]); + return e; +} + +static float64x2_t VPCS_ATTR NOINLINE +special_case (float64x2_t x, float64x2_t y, uint64x2_t special) +{ + return v_call_f64 (log10, x, y, special); +} + +/* Fast implementation of double-precision vector log10 + is a slight modification of double-precision vector log. + Max ULP error: < 2.5 ulp (nearest rounding.) + Maximum measured at 2.46 ulp for x in [0.96, 0.97] + _ZGVnN2v_log10(0x1.13192407fcb46p+0) got 0x1.fff6be3cae4bbp-6 + want 0x1.fff6be3cae4b9p-6. */ +float64x2_t VPCS_ATTR V_NAME_D1 (log10) (float64x2_t x) +{ + const struct data *d = ptr_barrier (&data); + uint64x2_t ix = vreinterpretq_u64_f64 (x); + uint64x2_t special + = vcgeq_u64 (vsubq_u64 (ix, d->min_norm), d->special_bound); + + /* x = 2^k z; where z is in range [OFF,2*OFF) and exact. + The range is split into N subintervals. + The ith subinterval contains z and c is near its center. */ + uint64x2_t tmp = vsubq_u64 (ix, Off); + uint64x2_t i + = vandq_u64 (vshrq_n_u64 (tmp, 52 - V_LOG10_TABLE_BITS), IndexMask); + int64x2_t k = vshrq_n_s64 (vreinterpretq_s64_u64 (tmp), 52); + uint64x2_t iz = vsubq_u64 (ix, vandq_u64 (tmp, d->sign_exp_mask)); + float64x2_t z = vreinterpretq_f64_u64 (iz); + + struct entry e = lookup (i); + + /* log10(x) = log1p(z/c-1)/log(10) + log10(c) + k*log10(2). */ + float64x2_t r = vfmaq_f64 (v_f64 (-1.0), z, e.invc); + float64x2_t kd = vcvtq_f64_s64 (k); + + /* hi = r / log(10) + log10(c) + k*log10(2). + Constants in v_log10_data.c are computed (in extended precision) as + e.log10c := e.logc * ivln10. */ + float64x2_t w = vfmaq_f64 (e.log10c, r, d->invln10); + + /* y = log10(1+r) + n * log10(2). */ + float64x2_t hi = vfmaq_f64 (w, kd, d->log10_2); + + /* y = r2*(A0 + r*A1 + r2*(A2 + r*A3 + r2*A4)) + hi. */ + float64x2_t r2 = vmulq_f64 (r, r); + float64x2_t y = v_pw_horner_4_f64 (r, r2, d->poly); + y = vfmaq_f64 (hi, r2, y); + + if (__glibc_unlikely (v_any_u64 (special))) + return special_case (x, y, special); + return y; +} diff --git a/sysdeps/aarch64/fpu/log10_sve.c b/sysdeps/aarch64/fpu/log10_sve.c new file mode 100644 index 0000000000..4522879f99 --- /dev/null +++ b/sysdeps/aarch64/fpu/log10_sve.c @@ -0,0 +1,77 @@ +/* Double-precision vector (SVE) log10 function + + Copyright (C) 2023 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include "sv_math.h" +#include "poly_sve_f64.h" + +#define Min 0x0010000000000000 +#define Max 0x7ff0000000000000 +#define Thres 0x7fe0000000000000 /* Max - Min. */ +#define Off 0x3fe6900900000000 +#define N (1 << V_LOG10_TABLE_BITS) + +#define T(s, i) __v_log10_data.s[i] + +static svfloat64_t NOINLINE +special_case (svfloat64_t x, svfloat64_t y, svbool_t special) +{ + return sv_call_f64 (log10, x, y, special); +} + +/* SVE log10 algorithm. + Maximum measured error is 2.46 ulps. + SV_NAME_D1 (log10)(0x1.131956cd4b627p+0) got 0x1.fffbdf6eaa669p-6 + want 0x1.fffbdf6eaa667p-6. */ +svfloat64_t SV_NAME_D1 (log10) (svfloat64_t x, const svbool_t pg) +{ + svuint64_t ix = svreinterpret_u64 (x); + svbool_t special = svcmpge (pg, svsub_x (pg, ix, Min), Thres); + + /* x = 2^k z; where z is in range [Off,2*Off) and exact. + The range is split into N subintervals. + The ith subinterval contains z and c is near its center. */ + svuint64_t tmp = svsub_x (pg, ix, Off); + svuint64_t i + = sv_mod_n_u64_x (pg, svlsr_x (pg, tmp, 52 - V_LOG10_TABLE_BITS), N); + svfloat64_t k = svcvt_f64_x (pg, svasr_x (pg, svreinterpret_s64 (tmp), 52)); + svfloat64_t z = svreinterpret_f64 ( + svsub_x (pg, ix, svand_x (pg, tmp, 0xfffULL << 52))); + + /* log(x) = k*log(2) + log(c) + log(z/c). */ + svfloat64_t invc = svld1_gather_index (pg, &T (invc, 0), i); + svfloat64_t logc = svld1_gather_index (pg, &T (log10c, 0), i); + + /* We approximate log(z/c) with a polynomial P(x) ~= log(x + 1): + r = z/c - 1 (we look up precomputed 1/c) + log(z/c) ~= P(r). */ + svfloat64_t r = svmad_x (pg, invc, z, -1.0); + + /* hi = log(c) + k*log(2). */ + svfloat64_t w = svmla_x (pg, logc, r, __v_log10_data.invln10); + svfloat64_t hi = svmla_x (pg, w, k, __v_log10_data.log10_2); + + /* y = r2*(A0 + r*A1 + r2*(A2 + r*A3 + r2*A4)) + hi. */ + svfloat64_t r2 = svmul_x (pg, r, r); + svfloat64_t y = sv_pw_horner_4_f64_x (pg, r, r2, __v_log10_data.poly); + y = svmla_x (pg, hi, r2, y); + + if (__glibc_unlikely (svptest_any (pg, special))) + return special_case (x, y, special); + return y; +} diff --git a/sysdeps/aarch64/fpu/log10f_advsimd.c b/sysdeps/aarch64/fpu/log10f_advsimd.c new file mode 100644 index 0000000000..371495b541 --- /dev/null +++ b/sysdeps/aarch64/fpu/log10f_advsimd.c @@ -0,0 +1,79 @@ +/* Single-precision vector (AdvSIMD) log10 function + + Copyright (C) 2023 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include "v_math.h" +#include "poly_advsimd_f32.h" + +static const struct data +{ + float32x4_t poly[8]; + float32x4_t inv_ln10, ln2; + uint32x4_t min_norm, special_bound, off, mantissa_mask; +} data = { + /* Use order 9 for log10(1+x), i.e. order 8 for log10(1+x)/x, with x in + [-1/3, 1/3] (offset=2/3). Max. relative error: 0x1.068ee468p-25. */ + .poly = { V4 (-0x1.bcb79cp-3f), V4 (0x1.2879c8p-3f), V4 (-0x1.bcd472p-4f), + V4 (0x1.6408f8p-4f), V4 (-0x1.246f8p-4f), V4 (0x1.f0e514p-5f), + V4 (-0x1.0fc92cp-4f), V4 (0x1.f5f76ap-5f) }, + .ln2 = V4 (0x1.62e43p-1f), + .inv_ln10 = V4 (0x1.bcb7b2p-2f), + .min_norm = V4 (0x00800000), + .special_bound = V4 (0x7f000000), /* asuint32(inf) - min_norm. */ + .off = V4 (0x3f2aaaab), /* 0.666667. */ + .mantissa_mask = V4 (0x007fffff), +}; + +static float32x4_t VPCS_ATTR NOINLINE +special_case (float32x4_t x, float32x4_t y, uint32x4_t cmp) +{ + /* Fall back to scalar code. */ + return v_call_f32 (log10f, x, y, cmp); +} + +/* Fast implementation of AdvSIMD log10f, + uses a similar approach as AdvSIMD logf with the same offset (i.e., 2/3) and + an order 9 polynomial. + Maximum error: 3.305ulps (nearest rounding.) + _ZGVnN4v_log10f(0x1.555c16p+0) got 0x1.ffe2fap-4 + want 0x1.ffe2f4p-4. */ +float32x4_t VPCS_ATTR V_NAME_F1 (log10) (float32x4_t x) +{ + const struct data *d = ptr_barrier (&data); + uint32x4_t u = vreinterpretq_u32_f32 (x); + uint32x4_t special + = vcgeq_u32 (vsubq_u32 (u, d->min_norm), d->special_bound); + + /* x = 2^n * (1+r), where 2/3 < 1+r < 4/3. */ + u = vsubq_u32 (u, d->off); + float32x4_t n = vcvtq_f32_s32 ( + vshrq_n_s32 (vreinterpretq_s32_u32 (u), 23)); /* signextend. */ + u = vaddq_u32 (vandq_u32 (u, d->mantissa_mask), d->off); + float32x4_t r = vsubq_f32 (vreinterpretq_f32_u32 (u), v_f32 (1.0f)); + + /* y = log10(1+r) + n * log10(2). */ + float32x4_t r2 = vmulq_f32 (r, r); + float32x4_t poly = v_pw_horner_7_f32 (r, r2, d->poly); + /* y = Log10(2) * n + poly * InvLn(10). */ + float32x4_t y = vfmaq_f32 (r, d->ln2, n); + y = vfmaq_f32 (vmulq_f32 (y, d->inv_ln10), poly, r2); + + if (__glibc_unlikely (v_any_u32 (special))) + return special_case (x, y, special); + return y; +} diff --git a/sysdeps/aarch64/fpu/log10f_sve.c b/sysdeps/aarch64/fpu/log10f_sve.c new file mode 100644 index 0000000000..22c63bb2fa --- /dev/null +++ b/sysdeps/aarch64/fpu/log10f_sve.c @@ -0,0 +1,94 @@ +/* Single-precision vector (SVE) log10 function + + Copyright (C) 2023 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include "sv_math.h" + +static const struct data +{ + float poly_0246[4]; + float poly_1357[4]; + float ln2, inv_ln10; +} data = { + .poly_1357 = { + /* Coefficients copied from the AdvSIMD routine, then rearranged so that coeffs + 1, 3, 5 and 7 can be loaded as a single quad-word, hence used with _lane + variant of MLA intrinsic. */ + 0x1.2879c8p-3f, 0x1.6408f8p-4f, 0x1.f0e514p-5f, 0x1.f5f76ap-5f + }, + .poly_0246 = { -0x1.bcb79cp-3f, -0x1.bcd472p-4f, -0x1.246f8p-4f, + -0x1.0fc92cp-4f }, + .ln2 = 0x1.62e43p-1f, + .inv_ln10 = 0x1.bcb7b2p-2f, +}; + +#define Min 0x00800000 +#define Max 0x7f800000 +#define Thres 0x7f000000 /* Max - Min. */ +#define Offset 0x3f2aaaab /* 0.666667. */ +#define MantissaMask 0x007fffff + +static svfloat32_t NOINLINE +special_case (svfloat32_t x, svfloat32_t y, svbool_t special) +{ + return sv_call_f32 (log10f, x, y, special); +} + +/* Optimised implementation of SVE log10f using the same algorithm and + polynomial as AdvSIMD log10f. + Maximum error is 3.31ulps: + SV_NAME_F1 (log10)(0x1.555c16p+0) got 0x1.ffe2fap-4 + want 0x1.ffe2f4p-4. */ +svfloat32_t SV_NAME_F1 (log10) (svfloat32_t x, const svbool_t pg) +{ + const struct data *d = ptr_barrier (&data); + svuint32_t ix = svreinterpret_u32 (x); + svbool_t special = svcmpge (pg, svsub_x (pg, ix, Min), Thres); + + /* x = 2^n * (1+r), where 2/3 < 1+r < 4/3. */ + ix = svsub_x (pg, ix, Offset); + svfloat32_t n = svcvt_f32_x ( + pg, svasr_x (pg, svreinterpret_s32 (ix), 23)); /* signextend. */ + ix = svand_x (pg, ix, MantissaMask); + ix = svadd_x (pg, ix, Offset); + svfloat32_t r = svsub_x (pg, svreinterpret_f32 (ix), 1.0f); + + /* y = log10(1+r) + n*log10(2) + log10(1+r) ~ r * InvLn(10) + P(r) + where P(r) is a polynomial. Use order 9 for log10(1+x), i.e. order 8 for + log10(1+x)/x, with x in [-1/3, 1/3] (offset=2/3). */ + svfloat32_t r2 = svmul_x (pg, r, r); + svfloat32_t r4 = svmul_x (pg, r2, r2); + svfloat32_t p_1357 = svld1rq (pg, &d->poly_1357[0]); + svfloat32_t q_01 = svmla_lane (sv_f32 (d->poly_0246[0]), r, p_1357, 0); + svfloat32_t q_23 = svmla_lane (sv_f32 (d->poly_0246[1]), r, p_1357, 1); + svfloat32_t q_45 = svmla_lane (sv_f32 (d->poly_0246[2]), r, p_1357, 2); + svfloat32_t q_67 = svmla_lane (sv_f32 (d->poly_0246[3]), r, p_1357, 3); + svfloat32_t q_47 = svmla_x (pg, q_45, r2, q_67); + svfloat32_t q_03 = svmla_x (pg, q_01, r2, q_23); + svfloat32_t y = svmla_x (pg, q_03, r4, q_47); + + /* Using hi = Log10(2)*n + r*InvLn(10) is faster but less accurate. */ + svfloat32_t hi = svmla_x (pg, r, n, d->ln2); + hi = svmul_x (pg, hi, d->inv_ln10); + y = svmla_x (pg, hi, r2, y); + + if (__glibc_unlikely (svptest_any (pg, special))) + return special_case (x, y, special); + return y; +} diff --git a/sysdeps/aarch64/fpu/test-double-advsimd-wrappers.c b/sysdeps/aarch64/fpu/test-double-advsimd-wrappers.c index d30dcd6f95..8d05498ec9 100644 --- a/sysdeps/aarch64/fpu/test-double-advsimd-wrappers.c +++ b/sysdeps/aarch64/fpu/test-double-advsimd-wrappers.c @@ -27,6 +27,7 @@ VPCS_VECTOR_WRAPPER (cos_advsimd, _ZGVnN2v_cos) VPCS_VECTOR_WRAPPER (exp_advsimd, _ZGVnN2v_exp) VPCS_VECTOR_WRAPPER (exp2_advsimd, _ZGVnN2v_exp2) VPCS_VECTOR_WRAPPER (log_advsimd, _ZGVnN2v_log) +VPCS_VECTOR_WRAPPER (log10_advsimd, _ZGVnN2v_log10) VPCS_VECTOR_WRAPPER (log2_advsimd, _ZGVnN2v_log2) VPCS_VECTOR_WRAPPER (sin_advsimd, _ZGVnN2v_sin) VPCS_VECTOR_WRAPPER (tan_advsimd, _ZGVnN2v_tan) diff --git a/sysdeps/aarch64/fpu/test-double-sve-wrappers.c b/sysdeps/aarch64/fpu/test-double-sve-wrappers.c index 22a8479100..b65bc6f1e6 100644 --- a/sysdeps/aarch64/fpu/test-double-sve-wrappers.c +++ b/sysdeps/aarch64/fpu/test-double-sve-wrappers.c @@ -36,6 +36,7 @@ SVE_VECTOR_WRAPPER (cos_sve, _ZGVsMxv_cos) SVE_VECTOR_WRAPPER (exp_sve, _ZGVsMxv_exp) SVE_VECTOR_WRAPPER (exp2_sve, _ZGVsMxv_exp2) SVE_VECTOR_WRAPPER (log_sve, _ZGVsMxv_log) +SVE_VECTOR_WRAPPER (log10_sve, _ZGVsMxv_log10) SVE_VECTOR_WRAPPER (log2_sve, _ZGVsMxv_log2) SVE_VECTOR_WRAPPER (sin_sve, _ZGVsMxv_sin) SVE_VECTOR_WRAPPER (tan_sve, _ZGVsMxv_tan) diff --git a/sysdeps/aarch64/fpu/test-float-advsimd-wrappers.c b/sysdeps/aarch64/fpu/test-float-advsimd-wrappers.c index e8f7f47c67..6ced0d4488 100644 --- a/sysdeps/aarch64/fpu/test-float-advsimd-wrappers.c +++ b/sysdeps/aarch64/fpu/test-float-advsimd-wrappers.c @@ -27,6 +27,7 @@ VPCS_VECTOR_WRAPPER (cosf_advsimd, _ZGVnN4v_cosf) VPCS_VECTOR_WRAPPER (expf_advsimd, _ZGVnN4v_expf) VPCS_VECTOR_WRAPPER (exp2f_advsimd, _ZGVnN4v_exp2f) VPCS_VECTOR_WRAPPER (logf_advsimd, _ZGVnN4v_logf) +VPCS_VECTOR_WRAPPER (log10f_advsimd, _ZGVnN4v_log10f) VPCS_VECTOR_WRAPPER (log2f_advsimd, _ZGVnN4v_log2f) VPCS_VECTOR_WRAPPER (sinf_advsimd, _ZGVnN4v_sinf) VPCS_VECTOR_WRAPPER (tanf_advsimd, _ZGVnN4v_tanf) diff --git a/sysdeps/aarch64/fpu/test-float-sve-wrappers.c b/sysdeps/aarch64/fpu/test-float-sve-wrappers.c index f5e9584265..2ed8d0659a 100644 --- a/sysdeps/aarch64/fpu/test-float-sve-wrappers.c +++ b/sysdeps/aarch64/fpu/test-float-sve-wrappers.c @@ -36,6 +36,7 @@ SVE_VECTOR_WRAPPER (cosf_sve, _ZGVsMxv_cosf) SVE_VECTOR_WRAPPER (expf_sve, _ZGVsMxv_expf) SVE_VECTOR_WRAPPER (exp2f_sve, _ZGVsMxv_exp2f) SVE_VECTOR_WRAPPER (logf_sve, _ZGVsMxv_logf) +SVE_VECTOR_WRAPPER (log10f_sve, _ZGVsMxv_log10f) SVE_VECTOR_WRAPPER (log2f_sve, _ZGVsMxv_log2f) SVE_VECTOR_WRAPPER (sinf_sve, _ZGVsMxv_sinf) SVE_VECTOR_WRAPPER (tanf_sve, _ZGVsMxv_tanf) diff --git a/sysdeps/aarch64/fpu/v_log10_data.c b/sysdeps/aarch64/fpu/v_log10_data.c new file mode 100644 index 0000000000..b0b07bafe7 --- /dev/null +++ b/sysdeps/aarch64/fpu/v_log10_data.c @@ -0,0 +1,175 @@ +/* Lookup table for double-precision log10(x) vector function. + + Copyright (C) 2023 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include "vecmath_config.h" + +const struct v_log10_data __v_log10_data = { + /* Computed from log's coefficients div by log(10) then rounded to double + precision. */ + .poly = { -0x1.bcb7b1526e506p-3, 0x1.287a7636be1d1p-3, -0x1.bcb7b158af938p-4, + 0x1.63c78734e6d07p-4, -0x1.287461742fee4p-4 }, + .invln10 = 0x1.bcb7b1526e50ep-2, + .log10_2 = 0x1.34413509f79ffp-2, + /* Algorithm: + + x = 2^k z + log10(x) = k log10(2) + log10(c) + poly(z/c - 1) / log(10) + + where z is in [a;2a) which is split into N subintervals (a=0x1.69009p-1, + N=128) and log(c) and 1/c for the ith subinterval comes from lookup + tables: + + invc[i] = 1/c + log10c[i] = (double)log10(c) + + where c is near the center of the subinterval and is chosen by trying + several floating point invc candidates around 1/center and selecting one + for which the error in (double)log(c) is minimized (< 0x1p-74), except the + subinterval that contains 1 and the previous one got tweaked to avoid + cancellation. NB: invc should be optimized to minimize error in + (double)log10(c) instead. */ + .invc = { 0x1.6a133d0dec120p+0, 0x1.6815f2f3e42edp+0, + 0x1.661e39be1ac9ep+0, 0x1.642bfa30ac371p+0, + 0x1.623f1d916f323p+0, 0x1.60578da220f65p+0, + 0x1.5e75349dea571p+0, 0x1.5c97fd387a75ap+0, + 0x1.5abfd2981f200p+0, 0x1.58eca051dc99cp+0, + 0x1.571e526d9df12p+0, 0x1.5554d555b3fcbp+0, + 0x1.539015e2a20cdp+0, 0x1.51d0014ee0164p+0, + 0x1.50148538cd9eep+0, 0x1.4e5d8f9f698a1p+0, + 0x1.4cab0edca66bep+0, 0x1.4afcf1a9db874p+0, + 0x1.495327136e16fp+0, 0x1.47ad9e84af28fp+0, + 0x1.460c47b39ae15p+0, 0x1.446f12b278001p+0, + 0x1.42d5efdd720ecp+0, 0x1.4140cfe001a0fp+0, + 0x1.3fafa3b421f69p+0, 0x1.3e225c9c8ece5p+0, + 0x1.3c98ec29a211ap+0, 0x1.3b13442a413fep+0, + 0x1.399156baa3c54p+0, 0x1.38131639b4cdbp+0, + 0x1.36987540fbf53p+0, 0x1.352166b648f61p+0, + 0x1.33adddb3eb575p+0, 0x1.323dcd99fc1d3p+0, + 0x1.30d129fefc7d2p+0, 0x1.2f67e6b72fe7dp+0, + 0x1.2e01f7cf8b187p+0, 0x1.2c9f518ddc86ep+0, + 0x1.2b3fe86e5f413p+0, 0x1.29e3b1211b25cp+0, + 0x1.288aa08b373cfp+0, 0x1.2734abcaa8467p+0, + 0x1.25e1c82459b81p+0, 0x1.2491eb1ad59c5p+0, + 0x1.23450a54048b5p+0, 0x1.21fb1bb09e578p+0, + 0x1.20b415346d8f7p+0, 0x1.1f6fed179a1acp+0, + 0x1.1e2e99b93c7b3p+0, 0x1.1cf011a7a882ap+0, + 0x1.1bb44b97dba5ap+0, 0x1.1a7b3e66cdd4fp+0, + 0x1.1944e11dc56cdp+0, 0x1.18112aebb1a6ep+0, + 0x1.16e013231b7e9p+0, 0x1.15b1913f156cfp+0, + 0x1.14859cdedde13p+0, 0x1.135c2dc68cfa4p+0, + 0x1.12353bdb01684p+0, 0x1.1110bf25b85b4p+0, + 0x1.0feeafd2f8577p+0, 0x1.0ecf062c51c3bp+0, + 0x1.0db1baa076c8bp+0, 0x1.0c96c5bb3048ep+0, + 0x1.0b7e20263e070p+0, 0x1.0a67c2acd0ce3p+0, + 0x1.0953a6391e982p+0, 0x1.0841c3caea380p+0, + 0x1.07321489b13eap+0, 0x1.062491aee9904p+0, + 0x1.05193497a7cc5p+0, 0x1.040ff6b5f5e9fp+0, + 0x1.0308d19aa6127p+0, 0x1.0203beedb0c67p+0, + 0x1.010037d38bcc2p+0, 1.0, + 0x1.fc06d493cca10p-1, 0x1.f81e6ac3b918fp-1, + 0x1.f44546ef18996p-1, 0x1.f07b10382c84bp-1, + 0x1.ecbf7070e59d4p-1, 0x1.e91213f715939p-1, + 0x1.e572a9a75f7b7p-1, 0x1.e1e0e2c530207p-1, + 0x1.de5c72d8a8be3p-1, 0x1.dae50fa5658ccp-1, + 0x1.d77a71145a2dap-1, 0x1.d41c51166623ep-1, + 0x1.d0ca6ba0bb29fp-1, 0x1.cd847e8e59681p-1, + 0x1.ca4a499693e00p-1, 0x1.c71b8e399e821p-1, + 0x1.c3f80faf19077p-1, 0x1.c0df92dc2b0ecp-1, + 0x1.bdd1de3cbb542p-1, 0x1.baceb9e1007a3p-1, + 0x1.b7d5ef543e55ep-1, 0x1.b4e749977d953p-1, + 0x1.b20295155478ep-1, 0x1.af279f8e82be2p-1, + 0x1.ac5638197fdf3p-1, 0x1.a98e2f102e087p-1, + 0x1.a6cf5606d05c1p-1, 0x1.a4197fc04d746p-1, + 0x1.a16c80293dc01p-1, 0x1.9ec82c4dc5bc9p-1, + 0x1.9c2c5a491f534p-1, 0x1.9998e1480b618p-1, + 0x1.970d9977c6c2dp-1, 0x1.948a5c023d212p-1, + 0x1.920f0303d6809p-1, 0x1.8f9b698a98b45p-1, + 0x1.8d2f6b81726f6p-1, 0x1.8acae5bb55badp-1, + 0x1.886db5d9275b8p-1, 0x1.8617ba567c13cp-1, + 0x1.83c8d27487800p-1, 0x1.8180de3c5dbe7p-1, + 0x1.7f3fbe71cdb71p-1, 0x1.7d055498071c1p-1, + 0x1.7ad182e54f65ap-1, 0x1.78a42c3c90125p-1, + 0x1.767d342f76944p-1, 0x1.745c7ef26b00ap-1, + 0x1.7241f15769d0fp-1, 0x1.702d70d396e41p-1, + 0x1.6e1ee3700cd11p-1, 0x1.6c162fc9cbe02p-1 }, + .log10c = { -0x1.345825f221684p-3, -0x1.2f71a1f0c554ep-3, + -0x1.2a91fdb30b1f4p-3, -0x1.25b9260981a04p-3, + -0x1.20e7081762193p-3, -0x1.1c1b914aeefacp-3, + -0x1.1756af5de404dp-3, -0x1.12985059c90bfp-3, + -0x1.0de0628f63df4p-3, -0x1.092ed492e08eep-3, + -0x1.0483954caf1dfp-3, -0x1.ffbd27a9adbcp-4, + -0x1.f67f7f2e3d1ap-4, -0x1.ed4e1071ceebep-4, + -0x1.e428bb47413c4p-4, -0x1.db0f6003028d6p-4, + -0x1.d201df6749831p-4, -0x1.c9001ac5c9672p-4, + -0x1.c009f3c78c79p-4, -0x1.b71f4cb642e53p-4, + -0x1.ae400818526b2p-4, -0x1.a56c091954f87p-4, + -0x1.9ca3332f096eep-4, -0x1.93e56a3f23e55p-4, + -0x1.8b3292a3903bp-4, -0x1.828a9112d9618p-4, + -0x1.79ed4ac35f5acp-4, -0x1.715aa51ed28c4p-4, + -0x1.68d2861c999e9p-4, -0x1.6054d40ded21p-4, + -0x1.57e17576bc9a2p-4, -0x1.4f7851798bb0bp-4, + -0x1.47194f5690ae3p-4, -0x1.3ec456d58ec47p-4, + -0x1.36794ff3e5f55p-4, -0x1.2e382315725e4p-4, + -0x1.2600b8ed82e91p-4, -0x1.1dd2fa85efc12p-4, + -0x1.15aed136e3961p-4, -0x1.0d94269d1a30dp-4, + -0x1.0582e4a7659f5p-4, -0x1.faf5eb655742dp-5, + -0x1.eaf888487e8eep-5, -0x1.db0d75ef25a82p-5, + -0x1.cb348a49e6431p-5, -0x1.bb6d9c69acdd8p-5, + -0x1.abb88368aa7ap-5, -0x1.9c1517476af14p-5, + -0x1.8c833051bfa4dp-5, -0x1.7d02a78e7fb31p-5, + -0x1.6d93565e97c5fp-5, -0x1.5e351695db0c5p-5, + -0x1.4ee7c2ba67adcp-5, -0x1.3fab35ba16c01p-5, + -0x1.307f4ad854bc9p-5, -0x1.2163ddf4f988cp-5, + -0x1.1258cb5d19e22p-5, -0x1.035defdba3188p-5, + -0x1.e8e651191bce4p-6, -0x1.cb30a62be444cp-6, + -0x1.ad9a9b3043823p-6, -0x1.9023ecda1ccdep-6, + -0x1.72cc592bd82dp-6, -0x1.55939eb1f9c6ep-6, + -0x1.38797ca6cc5ap-6, -0x1.1b7db35c2c072p-6, + -0x1.fd400812ee9a2p-7, -0x1.c3c05fb4620f1p-7, + -0x1.8a7bf3c40e2e3p-7, -0x1.517249c15a75cp-7, + -0x1.18a2ea5330c91p-7, -0x1.c01abc8cdc4e2p-8, + -0x1.4f6261750dec9p-8, -0x1.be37b6612afa7p-9, + -0x1.bc3a8398ac26p-10, 0.0, + 0x1.bb796219f30a5p-9, 0x1.b984fdcba61cep-8, + 0x1.49cf12adf8e8cp-7, 0x1.b6075b5217083p-7, + 0x1.10b7466fc30ddp-6, 0x1.4603e4db6a3a1p-6, + 0x1.7aeb10e99e105p-6, 0x1.af6e49b0f0e36p-6, + 0x1.e38f064f41179p-6, 0x1.0ba75abbb7623p-5, + 0x1.25575ee2dba86p-5, 0x1.3ed83f477f946p-5, + 0x1.582aa79af60efp-5, 0x1.714f400fa83aep-5, + 0x1.8a46ad3901cb9p-5, 0x1.a311903b6b87p-5, + 0x1.bbb086f216911p-5, 0x1.d4242bdda648ep-5, + 0x1.ec6d167c2af1p-5, 0x1.0245ed8221426p-4, + 0x1.0e40856c74f64p-4, 0x1.1a269a31120fep-4, + 0x1.25f8718fc076cp-4, 0x1.31b64ffc95bfp-4, + 0x1.3d60787ca5063p-4, 0x1.48f72ccd187fdp-4, + 0x1.547aad6602f1cp-4, 0x1.5feb3989d3acbp-4, + 0x1.6b490f3978c79p-4, 0x1.76946b3f5e703p-4, + 0x1.81cd895717c83p-4, 0x1.8cf4a4055c30ep-4, + 0x1.9809f4c48c0ebp-4, 0x1.a30db3f9899efp-4, + 0x1.ae001905458fcp-4, 0x1.b8e15a2e3a2cdp-4, + 0x1.c3b1ace2b0996p-4, 0x1.ce71456edfa62p-4, + 0x1.d9205759882c4p-4, 0x1.e3bf1513af0dfp-4, + 0x1.ee4db0412c414p-4, 0x1.f8cc5998de3a5p-4, + 0x1.019da085eaeb1p-3, 0x1.06cd4acdb4e3dp-3, + 0x1.0bf542bef813fp-3, 0x1.11159f14da262p-3, + 0x1.162e761c10d1cp-3, 0x1.1b3fddc60d43ep-3, + 0x1.2049ebac86aa6p-3, 0x1.254cb4fb7836ap-3, + 0x1.2a484e8d0d252p-3, 0x1.2f3ccce1c860bp-3 } +}; diff --git a/sysdeps/aarch64/fpu/vecmath_config.h b/sysdeps/aarch64/fpu/vecmath_config.h index 82a68fd35d..4eb7b587b3 100644 --- a/sysdeps/aarch64/fpu/vecmath_config.h +++ b/sysdeps/aarch64/fpu/vecmath_config.h @@ -58,4 +58,13 @@ extern const struct v_log2_data double log2c[1 << V_LOG2_TABLE_BITS]; } __v_log2_data attribute_hidden; +#define V_LOG10_TABLE_BITS 7 +extern const struct v_log10_data +{ + double poly[5]; + double invln10, log10_2; + double invc[1 << V_LOG10_TABLE_BITS]; + double log10c[1 << V_LOG10_TABLE_BITS]; +} __v_log10_data attribute_hidden; + #endif diff --git a/sysdeps/aarch64/libm-test-ulps b/sysdeps/aarch64/libm-test-ulps index 7a5af571e2..6641c7fa0b 100644 --- a/sysdeps/aarch64/libm-test-ulps +++ b/sysdeps/aarch64/libm-test-ulps @@ -1180,11 +1180,19 @@ double: 2 float: 2 ldouble: 2 +Function: "log10_advsimd": +double: 1 +float: 2 + Function: "log10_downward": double: 2 float: 3 ldouble: 1 +Function: "log10_sve": +double: 1 +float: 2 + Function: "log10_towardzero": double: 2 float: 2 diff --git a/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist b/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist index 657edab7ae..f8776a6bea 100644 --- a/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist +++ b/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist @@ -15,13 +15,17 @@ GLIBC_2.38 _ZGVsMxv_logf F GLIBC_2.38 _ZGVsMxv_sin F GLIBC_2.38 _ZGVsMxv_sinf F GLIBC_2.39 _ZGVnN2v_exp2 F +GLIBC_2.39 _ZGVnN2v_log10 F GLIBC_2.39 _ZGVnN2v_log2 F GLIBC_2.39 _ZGVnN2v_tan F GLIBC_2.39 _ZGVnN4v_exp2f F +GLIBC_2.39 _ZGVnN4v_log10f F GLIBC_2.39 _ZGVnN4v_log2f F GLIBC_2.39 _ZGVnN4v_tanf F GLIBC_2.39 _ZGVsMxv_exp2 F GLIBC_2.39 _ZGVsMxv_exp2f F +GLIBC_2.39 _ZGVsMxv_log10 F +GLIBC_2.39 _ZGVsMxv_log10f F GLIBC_2.39 _ZGVsMxv_log2 F GLIBC_2.39 _ZGVsMxv_log2f F GLIBC_2.39 _ZGVsMxv_tan F