From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754144AbbHFCbz (ORCPT ); Wed, 5 Aug 2015 22:31:55 -0400 Received: from mail-by2on0130.outbound.protection.outlook.com ([207.46.100.130]:10706 "EHLO na01-by2-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754076AbbHFCbv (ORCPT ); Wed, 5 Aug 2015 22:31:51 -0400 Authentication-Results: spf=none (sender IP is ) smtp.mailfrom=scottwood@freescale.com; Message-ID: <1438828301.2097.126.camel@freescale.com> Subject: Re: [PATCH v2 2/2] powerpc32: optimise csum_partial() loop From: Scott Wood To: Segher Boessenkool CC: Christophe Leroy , Benjamin Herrenschmidt , Paul Mackerras , "Michael Ellerman" , , Date: Wed, 5 Aug 2015 21:31:41 -0500 In-Reply-To: <20150806003059.GD18479@gate.crashing.org> References: <67cf476f657e87b2ea586951a57ae3ba3c1e3c0c.1435655733.git.christophe.leroy@c-s.fr> <20150806003059.GD18479@gate.crashing.org> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.16.0-fta1 MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Originating-IP: [2601:448:8100:f9f:12bf:48ff:fe84:c9a0] X-ClientProxiedBy: BY2PR04CA0070.namprd04.prod.outlook.com (10.255.247.38) To BN3PR03MB1480.namprd03.prod.outlook.com (25.163.35.143) X-Microsoft-Exchange-Diagnostics: 1;BN3PR03MB1480;2:5dcNqaFlID4UyHhK2rJi4UAdoAJogRpsiuB8mi/nbqmriSEr4acaybSN4PGfgSPQzzhaLhheoG2/+MD2l4smikfzTezwd90wEFZTcfm64RW9riwWVSf0jG8xjxRNxcz9jaBdi0Fla/lgjd1x5EMB5+ezeL02dQVyiLSPYfAMhTc=;3:mFGfNcjNRMS1UtqM2yJrVxK0IvdpNtixSRBZkMZRkn7uoDe9G1X2v7brBqGBGXnuUHA8gm/W948PK3eN7nwcSTNB8jPVmzlaMloGn+rcwBkFqR1kHNXEwEyukpcka0LW86R7hfvwL3wHPUJjAz+O4w==;25:nCFIOa33KZo7WOmRhtNWYBd6MuKCZjp8WgYbb144D/bBtzh0MGWF5HR/jZD5nCw1lpmPPIrdImnphpaLcH6dSog/gH3K8uqsBvsfhU86pATS2Nf+XpKNQjV+IitpZe+ubS4n1WZDxXp9/u+zxCsPuE+BxvVOELIdXq4nwgLLNrCifG1Cp1Z7qPC/g5FPh/kXeu0ps7RcouIRA9FrNiyMlL6FP+AP75VRBdso9Z9VhvQgY9B8xX+4AX7TYKCbypOoMGtHEqZshycjX1Zqc2DhmQ== X-Microsoft-Antispam: UriScan:;BCL:0;PCL:0;RULEID:;SRVR:BN3PR03MB1480; X-Microsoft-Exchange-Diagnostics: 1;BN3PR03MB1480;20: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;4:X3PWXTh5iRDaNyp7ZvxGvcFZTh5nw+VRoyjH/uK1Bbk7z0bKcZAiwPe9Cuee2tnLCKlFV6SlAhub91BHc/ePLpm92ih6S7NpijG/ionQfTHaSeF60Vv5l9HxTdZW8f+bfelV1PvSLZgwDfe6mk+ZekJBfdxIC2uX08cwsbQuejotax5jWlNJxMFoiYIWn5CyTSQx+29uz8NA0spoPw0UGCkEIp7JxdL9j/pmm/PnL4SjuxNFxHENvboiIl5iDoNWpfPLvLt9JnOk5APj4ZorZR/ELegClBknpowLiCosr7A= X-Microsoft-Antispam-PRVS: X-Exchange-Antispam-Report-Test: UriScan:; X-Exchange-Antispam-Report-CFA-Test: BCL:0;PCL:0;RULEID:(601004)(5005006)(3002001);SRVR:BN3PR03MB1480;BCL:0;PCL:0;RULEID:;SRVR:BN3PR03MB1480; X-Forefront-PRVS: 06607E485E X-Forefront-Antispam-Report: SFV:NSPM;SFS:(10019020)(6009001)(189002)(377424004)(24454002)(199003)(2950100001)(86362001)(33646002)(103116003)(68736005)(77096005)(50226001)(92566002)(87976001)(101416001)(47776003)(76176999)(64706001)(189998001)(122386002)(77156002)(5001830100001)(106356001)(50466002)(97736004)(81156007)(50986999)(5820100001)(5001860100001)(46102003)(62966003)(5001920100001)(23676002)(40100003)(105586002)(36756003)(110136002)(5001960100002)(4001540100001)(42186005)(99106002)(3826002)(5001840100002);DIR:OUT;SFP:1102;SCL:1;SRVR:BN3PR03MB1480;H:[IPv6:2601:448:8100:f9f:12bf:48ff:fe84:c9a0];FPR:;SPF:None;PTR:InfoNoRecords;MX:1;A:1;LANG:en; X-Microsoft-Exchange-Diagnostics: =?utf-8?B?MTtCTjNQUjAzTUIxNDgwOzIzOmZBTmtrcDlTTGFzT01JSFhoZFZsZm1nU0wr?= =?utf-8?B?VXk3U3NBVENna2Y0dVNKdGtjVm5nYi8yQ0htUk1NcmxkQXR3elVEVEtDZUtT?= =?utf-8?B?Vk82bUZCdGxCZ1R3SzR0VTM3dnVGVnVvcGowQXBTZHVYeG5UOWJmdS9HS3Zq?= =?utf-8?B?bVBaL3JjS0E5N2phNzRvbkRLQ1BjZHg5eHQwaU8xOWI1a3VLbUR1eGxiNXVO?= =?utf-8?B?Qk9TM2hnSkhGWldBQXVHakxJWktpNmhnZmVLWC8rQmVWQTVEazRVY0Z4S3or?= =?utf-8?B?ZkF5UmpCMmkvUFRjRTNxRUgvZE5qa0lWVENVa1RYdXRWb0d5U2NLWjFMMlhm?= =?utf-8?B?TldidXNaUGo2d3hEaW13MlFjcTRDRkVRK0VRMXpMTm93RnpIWUJablJJaC9X?= =?utf-8?B?Z01uZGoxMVZrcW1WV041K1dkblZ4MlRtakxuRHQ5cXgvR1VNOUFlWWRjL0ZE?= =?utf-8?B?WXdicXVDTFlKcUdIVTlNY2w0aG1tVkYySVdpLzhwbXNSY1JQS2ZCUzl0ZVc3?= =?utf-8?B?OFltRUNJbW80MmRlZUpxRE42QkQrOHdtUE8xVTlYdjBkZWVMYUVVMm83REFn?= =?utf-8?B?cjdhQ1h6RTBZeWcrYmlVRi9iRkZreHRNbElZYXFnQjhXTWZZL09TYnFhaERU?= =?utf-8?B?NTJMNHZRbnQyeVNieGJuMzRSRVJvYkhCWlh5aUpIT043cEgrQkJ0aTNHL1Ri?= =?utf-8?B?cWNlQXZic28rM0ZLUWhqTGljNHJKNms3S1RtNkNiKzNncDlYcS9kUFdqSWM0?= =?utf-8?B?MnZONTNxSjAwMHVHaFJaWkt1KzlZZnpHVXlyckpDTklvWXNUdldWWndEbmRP?= =?utf-8?B?OGRLcjZEdmxLUnUxajJ4WlczdWxRRlRlWHNnVWp4YU5MNkMvNlNuTDdoMmdC?= =?utf-8?B?WHFvTnF1RlhGaXZkN1YzamZFelJtUWtoK3l2VWpzRW80K0NLWTYreEYyNG1Q?= =?utf-8?B?VGpvZUVKaXQ0MENON21neW9PM0gwa2JuZ00yQi9tUDNUVkI1WHo1akRDSnU0?= =?utf-8?B?ZzZjTkVCV04va2E3THRjTkhpNUJjOUY2UUx1TEhYYjVkaTlEcVY5dzFNM1RT?= =?utf-8?B?L1lNazI2T3JDZmNaU0pjVDlxbmIzVWsxa0Zma1FMalFXM29lbkF1RUowMkFQ?= =?utf-8?B?QThqUkFMeUtFNGF6eDJrL2V6U2RiMG9LVnRSUGZGckkvOERlaC9EMUw3d0Rj?= =?utf-8?B?UmhmZ2FNdUZwQTQ3QXdmWXFzMjYyV21zQXpmSkdhL25rNkxRWkJ4Vy80ZWc1?= =?utf-8?B?RXlhaGoxUFRSaUtCdXErS3poTzFZRldJR3d2TzFWRHJZZTV3WWg0MVF0R3Zj?= =?utf-8?B?NG5SaCtUQWhTL2RTcFY4K3gvY0JGSHFYN216RlVqUXNCcHNUNGZVNXplVDg4?= =?utf-8?B?N3dQZmlpS0pTWE0wTk5ha0dFcU5oM2hSSjhiN3RXUy90R0tzYU95UXNGWWJ3?= =?utf-8?B?VkxIM1lZVVBrc2JMQVRBVWRCTy80bXR6c2hwZG8zdGR0bVU3Nm5SaG1mVFhF?= =?utf-8?Q?15dT7SrLaakQt+b+/tnLI31+hf8nH2WULQOXzS6RO6zR0T?= X-Microsoft-Exchange-Diagnostics: 1;BN3PR03MB1480;5:T8eGUMkH+5EDKhk0GndIl9MunvajSFfMVHv5XSuOYGAwhw+VDUYa9/jZdxdDnCotHK6W3O2Fc5JiCjR5XSFSkz+4Qxq3DNm0xIz3tQJ4VXd3sa8VQEkry7n+vSyGrI+IXW10S15HPzubtczYZt7vrA==;24:kXFX0DVW0fDxk+g+dmvFhFvaGn9eHoTaASCM0GXnl5Ui1o2wNQpn8JOLxVK/cXPUxrp2y9pPRWTAQRltF5pmIOhDrsFKJURSmnifSzpZdXY=;20:EHGPrn2rM23CEX6UuLkl5kjgB3qBMFsQ783EYeJwzW6JM5G9lr3V2twYJggWROUIwYw4gCu6cpc3IcYJwlXJbw== X-OriginatorOrg: freescale.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 06 Aug 2015 02:31:47.9598 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-Transport-CrossTenantHeadersStamped: BN3PR03MB1480 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 2015-08-05 at 19:30 -0500, Segher Boessenkool wrote: > On Wed, Aug 05, 2015 at 03:29:35PM +0200, Christophe Leroy wrote: > > On the 8xx, load latency is 2 cycles and taking branches also takes > > 2 cycles. So let's unroll the loop. > > This is not true for most other 32-bit PowerPC; this patch makes > performance worse on e.g. 6xx/7xx/7xxx. Let's not! Chips with a load latency greater than 2 cycles should also benefit from the unrolling. Have you benchmarked this somewhere and seen it reduce performance? Do you know of any 32-bit PPC chips with a load latency less than 2 cycles? -Scott