From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 55863 invoked by alias); 21 Aug 2015 05:20:20 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Received: (qmail 41443 invoked by uid 89); 21 Aug 2015 05:19:43 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.5 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_NONE,SPF_HELO_PASS,SPF_PASS autolearn=ham version=3.3.2 X-Spam-User: qpsmtpd, 2 recipients X-HELO: na01-bl2-obe.outbound.protection.outlook.com Received: from mail-bl2on0093.outbound.protection.outlook.com (HELO na01-bl2-obe.outbound.protection.outlook.com) (65.55.169.93) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-SHA256 encrypted) ESMTPS; Fri, 21 Aug 2015 05:18:40 +0000 Received: from BN1AFFO11FD052.protection.gbl (10.58.52.33) by BN1AFFO11HUB029.protection.gbl (10.58.52.139) with Microsoft SMTP Server (TLS) id 15.1.249.14; Fri, 21 Aug 2015 05:18:37 +0000 Authentication-Results: spf=pass (sender IP is 149.199.60.83) smtp.mailfrom=xilinx.com; gcc.gnu.org; dkim=none (message not signed) header.d=none; Received-SPF: Pass (protection.outlook.com: domain of xilinx.com designates 149.199.60.83 as permitted sender) receiver=protection.outlook.com; client-ip=149.199.60.83; helo=xsj-pvapsmtpgw01; Received: from xsj-pvapsmtpgw01 (149.199.60.83) by BN1AFFO11FD052.mail.protection.outlook.com (10.58.53.67) with Microsoft SMTP Server (TLS) id 15.1.249.14 via Frontend Transport; Fri, 21 Aug 2015 05:18:37 +0000 Received: from unknown-38-66.xilinx.com ([149.199.38.66] helo=xsj-pvapsmtp01) by xsj-pvapsmtpgw01 with esmtp (Exim 4.63) (envelope-from ) id 1ZSeiq-0006U9-HF; Thu, 20 Aug 2015 22:18:36 -0700 Received: from [127.0.0.1] (helo=xsj-smtp-dlp2.xlnx.xilinx.com) by xsj-pvapsmtp01 with esmtp (Exim 4.63) (envelope-from ) id 1ZSeiq-0004cq-Ca; Thu, 20 Aug 2015 22:18:36 -0700 Received: from xsj-pvapsmtp01 (smtp3.xilinx.com [149.199.38.66]) by xsj-smtp-dlp2.xlnx.xilinx.com (8.13.8/8.13.1) with ESMTP id t7L5IAjG004532; Thu, 20 Aug 2015 22:18:10 -0700 Received: from [172.22.159.25] (helo=XAP-PVEXCAS01.xlnx.xilinx.com) by xsj-pvapsmtp01 with esmtp (Exim 4.63) (envelope-from ) id 1ZSeip-0004cb-1o; Thu, 20 Aug 2015 22:18:35 -0700 Received: from XAP-PVEXMBX02.xlnx.xilinx.com ([fe80::6c95:7dae:8014:5ca1]) by XAP-PVEXCAS01.xlnx.xilinx.com ([::1]) with mapi id 14.03.0195.001; Fri, 21 Aug 2015 13:18:34 +0800 From: Ajit Kumar Agarwal To: Jeff Law , Richard Biener , GCC Patches , "gcc@gcc.gnu.org" CC: Vinod Kathail , Shail Aditya Gupta , Vidhumouli Hunsigida , "Nagaraju Mekala" Subject: [RFC]: Vectorization cost benefit changes. Date: Fri, 21 Aug 2015 05:29:00 -0000 Message-ID: <37378DC5BCD0EE48BA4B082E0B55DFAA4295E5C2@XAP-PVEXMBX02.xlnx.xilinx.com> Content-Type: multipart/mixed; boundary="_002_37378DC5BCD0EE48BA4B082E0B55DFAA4295E5C2XAPPVEXMBX02xln_" MIME-Version: 1.0 X-RCIS-Action: ALLOW X-EOPAttributedMessage: 0 X-Microsoft-Exchange-Diagnostics: 1;BN1AFFO11FD052;1:AlFtDPz5Pz4BajetDrVgwar5HQZBt6Hx6PDnZvdIVlrKx8kcQyrhOJz5TFNq60SQiYybcWJ7gQJNtuXOMrdbCEU7BO0LolPE9FgOM1vbWdVm/LXTuyVmjWiJmtSd6XImp/8Cc5gWBtyGV9eFhpW6evMuInNSTAKqXv0N1yoe86IYQiJELnE/DxsJh+XbqSGO3T2+EFsQXeWGXmvTP5EgFvZKISSHr5zSOyHSDl6qqUns6aMYNcUVvLq1FM6BEOCJZmmE6WnVpp6lJIQXOe5bT/zl9ilmHVR9WpyVL+JAPlYERbHzqrLKB6riSfFfEYJVWbBRWeAF/z1Hb3EDAQiw7zNGzfEKpBFCt4/TSWrkUrEzgerzRtWJmUhvjakRrB+X X-Forefront-Antispam-Report: CIP:149.199.60.83;CTRY:US;IPV:NLI;EFV:NLI;SFV:NSPM;SFS:(10009020)(6009001)(2980300002)(438002)(3050300001)(189002)(199003)(50986999)(62966003)(54356999)(87936001)(77156002)(512954002)(86362001)(46102003)(19580395003)(6806004)(55846006)(5890100001)(99936001)(84326002)(63266004)(229853001)(2501003)(5250100002)(106466001)(5260100001)(5007970100001)(2930100002)(92566002)(5003600100002)(2900100001)(2920100001)(4610100001)(81156007)(64706001)(4001540100001)(2656002)(5001830100001)(5001770100001)(5001920100001)(5001860100001)(102836002)(568964001)(189998001)(21490400002)(33656002)(107886002)(5004730100002)(5001960100002)(107986001)(4001430100001);DIR:OUT;SFP:1101;SCL:1;SRVR:BN1AFFO11HUB029;H:xsj-pvapsmtpgw01;FPR:;SPF:Pass;PTR:unknown-60-83.xilinx.com;MX:1;A:1;LANG:en; X-Microsoft-Exchange-Diagnostics: 1;BN1AFFO11HUB029;2:9Sq0Tt8el+3oKKDbwPXI1Ka1BLvqQFdSkxiJAQ0WBUAThB/5+NeLBKT1WOnq9FxIW0Kejhf/RC4urGS7QYtx2uw4D9A1UxCZvHz29Z3nslRCHdjBtWhaZjLYyiIjfq0BsCGI2vW3JLvZ3O2Q5wnA/XeVHqBW1BRGBDg+a+9RDdI=;3:mwPjuukZl/OfrHYHjmYhsgIaUJCwZ4VN8cL2bMsFe22sVqRypIivyzdlb7lZWkp1wUfkq+QHyy8NyD/cT2hQYoBiwPyJ07ntlZ0Jl+gQnpB1Jq5AUisL3FJnzbwE/ONVWDfL9g5z+w9GWJFvR+Qq2wH9LWZpiU+we+5SY4qZ1s4fkvZvBeEbHMMTMiniZRVsoVy5fAzLezhmM2JJV8Lwof8io3rBuE6lvu1nlw+LuNc=;25:9y48UpCnA4SUNfI2D/6nc/CBqN5kQDe7crosUjhpkEZsgaPxmxNECW8TueLkGGtTnECGz+6McAmpsAxMsgrwKRWM1qEIIqkRk3bD9gJJ9jV6/oUfsidC/Sk1h8WDDrsANOe/v1eVdj5LOQC3jXCcf0WssQxB5sfs01a+ZsddTPEniV7wZzlTJrxkKb86vStHXOny2v31a7GpHROVOdAzTzd2meuM/Hd0flhn67DfAlbn22dNZDbU7uVRvYhrLwoANT1SpjiDkLOxOX6LD4hF+A== X-Microsoft-Antispam: UriScan:;BCL:0;PCL:0;RULEID:;SRVR:BN1AFFO11HUB029; X-Microsoft-Exchange-Diagnostics: 1;BN1AFFO11HUB029;20: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;4:iKY8tSu9QHihW3NsoeNfIJ97WCsrU8JpII6hY3O0pC/Y8rTapaCsbbmMQsiffaRkliFG8lrkDGfDoalaNFoSAuR2uDJOF/MHae7SXKey3YFvECTlQGs+ZVtSJZnZmDSoK/TkZF6oVa0uuC16RPEESAPGtXOzvcBiRAAc5YEE4bAUwbLfJmgFZvoQ1P9kWfY3xH5lwgQTsayKlvtu1ssmtJyI/QoQplcwI0jwU9sO7UKZeowlcuo7suJWrZvLJKSzaxPSqaeYxQY86xP2nQhzDopg89LOhuSND2rcF3yoHSSHgXmSdZvO+hWgGn/xdCLQp8s3O4z5f6ZZ98vsIvJXlg== X-Microsoft-Antispam-PRVS: X-Exchange-Antispam-Report-Test: UriScan:; X-Exchange-Antispam-Report-CFA-Test: BCL:0;PCL:0;RULEID:(601004)(5005006)(8121501046)(3002001);SRVR:BN1AFFO11HUB029;BCL:0;PCL:0;RULEID:;SRVR:BN1AFFO11HUB029; X-Forefront-PRVS: 067553F396 X-Microsoft-Exchange-Diagnostics: =?us-ascii?Q?1;BN1AFFO11HUB029;23:pD+pcbbUAo6WgTaBlziQDfzj4p0c053BhtjkLdr?= =?us-ascii?Q?eFuv/0d0q3q07QGzPAfEAhwB0KnD0jfkuaJ8WoHmQ/akrgQdXGx2nHRack7Q?= =?us-ascii?Q?rigPw5aWeWzEtFylfjf3QLPiC1n/X/ndOswpRkVn1/iY3ksfHGldbCPQfPY0?= =?us-ascii?Q?h9MTNFuP7SqMODO6WBz18YUYvzlkxfukEtcvYmDEQUQcU14CtJU3gAne5/8B?= =?us-ascii?Q?UzQvdwM/QSBVIoRBY9riPIaqVlaVGcxPHYNF+XTAFXyuyephkGTg0NdYq9R3?= =?us-ascii?Q?KW7o2Hax7a2jWRX9rohXcjoD/dbADmFJnlgHDSPHEIXPa/Gzv5YbwTgC9J22?= =?us-ascii?Q?YNv5s7g8mLhZsEO8J1E6cn2PYBfbnKr+FHxTYCPP/Ru1V1pEJo1kThdZfsic?= =?us-ascii?Q?RiI7jIgqL5woZMJaQIEdV1pprwv1QBSalJpcCdxhTsSR9ZB+h9RkmOzK0dSp?= =?us-ascii?Q?QhAHhzXZLvenkONZynsAkATL8s98U70CZIZkjWlJHH/B92V8drz8xOr/pWv8?= =?us-ascii?Q?RMALJ8KjrUOC6ClmiinDNKJ4rslMh3vwH+vVtY8DRNKgLx2fSWQL4F8QSgpM?= =?us-ascii?Q?SveSQ9lr0nDhOwSIBDPZ2LrJJETHhb7+yB2LOjvSj4/SIPvP4vftjBPtQ8Fa?= =?us-ascii?Q?rFpvBj6mb/czFntsjbA5QC+CgBue678MNrTaDpKmmGyBUX7u2DSHXYPJesUM?= =?us-ascii?Q?o/pDJtDAiDanfiSJ49cdqg8jOtrFss6SFXOF03VM6GADWJd6SfIyWUae1ki7?= =?us-ascii?Q?n9Oq6YA9+2QMxAnpRdupyh/T19hHv0q+dFtCs4aBitBNwrs781pOOrlNiIAW?= =?us-ascii?Q?8IVECTyZJpiq2mcuODtoVZt3cBFaMc9jwv/1aGNvwRh99HNjtHzrzaIJkTHp?= =?us-ascii?Q?ASwt0CIxz6Vp3sgCwxbt71D17mmUzcHrdeMk+llu1+hg5Hzi8+sucHPIWGYw?= =?us-ascii?Q?U0fbuBINIKyR8XRqsTdPbjjuRbucUVKr6oLrMrV2vp+Y+Lq9CgPF+1XqpAVL?= =?us-ascii?Q?44QAprqp1U3D/GdTNI3NBUS50HOZmgQMrI5j2xG05HKC9L3K1TIiQlkHKyiW?= =?us-ascii?Q?scWXDjbm/zQclkehr1HqJYGnV1mdGmNNUmRJwrC+ZfE+zMaLwFwRq1VN9aKm?= =?us-ascii?Q?BJ0hsIYFXeqXcSuXzxQ3fwfu6NW76dM586f3BRzpWy7ogmvx0Sa5EZV/1E4D?= =?us-ascii?Q?LpdB3TQc0nYNS17IUMxtilh1w8eM8z6zpp/aN16gpfrORwDZgHsbWXm6yhIO?= =?us-ascii?Q?pAexGCd0fUsYyxnEylmCjyZReNtwUlm9O3xB2PF+PPcvJxTdqadpZsXEI/Hg?= =?us-ascii?Q?dVqiA0vpACa73VzHHDdnqr8EOVI4uKv+eJKKSfks3R+TII2jvOQ0RgpWz1QC?= =?us-ascii?Q?4YNMBMMgj5qYORN4vGdrboL8SiYcBfvF4G4Egyy/YVQAs+cCl?= X-Microsoft-Exchange-Diagnostics: 1;BN1AFFO11HUB029;5:B9a+j6Tf3VB1tv3KMTbJt2R06wOvwr65iAOwAXN7c6taBqByl+VL8CHS39tviOUIVQg++mNYs2X7+Im3bls0jCpJD8ahicGNkjIjoxW9aPN8hzckVHT8S8OIxb1V2A2ua2Qp3w0oRUh2ncFMcvhQrQ==;24:apsk49fDTYtrpkuSzKOtqp8G5zdVDZ/ye0R5c8QYaKCuJ2/NRqM7ViS+a0II9jlFFEZaYJXX4r19Urru56wmH9iYrbACEUgWRQ26cpv+Oos= SpamDiagnosticOutput: 1:23 SpamDiagnosticMetadata: NSPM X-OriginatorOrg: xilinx.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 21 Aug 2015 05:18:37.6229 (UTC) X-MS-Exchange-CrossTenant-Id: 657af505-d5df-48d0-8300-c31994686c5c X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=657af505-d5df-48d0-8300-c31994686c5c;Ip=[149.199.60.83];Helo=[xsj-pvapsmtpgw01] X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: BN1AFFO11HUB029 X-SW-Source: 2015-08/txt/msg01272.txt.bz2 --_002_37378DC5BCD0EE48BA4B082E0B55DFAA4295E5C2XAPPVEXMBX02xln_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Content-length: 2414 All: I have done the vectorization cost changes as given below. I have considere= d only the cost associated with the inner instead of outside. The consideration of inside scalar and vector cost is done as the inner cos= t are the most cost effective than the outside cost. min_profitable_iters =3D ((scalar_single_iter_cost - vec_inside_cost) *vf); The Scalar_single_iter_cost consider the hardcoded value 50 which is used f= or most of the targets and the scalar cost is multiplied With 50. This scalar cost is subtracted with vector cost and as the scalar = cost is increased the chances of vectorization is more with same Vectorization factor and more loops will be vectorized. I have not changed the iteration count which is hardcoded with 50 and I wil= l do the changes to replace the 50 with the static Estimates of iteration count if you agree upon the below changes.=20 I have ran the SPEC cpu 2000 benchmarks with the below changes for i386 tar= gets and the significant gains are achieved with respect To INT and FP benchmarks. Here is the data. Ratio of vectorization cost changes(FP benchmarks) vs Ratio of without vect= orization cost changes( FP benchmarks) =3D 4640.102 vs 4583.379. Ratio of vectorization cost changes (INT benchmarks ) vs Ratio of without v= ectorization cost changes( INT benchmarks0 =3D 3812.883 vs 3778.558 Please give your feedback on the below changes for vectorization cost benef= it. diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c index 422b883..35d538f 100644 --- a/gcc/tree-vect-loop.c +++ b/gcc/tree-vect-loop.c @@ -2987,11 +2987,8 @@ vect_estimate_min_profitable_iters (loop_vec_info lo= op_vinfo, min_profitable_iters =3D 1; else { - min_profitable_iters =3D ((vec_outside_cost - scalar_outside_cos= t) * vf - - vec_inside_cost * peel_iters_prologue - - vec_inside_cost * peel_iters_epilogue) - / ((scalar_single_iter_cost * vf) - - vec_inside_cost); + min_profitable_iters =3D ((scalar_single_iter_cost + - vec_inside_cost) *vf); if ((scalar_single_iter_cost * vf * min_profitable_iters) <=3D (((int) vec_inside_cost * min_profitable_iters) Thanks & Regards Ajit --_002_37378DC5BCD0EE48BA4B082E0B55DFAA4295E5C2XAPPVEXMBX02xln_ Content-Type: application/octet-stream; name="vect.diff" Content-Description: vect.diff Content-Disposition: attachment; filename="vect.diff"; size=874; creation-date="Fri, 21 Aug 2015 05:14:05 GMT"; modification-date="Fri, 21 Aug 2015 04:58:06 GMT" Content-Transfer-Encoding: base64 Content-length: 1188 ZGlmZiAtLWdpdCBhL2djYy90cmVlLXZlY3QtbG9vcC5jIGIvZ2NjL3RyZWUt dmVjdC1sb29wLmMKaW5kZXggNDIyYjg4My4uMzVkNTM4ZiAxMDA2NDQKLS0t IGEvZ2NjL3RyZWUtdmVjdC1sb29wLmMKKysrIGIvZ2NjL3RyZWUtdmVjdC1s b29wLmMKQEAgLTI5ODcsMTEgKzI5ODcsOCBAQCB2ZWN0X2VzdGltYXRlX21p bl9wcm9maXRhYmxlX2l0ZXJzIChsb29wX3ZlY19pbmZvIGxvb3BfdmluZm8s CiAgICAgICAgIG1pbl9wcm9maXRhYmxlX2l0ZXJzID0gMTsKICAgICAgIGVs c2UKICAgICAgICAgewotICAgICAgICAgIG1pbl9wcm9maXRhYmxlX2l0ZXJz ID0gKCh2ZWNfb3V0c2lkZV9jb3N0IC0gc2NhbGFyX291dHNpZGVfY29zdCkg KiB2ZgotCQkJCSAgLSB2ZWNfaW5zaWRlX2Nvc3QgKiBwZWVsX2l0ZXJzX3By b2xvZ3VlCi0gICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgLSB2 ZWNfaW5zaWRlX2Nvc3QgKiBwZWVsX2l0ZXJzX2VwaWxvZ3VlKQotICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgLyAoKHNjYWxhcl9zaW5nbGVf aXRlcl9jb3N0ICogdmYpCi0gICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAtIHZlY19pbnNpZGVfY29zdCk7CisgICAgICAgICAgbWluX3By b2ZpdGFibGVfaXRlcnMgPSAoKHNjYWxhcl9zaW5nbGVfaXRlcl9jb3N0ICAK KyAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIC0gdmVjX2lu c2lkZV9jb3N0KSAqdmYpOwogCiAgICAgICAgICAgaWYgKChzY2FsYXJfc2lu Z2xlX2l0ZXJfY29zdCAqIHZmICogbWluX3Byb2ZpdGFibGVfaXRlcnMpCiAg ICAgICAgICAgICAgIDw9ICgoKGludCkgdmVjX2luc2lkZV9jb3N0ICogbWlu X3Byb2ZpdGFibGVfaXRlcnMpCg== --_002_37378DC5BCD0EE48BA4B082E0B55DFAA4295E5C2XAPPVEXMBX02xln_--