From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Monjalon Subject: Re: [PATCH] eal/x86: revert select optimized memcpy at run-time Date: Tue, 07 Nov 2017 01:15:27 +0100 Message-ID: <2907620.AlD9JGsWPK@xps> References: <1509713243-28460-1-git-send-email-xiaoyun.li@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Cc: dev@dpdk.org, zhihong.wang@intel.com, konstantin.ananyev@intel.com, bruce.richardson@intel.com, wenzhuo.lu@intel.com, helin.zhang@intel.com, ophirmu@mellanox.com To: Xiaoyun Li Return-path: Received: from out4-smtp.messagingengine.com (out4-smtp.messagingengine.com [66.111.4.28]) by dpdk.org (Postfix) with ESMTP id 933571B2A1 for ; Tue, 7 Nov 2017 01:15:29 +0100 (CET) In-Reply-To: <1509713243-28460-1-git-send-email-xiaoyun.li@intel.com> List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" 03/11/2017 13:47, Xiaoyun Li: > This patch reverts the patchset run-time Linking support including the > following 3 commits: > commit 84cc318424d4 ("eal/x86: select optimized memcpy at run-time") > commit c7fbc80fe60f ("test: select memcpy alignment unit at run-time") > commit 5f180ae32962 ("efd: move AVX2 lookup in its own compilation unit") > > The patchset would cause perf drop in vhost/virtio loopback performance > test. Because the run-time dispatch must cost at least a function call > comparing to the compile-time dispatch. And the reference cpu cycles value > is small. And in the test, when using 128-256 bytes packet, it would cause > 16%-20% perf drop with mergeble path. When using 256 bytes packet, it would > cause 13% perf drop with vector path. > So revert that patchset. > > Signed-off-by: Xiaoyun Li Revert applied