From mboxrd@z Thu Jan 1 00:00:00 1970 From: Adrien Mazarguil Subject: Re: [PATCH] net/mlx5: remove excessive data prefetch Date: Tue, 13 Mar 2018 13:44:59 +0100 Message-ID: <20180313124459.GI3994@6wind.com> References: <20180312170545.16165-1-yskoh@mellanox.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: nelio.laranjeiro@6wind.com, dev@dpdk.org To: Yongseok Koh Return-path: Received: from mail-wr0-f173.google.com (mail-wr0-f173.google.com [209.85.128.173]) by dpdk.org (Postfix) with ESMTP id 238A2A486 for ; Tue, 13 Mar 2018 13:45:14 +0100 (CET) Received: by mail-wr0-f173.google.com with SMTP id r8so7122920wrg.0 for ; Tue, 13 Mar 2018 05:45:14 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20180312170545.16165-1-yskoh@mellanox.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" On Mon, Mar 12, 2018 at 10:05:45AM -0700, Yongseok Koh wrote: > In Enhanced Multi-Packet Send (eMPW), entire packet data is prefetched to > LLC if it isn't inlined. Even though this helps reducing jitter when HW > fetches data by DMA, this can thresh the LLC with evicting precious data. > And if the size of queue is large and there are many queues, this might not > be effective. Also, if application runs on a remote node from the PCIe > link, it may not be helpful and can even cause bad results. > > Signed-off-by: Yongseok Koh Acked-by: Adrien Mazarguil -- Adrien Mazarguil 6WIND