From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by dpdk.org (Postfix) with ESMTP id 369835A8C for ; Fri, 17 Jul 2015 09:34:56 +0200 (CEST) Received: from orsmga001.jf.intel.com ([10.7.209.18]) by fmsmga102.fm.intel.com with ESMTP; 17 Jul 2015 00:34:42 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.15,494,1432623600"; d="scan'208";a="730600625" Received: from irsmsx104.ger.corp.intel.com ([163.33.3.159]) by orsmga001.jf.intel.com with ESMTP; 17 Jul 2015 00:34:41 -0700 Received: from irsmsx108.ger.corp.intel.com ([169.254.11.201]) by IRSMSX104.ger.corp.intel.com ([169.254.5.171]) with mapi id 14.03.0224.002; Fri, 17 Jul 2015 08:34:40 +0100 From: "De Lara Guarch, Pablo" To: Tony Lu , "dev@dpdk.org" Thread-Topic: [dpdk-dev] [PATCH v7 1/7] hash: replace existing hash library with cuckoo hash implementation Thread-Index: AQHQu29EOFIwuy285UGf11xGpF0wDp3dzxKAgADIHeCAAFH/8IAAYyMA Date: Fri, 17 Jul 2015 07:34:39 +0000 Message-ID: References: <1436571020-16252-1-git-send-email-pablo.de.lara.guarch@intel.com> <1436573936-15956-1-git-send-email-pablo.de.lara.guarch@intel.com> <1436573936-15956-2-git-send-email-pablo.de.lara.guarch@intel.com> <001601d0bfab$6c4071e0$44c155a0$@com> <001c01d0c041$98b01c30$ca105490$@com> In-Reply-To: <001c01d0c041$98b01c30$ca105490$@com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [163.33.239.182] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH v7 1/7] hash: replace existing hash library with cuckoo hash implementation X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 17 Jul 2015 07:34:57 -0000 > -----Original Message----- > From: Tony Lu [mailto:zlu@ezchip.com] > Sent: Friday, July 17, 2015 4:35 AM > To: De Lara Guarch, Pablo; dev@dpdk.org > Subject: RE: [dpdk-dev] [PATCH v7 1/7] hash: replace existing hash librar= y > with cuckoo hash implementation >=20 > Hi, Pablo >=20 > >-----Original Message----- > >From: De Lara Guarch, Pablo [mailto:pablo.de.lara.guarch@intel.com] > >Sent: Friday, July 17, 2015 4:42 AM > >To: Tony Lu; dev@dpdk.org > >Subject: RE: [dpdk-dev] [PATCH v7 1/7] hash: replace existing hash libra= ry > with > >cuckoo hash implementation > > > >Hi Tony, > > > >> -----Original Message----- > >> From: Tony Lu [mailto:zlu@ezchip.com] > >> Sent: Thursday, July 16, 2015 10:40 AM > >> To: De Lara Guarch, Pablo; dev@dpdk.org > >> Subject: RE: [dpdk-dev] [PATCH v7 1/7] hash: replace existing hash > >> library with cuckoo hash implementation > >> > >> >diff --git a/lib/librte_hash/rte_cuckoo_hash.c > >> b/lib/librte_hash/rte_cuckoo_hash.c > >> >new file mode 100644 > >> >index 0000000..50e3acd > >> >--- /dev/null > >> >+++ b/lib/librte_hash/rte_cuckoo_hash.c > >> >@@ -0,0 +1,1027 @@ > >> ... > >> >+ > >> >+/* Functions to compare multiple of 16 byte keys (up to 128 bytes) > >> >+*/ static int rte_hash_k16_cmp_eq(const void *key1, const void > >> >+*key2, size_t key_len > >> >__rte_unused) > >> >+{ > >> >+ const __m128i k1 =3D _mm_loadu_si128((const __m128i *) key1); > >> >+ const __m128i k2 =3D _mm_loadu_si128((const __m128i *) key2); > >> >+ const __m128i x =3D _mm_xor_si128(k1, k2); > >> >+ > >> >+ return !_mm_test_all_zeros(x, x); > >> >+} > >> ... > >> > >> When compiling the latest dev DPDK for non-x86 arch, it fails on the > >> above code, as the SSE is x86 specific defined in . Is > >> it possible to replace this function with platform independent code? > > > >Thanks for spotting this. I just sent a patch that should fix the proble= m. > >Can you check if it works? >=20 > Thanks for your quick response, but __m128i and all the _mm_ related > functions > are X86 specific defined in . This header file is only > available in X86 > compiler library, but no-X86 archs do not have this file. So if we can > replace all > the X86 specific code in the above function, that would be great. >=20 With the patch that I sent, if you are compiling for a non-x86 arch, you sh= ould not have any problem, since all that code will only be used if using x86 arch. Have you tried com= piling DPDK with the patch? Pablo > Thanks > -Tony >=20 >=20 > >Thanks, > >Pablo > >> > >> Thanks > >> -Zhigang Lu >=20