From patchwork Wed May 18 17:26:34 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Noah Goldstein X-Patchwork-Id: 54172 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 2A7853856DC5 for ; Wed, 18 May 2022 17:29:13 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 2A7853856DC5 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1652894953; bh=kCaK88SxrRZGN8OM6+L3uZIO7mcP3fV2eiWX4EuZI8c=; h=To:Subject:Date:In-Reply-To:References:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=NN4Zcd2TqgYQlAaKUfNvQezlkWag+5tnzYeFlY6RgZMviZA+/x1eEerHTX8SEIh7g hpI1L8Dh80LcZCQtQCxecqNsXC7eGiLHtCiDAGmoJbpTdLAJNG/sSaKYnxSrSqwX+p K7sOSPsE4wFq37I5gCFXtHjROjhHhiOrgnuhmv2w= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mail-io1-xd2e.google.com (mail-io1-xd2e.google.com [IPv6:2607:f8b0:4864:20::d2e]) by sourceware.org (Postfix) with ESMTPS id C7BE63856DC5 for ; Wed, 18 May 2022 17:26:45 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org C7BE63856DC5 Received: by mail-io1-xd2e.google.com with SMTP id y12so3037430ior.7 for ; Wed, 18 May 2022 10:26:45 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=kCaK88SxrRZGN8OM6+L3uZIO7mcP3fV2eiWX4EuZI8c=; b=VQOHGFbBxgKJzTxN96mPDWc+Zk5a3NCabBGLp5FFsCA+vDJlQ0kGPwjGZ9nuCp4Mr7 g73IISVx95+7vUFEjeT+T1k2lNmkZLcKW4TZImt2u5ipk82YMyjwzUMtS0uTFqZIfXJ7 /Tp6+4GLNxXQbN35lPXIPPu3PBZKBm/S58nVeVsZgdiV1Ulg4GYfVyB4UiV3j8G7viol C71q6sdVBnkDxryzvNXLLheTCmm/jCkNpRegVdmDu8WV9mNorYKKvOKDnaGm9YrMmfMl 29cWI2/lNUysw2ChSjzzYoqFSXivRUFUUFKnnu9NN8jtTJ6Vu62aw3puBHvauCFrNUe1 MJlA== X-Gm-Message-State: AOAM5337GNKIrk2ZmmUXeGQfNPEOoBHTSUsFGRFjAQMmjbbRwGsi3h3S 8rf8DzgepgTBiYe7NzoPfJOv7LRwOI4= X-Google-Smtp-Source: ABdhPJwH9737AesEzkcrksVVWizq5DA6+WCLGk8yVM1Hx9ifV27E8EjUbh2vGGIF4RHyum/yBreVgQ== X-Received: by 2002:a05:6638:160d:b0:32b:d9d2:f2f2 with SMTP id x13-20020a056638160d00b0032bd9d2f2f2mr346609jas.68.1652894804879; Wed, 18 May 2022 10:26:44 -0700 (PDT) Received: from noah-tgl.. (node-17-161.flex.volo.net. [76.191.17.161]) by smtp.gmail.com with ESMTPSA id r2-20020a02b102000000b0032b3a78179fsm18598jah.99.2022.05.18.10.26.44 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 18 May 2022 10:26:44 -0700 (PDT) To: libc-alpha@sourceware.org Subject: [PATCH v10 5/6] nss: Optimize nss_hash in nss_hash.c Date: Wed, 18 May 2022 12:26:34 -0500 Message-Id: <20220518172635.291119-5-goldstein.w.n@gmail.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20220518172635.291119-1-goldstein.w.n@gmail.com> References: <20220414041231.926415-1-goldstein.w.n@gmail.com> <20220518172635.291119-1-goldstein.w.n@gmail.com> MIME-Version: 1.0 X-Spam-Status: No, score=-11.3 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, FREEMAIL_FROM, GIT_PATCH_0, LIKELY_SPAM_BODY, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Noah Goldstein via Libc-alpha From: Noah Goldstein Reply-To: Noah Goldstein Errors-To: libc-alpha-bounces+patchwork=sourceware.org@sourceware.org Sender: "Libc-alpha" The prior unrolling didn't really do much as it left the dependency chain between iterations. Unrolled the loop for 4 so 4x multiplies could be pipelined in out-of-order machines. Results for __nss_hash Benchmarked on Tigerlake: 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz Time as Geometric Mean of N=25 runs Geometric of all benchmark New / Old: 0.845 type, length, New Time, Old Time, New Time / Old Time fixed, 0, 4.019, 3.729, 1.078 fixed, 1, 4.95, 5.707, 0.867 fixed, 2, 5.152, 5.657, 0.911 fixed, 3, 4.641, 5.721, 0.811 fixed, 4, 5.551, 5.81, 0.955 fixed, 5, 6.525, 6.552, 0.996 fixed, 6, 6.711, 6.561, 1.023 fixed, 7, 6.715, 6.767, 0.992 fixed, 8, 7.874, 7.915, 0.995 fixed, 9, 8.888, 9.767, 0.91 fixed, 10, 8.959, 9.762, 0.918 fixed, 11, 9.188, 9.987, 0.92 fixed, 12, 9.708, 10.618, 0.914 fixed, 13, 10.393, 11.14, 0.933 fixed, 14, 10.628, 12.097, 0.879 fixed, 15, 10.982, 12.965, 0.847 fixed, 16, 11.851, 14.429, 0.821 fixed, 32, 24.334, 34.414, 0.707 fixed, 64, 55.618, 86.688, 0.642 fixed, 128, 118.261, 224.36, 0.527 fixed, 256, 256.183, 538.629, 0.476 random, 2, 11.194, 11.556, 0.969 random, 4, 17.516, 17.205, 1.018 random, 8, 23.501, 20.985, 1.12 random, 16, 28.131, 29.212, 0.963 random, 32, 35.436, 38.662, 0.917 random, 64, 45.74, 58.868, 0.777 random, 128, 75.394, 121.963, 0.618 random, 256, 139.524, 260.726, 0.535 --- nss/nss_hash.c | 92 ++++++++++++++++++++++---------------------------- 1 file changed, 41 insertions(+), 51 deletions(-) diff --git a/nss/nss_hash.c b/nss/nss_hash.c index f9e17d068a..1d3787e675 100644 --- a/nss/nss_hash.c +++ b/nss/nss_hash.c @@ -19,74 +19,64 @@ /* This is from libc/db/hash/hash_func.c, hash3 is static there */ /* - * This is INCREDIBLY ugly, but fast. We break the string up into 8 byte + * This is INCREDIBLY ugly, but fast. We break the string up into 4 byte * units. On the first time through the loop we get the "leftover bytes" - * (strlen % 8). On every other iteration, we perform 8 HASHC's so we handle - * all 8 bytes. Essentially, this saves us 7 cmp & branch instructions. If - * this routine is heavily used enough, it's worth the ugly coding. + * (len % 4). On every other iteration, we perform a 4x unrolled version + * HASHC. Further unrolling does not appear to help. * * OZ's original sdbm hash */ uint32_t __nss_hash (const void *keyarg, size_t len) { + enum + { + HASH_CONST_P0 = 1, /* (uint32_t)(65599 ^ 0). */ + HASH_CONST_P1 = 65599, /* (uint32_t)(65599 ^ 1). */ + HASH_CONST_P2 = 8261505, /* (uint32_t)(65599 ^ 2). */ + HASH_CONST_P3 = 780587199, /* (uint32_t)(65599 ^ 3). */ + HASH_CONST_P4 = 1139564289 /* (uint32_t)(65599 ^ 4). */ + }; + const unsigned char *key; - size_t loop; uint32_t h; -#define HASHC h = *key++ + 65599 * h +#define HASHC h = *key++ + HASH_CONST_P1 * h h = 0; key = keyarg; if (len > 0) { - loop = (len + 8 - 1) >> 3; - switch (len & (8 - 1)) - { - case 0: - do - { - HASHC; - /* FALLTHROUGH */ - case 7: - HASHC; - /* FALLTHROUGH */ - case 6: - HASHC; - /* FALLTHROUGH */ - case 5: - HASHC; - /* FALLTHROUGH */ - case 4: - HASHC; - /* FALLTHROUGH */ - case 3: - HASHC; - /* FALLTHROUGH */ - case 2: - HASHC; - /* FALLTHROUGH */ - case 1: - HASHC; - } - while (--loop); - } - } - return h; -} + switch ((len & (4 - 1))) + { + case 0: + /* h starts out as zero so no need to include the multiply. */ + h = *key++; + /* FALLTHROUGH */ + case 3: + HASHC; + /* FALLTHROUGH */ + case 2: + HASHC; + /* FALLTHROUGH */ + case 1: + HASHC; + /* FALLTHROUGH */ + } -/* For testing/benchmarking purposes. */ -static uint32_t -__simple_nss_hash (const void *keyarg, size_t len) -{ - const unsigned char *key; - size_t i; - uint32_t h = 0; - key = keyarg; - - for (i = 0; i < len; ++i) - h = *key++ + 65599 * h; + uint32_t c0, c1, c2, c3; + for (--len; len >= 4; len -= 4) + { + c0 = (unsigned char) *(key + 0); + c1 = (unsigned char) *(key + 1); + c2 = (unsigned char) *(key + 2); + c3 = (unsigned char) *(key + 3); + h = HASH_CONST_P4 * h + HASH_CONST_P3 * c0 + HASH_CONST_P2 * c1 + + HASH_CONST_P1 * c2 + HASH_CONST_P0 * c3; + key += 4; + } + } return h; }