From patchwork Wed Nov 14 14:45:13 2018
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
X-Patchwork-Submitter: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
X-Patchwork-Id: 30143
Received: (qmail 57207 invoked by alias); 14 Nov 2018 14:45:40 -0000
Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Id: <libc-alpha.sourceware.org>
List-Unsubscribe: <mailto:libc-alpha-unsubscribe-##L=##H@sourceware.org>
List-Subscribe: <mailto:libc-alpha-subscribe@sourceware.org>
List-Archive: <http://sourceware.org/ml/libc-alpha/>
List-Post: <mailto:libc-alpha@sourceware.org>
List-Help: <mailto:libc-alpha-help@sourceware.org>,
	<http://sourceware.org/ml/#faqs>
Sender: libc-alpha-owner@sourceware.org
Delivered-To: mailing list libc-alpha@sourceware.org
Received: (qmail 57029 invoked by uid 89); 14 Nov 2018 14:45:32 -0000
Authentication-Results: sourceware.org; auth=none
X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00, GIT_PATCH_0,
	GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, RCVD_IN_DNSWL_NONE,
	SPF_HELO_PASS,
	SPF_PASS autolearn=ham version=3.3.2 spammy=Hx-languages-length:1310,
	bhs, biased
X-HELO: EUR03-VE1-obe.outbound.protection.outlook.com
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=armh.onmicrosoft.com;
	s=selector1-arm-com;
	h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck;
	bh=ElHsn5BjKgyWhtoKI8TUbbrQIECOtNKZvNmgIgGykOc=;
	b=pFb+yjeCwT5038Lh90A/YwU+IsFpcx89/CjPMCwbtdSDuYSNx4DpYDzbVDKSynRO1gJ5LmypvpSFusgXinCqLD4iGLfn6E0COLb3KATVXPDvaH3wjzVUJqKkpyRc2qEBijSsRNZ8Aw6/qPYYNRkRsgWdvVtVGh3ptph9Zwt6MKs=
From: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
To: 'GNU C Library' <libc-alpha@sourceware.org>, Szabolcs Nagy
	<Szabolcs.Nagy@arm.com>
CC: nd <nd@arm.com>
Subject: Re: [PATCH][AArch64] Adjust writeback in non-zero memset
Date: Wed, 14 Nov 2018 14:45:13 +0000
Message-ID: 
 <DB5PR08MB103082464EB664B31AFC5C5783C30@DB5PR08MB1030.eurprd08.prod.outlook.com>
References: 
 <DB5PR08MB103042C88641FD9D3A04C3F983C40@DB5PR08MB1030.eurprd08.prod.outlook.com>
In-Reply-To: 
 <DB5PR08MB103042C88641FD9D3A04C3F983C40@DB5PR08MB1030.eurprd08.prod.outlook.com>
authentication-results: spf=none (sender IP is )
	smtp.mailfrom=Wilco.Dijkstra@arm.com;
received-spf: None (protection.outlook.com: arm.com does not designate
	permitted sender hosts)
MIME-Version: 1.0

v2: Also bias dst in the zva_other code to avoid issues with zva sizes >= 256.

This fixes an ineffiency in the non-zero memset.  Delaying the writeback
until the end of the loop is slightly faster on some cores - this shows
~5% performance gain on Cortex-A53 when doing large non-zero memsets.

Tested against the GLIBC testsuite, OK for commit?

ChangeLog:
2018-11-14  Wilco Dijkstra  <wdijkstr@arm.com>

        * sysdeps/aarch64/memset.S (MEMSET): Improve non-zero memset loop.

diff --git a/sysdeps/aarch64/memset.S b/sysdeps/aarch64/memset.S
index 4a454593618f78e22c55520d56737fab5d8f63a4..9738cf5fd55a1d937fb3392cec46f37b4d5fb51d 100644
--- a/sysdeps/aarch64/memset.S
+++ b/sysdeps/aarch64/memset.S
@@ -89,10 +89,10 @@ L(set_long):
     b.eq    L(try_zva)
 L(no_zva):
     sub    count, dstend, dst    /* Count is 16 too large.  */
-    add    dst, dst, 16
+    sub    dst, dst, 16        /* Dst is biased by -32.  */
     sub    count, count, 64 + 16    /* Adjust count and bias for loop.  */
-1:    stp    q0, q0, [dst], 64
-    stp    q0, q0, [dst, -32]
+1:    stp    q0, q0, [dst, 32]
+    stp    q0, q0, [dst, 64]!
 L(tail64):
     subs    count, count, 64
     b.hi    1b
@@ -183,6 +183,7 @@ L(zva_other):
     subs    count, count, zva_len
     b.hs    3b
 4:    add    count, count, zva_len
+    sub    dst, dst, 32        /* Bias dst for tail loop.  */
     b    L(tail64)
 #endif