From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:46842)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1bz8nQ-0001k4-Ak
	for qemu-devel@nongnu.org; Tue, 25 Oct 2016 16:58:09 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1bz8nN-00077c-9w
	for qemu-devel@nongnu.org; Tue, 25 Oct 2016 16:58:08 -0400
Received: from mail-wm0-x242.google.com ([2a00:1450:400c:c09::242]:40709)
	by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16)
	(Exim 4.71) (envelope-from <paolo.bonzini@gmail.com>)
	id 1bz8nN-00077V-35
	for qemu-devel@nongnu.org; Tue, 25 Oct 2016 16:58:05 -0400
Received: by mail-wm0-x242.google.com with SMTP id b80so2060741wme.7
	for <qemu-devel@nongnu.org>; Tue, 25 Oct 2016 13:58:04 -0700 (PDT)
Sender: Paolo Bonzini <paolo.bonzini@gmail.com>
References: <20161025153507.27110-1-bobby.prani@gmail.com>
	<5b91345d-fea6-8d98-69a9-eda1e256c3e1@redhat.com>
	<CAJhHMCCgKmT2=Vpzhi+pF=5b3Mjojq4jjkJrWwLiMmY-2yniPw@mail.gmail.com>
	<54c9a235-4ab4-1731-b076-116ecde91442@redhat.com>
	<87shrkckfv.fsf@gmail.com>
	<5190ece5-7167-dbe1-5b5a-38247cba5be3@redhat.com>
	<CAJhHMCD7zsgfXpVWE7S9CaN3TOz9KR-iiTX2kFcHMGqO5ggkCg@mail.gmail.com>
	<20161025204515.GA22860@flamenco>
From: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <7012eb7a-929a-519b-cf0d-48c62d594589@redhat.com>
Date: Tue, 25 Oct 2016 22:58:00 +0200
MIME-Version: 1.0
In-Reply-To: <20161025204515.GA22860@flamenco>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [RFC PATCH] qht: Align sequence lock to cache line
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel/>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: "Emilio G. Cota" <cota@braap.org>, Pranith Kumar <bobby.prani@gmail.com>
Cc: Richard Henderson <rth@twiddle.net>, =?UTF-8?Q?Alex_Benn=c3=a9e?= <alex.bennee@linaro.org>, "open list:All patches CC here" <qemu-devel@nongnu.org>, Markus Armbruster <armbru@redhat.com>



On 25/10/2016 22:45, Emilio G. Cota wrote:
> On Tue, Oct 25, 2016 at 16:35:48 -0400, Pranith Kumar wrote:
>> On Tue, Oct 25, 2016 at 4:02 PM, Paolo Bonzini <pbonzini@redhat.com> wrote:
>>>
>>>
>>>> I've written a patch (see below) to take the per-bucket sequence locks.
>>>
>>> What's the performance like?
>>>
>>
>> Applying only this patch, the perf numbers are similar to the 128
>> cache line alignment you suggested.
> 
> That makes sense. Having a single seqlock per bucket is simple and fast;
> note that bucket chains should be very short (we use good hashing and
> automatic resize for this purpose).

But why do we get such worse performance in the 100% reader case?  (And
even more puzzling, why does Pranith's original patch improve
performance instead of causing more cache misses?)

Thanks,

Paolo