Re: [PATCH v3 2/7] locking/pvqspinlock: Add pending bit support

From: Waiman Long <waiman.long@hp.com>
To: Davidlohr Bueso <dave@stgolabs.net>
Cc: Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	"H. Peter Anvin" <hpa@zytor.com>,
	x86@kernel.org, linux-kernel@vger.kernel.org,
	Scott J Norton <scott.norton@hp.com>,
	Douglas Hatch <doug.hatch@hp.com>
Subject: Re: [PATCH v3 2/7] locking/pvqspinlock: Add pending bit support
Date: Mon, 27 Jul 2015 13:11:23 -0400	[thread overview]
Message-ID: <55B6663B.2060707@hp.com> (raw)
In-Reply-To: <1437952192.25997.3.camel@stgolabs.net>

On 07/26/2015 07:09 PM, Davidlohr Bueso wrote:
> On Wed, 2015-07-22 at 16:12 -0400, Waiman Long wrote:
>> Using a locking microbenchmark on the same system, the locking
>> rates in (kops/s) were:
>>
>>    Threads		Rate w/o patch	Rate with patch
>>    -------		--------------	---------------
>>    2 (same socket)	  6,515,265	  7,077,476
>>    2 (diff sockets)	  2,967,145	  4,353,851
> At this level stddev would be nice to have, as these sort of things have
> a lot of variation.
>
> Out of general interest, are these microbenchmarks available? Would you
> care to add something like that to perf-bench? I assume they would
> probably be x86 specific.

Ingo has asked similar question and he suggested doing an user-space 
implementation in perf bench. I will take a further look at that when I 
have time.

Cheers,
Longman