Received-SPF: pass (google.com: domain of linux-crypto+bounces-779-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Wed, 13 Dec 2023 11:01:06 +0800
From: Wang Jinchao <wangjinchao@xfusion.com>
To: Tim Chen <tim.c.chen@linux.intel.com>
CC: Herbert Xu <herbert@gondor.apana.org.au>, "David S. Miller"
	<davem@davemloft.net>, Steffen Klassert <steffen.klassert@secunet.com>,
	Daniel Jordan <daniel.m.jordan@oracle.com>, <linux-crypto@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>, <stone.xulei@xfusion.com>
Subject: Re: [PATCH] crypto: tcrypt - add script tcrypt_speed_compare.py
Message-ID: <ZXkecmQGYT81R17C@fedora>
References: <202312101758+0800-wangjinchao@xfusion.com>
 <118d0b232f7daffc56db814f90dba7bad266c5ab.camel@linux.intel.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <118d0b232f7daffc56db814f90dba7bad266c5ab.camel@linux.intel.com>

On Tue, Dec 12, 2023 at 01:56:56PM -0800, Tim Chen wrote:
> On Sun, 2023-12-10 at 18:19 +0800, WangJinchao wrote:
> > Create a script for comparing tcrypt speed test logs.
> > The script will systematically analyze differences item
> > by item and provide a summary (average).
> > This tool is useful for evaluating the stability of
> > cryptographic module algorithms and assisting with
> > performance optimization.
> 
> I have found that for such comparison, the stability is
> dependent on whether we allow the frequency to
> float or we pin the frequency.  So in the past when
> I use tcrypt, sometimes I have
> to pin the frequency of CPU for stable results.
> 
> One suggestion I have is for for you to also dump the
> frequency governor and P-state info so we know
> for the runs being compared, whether they are running
> with the same CPU frequency.
> 
> Tim 
> 
Thank you for the feedback. This information is valuable for stability testing
and performance optimization.

However, I am uncertain about how to dump P-state information, or I believe that
the script is unable to do so. The reasons are as follows:

1. The primary purpose of this script is to compare tcrypt logs, and it is
executed after the completion of the tcrypt tests. Consequently, it cannot
dump P-state information during tcrypt's runtime.

2. In virtualized environments, there is no available information in the
`/sys/devices/system/cpu/cpufreq` directory pertaining to P-state details.

Am I correct in my understanding?
I am considering documenting your suggestion in the script's comments.
What are your thoughts?
>