DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>
To: Neil Horman <nhorman@tuxdriver.com>,
	"Richardson, Bruce" <bruce.richardson@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH v2 1/4] app/test: unit test for rx and tx cycles/packet
Date: Tue, 21 Oct 2014 22:43:45 +0000
Message-ID: <2601191342CEEE43887BDE71AB9772582139510A@IRSMSX105.ger.corp.intel.com> (raw)
In-Reply-To: <20141021133738.GC12795@hmsreliant.think-freely.org>



> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Neil Horman
> Sent: Tuesday, October 21, 2014 2:38 PM
> To: Richardson, Bruce
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 1/4] app/test: unit test for rx and tx cycles/packet
> 
> On Tue, Oct 21, 2014 at 10:43:03AM +0000, Richardson, Bruce wrote:
> >
> >
> > > -----Original Message-----
> > > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Neil Horman
> > > Sent: Tuesday, October 21, 2014 11:33 AM
> > > To: Liang, Cunming
> > > Cc: dev@dpdk.org
> > > Subject: Re: [dpdk-dev] [PATCH v2 1/4] app/test: unit test for rx and tx
> > > cycles/packet
> > >
> > > > >
> > > > > > +	if (count == 0)
> > > > > > +		return -1;
> > > > > > +
> > > > > > +	printf("%lu packet, %lu drop, %lu idle\n", count, drop, idle);
> > > > > > +	printf("Result: %ld cycles per packet\n", (cur_tsc - prev_tsc) / count);
> > > > > > +
> > > > > Bad math here.  Theres no guarantee that the tsc hasn't wrapped (potentially
> > > > > more than once) depending on your test length.  you need to check the tsc
> > > before
> > > > > and after each burst and record an average of deltas instead, accounting in
> > > each
> > > > > instance for the possibility of wrap.
> > > > [Liang, Cunming] I'm not sure catch your point correctly.
> > > > I think both cur_tsc and prev_tsc are 64 bits width.
> > > > For 3GHz, I think it won't wrapped so quick.
> > > > As it's uint64_t, so even get wrapped, the delta should still be correct.
> > > But theres no guarantee that the tsc starts at zero when you begin your test.
> > > The system may have been up for a long time and near wrapping already.
> > > Regardless, you need to account for the possibility that cur_tsc is smaller
> > > than prev_tsc, or this breaks.
> > >
> >
> > The tsc. is 64-bit and so only wraps around every couple of hundred years or so on a 2GHz machine, so I don't think it's necessary to
> handle that case.
> >
> But that presumes that no one has written the TSC via IA32_TIME_STAMP_COUNTER.

Then the test app would just print the wrong number :)
I suppose user will just repeat the test.
Again, imagine someone wrmsr(TSC), but set the value just much bigger than current.
Then your statistics will be incorrect, but you have no wait to figure that out. 
I think that trying to handle all such hypothetical situations is a pure overkill.

> Assuming that something will never wrap just seems like bad practice here.  We
> should have a general purpose macro to handle wrapping counters like this, if
> not for this case specficially, then in general.
> 
> Neil
> 
> > /Bruce
> >

  reply	other threads:[~2014-10-21 22:35 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-08-25  6:12 [dpdk-dev] [PATCH 0/5] app/test: unit test to measure cycles per packet Cunming Liang
2014-08-25  6:12 ` [dpdk-dev] [PATCH 1/5] app/test: unit test for rx and tx cycles/packet Cunming Liang
2014-08-25  6:12 ` [dpdk-dev] [PATCH 2/5] app/test: measure standalone rx or " Cunming Liang
2014-08-25  6:12 ` [dpdk-dev] [PATCH 3/5] ixgbe/vpmd: add fix to store unaligned mbuf point array Cunming Liang
2014-08-25  6:12 ` [dpdk-dev] [PATCH 4/5] app/test: add unit test to measure RX burst cycles Cunming Liang
2014-08-25  6:12 ` [dpdk-dev] [PATCH 5/5] app/test: allow to create packets in different sizes Cunming Liang
2014-09-03  1:46 ` [dpdk-dev] [PATCH 0/5] app/test: unit test to measure cycles per packet Zhan, Zhaochen
2014-10-10 12:29 ` [dpdk-dev] [PATCH v2 0/4] " Cunming Liang
2014-10-10 12:29   ` [dpdk-dev] [PATCH v2 1/4] app/test: unit test for rx and tx cycles/packet Cunming Liang
2014-10-10 17:52     ` Neil Horman
2014-10-12 11:10       ` Liang, Cunming
     [not found]         ` <E115CCD9D858EF4F90C690B0DCB4D8972262B720@IRSMSX108.ger.corp.intel.com>
2014-10-14  0:54           ` Liang, Cunming
2014-10-21 10:33         ` Neil Horman
2014-10-21 10:43           ` Richardson, Bruce
2014-10-21 13:37             ` Neil Horman
2014-10-21 22:43               ` Ananyev, Konstantin [this message]
2014-10-21 13:17           ` Liang, Cunming
2014-10-22 14:03             ` Neil Horman
2014-10-22 14:48               ` Liang, Cunming
2014-10-22 14:53               ` Ananyev, Konstantin
2014-10-22 15:09                 ` Richardson, Bruce
2014-10-24  3:06                   ` Liang, Cunming
2014-10-10 12:29   ` [dpdk-dev] [PATCH v2 2/4] app/test: measure standalone rx or " Cunming Liang
2014-10-10 12:30   ` [dpdk-dev] [PATCH v2 3/4] app/test: add unit test to measure RX burst cycles Cunming Liang
2014-10-10 12:30   ` [dpdk-dev] [PATCH v2 4/4] app/test: allow to create packets in different sizes Cunming Liang
2014-10-20  8:13   ` [dpdk-dev] [PATCH v3 0/2] app/test: unit test to measure cycles per packet Cunming Liang
2014-10-20  8:13     ` [dpdk-dev] [PATCH v3 1/2] app/test: allow to create packets in different sizes Cunming Liang
2014-10-20  8:13     ` [dpdk-dev] [PATCH v3 2/2] app/test: measure the cost of rx/tx routines by cycle number Cunming Liang
2014-10-21  2:40     ` [dpdk-dev] [PATCH v3 0/2] app/test: unit test to measure cycles per packet Liu, Yong
2014-10-24  5:39     ` [dpdk-dev] [PATCH v4 0/3] " Cunming Liang
2014-10-24  5:39       ` [dpdk-dev] [PATCH v4 1/3] app/test: allow to create packets in different sizes Cunming Liang
2014-10-24  5:39       ` [dpdk-dev] [PATCH v4 2/3] app/test: measure the cost of rx/tx routines by cycle number Cunming Liang
2014-10-24  5:39       ` [dpdk-dev] [PATCH v4 3/3] ethdev: fix wrong error return refer to API definition Cunming Liang
2014-10-24  5:57       ` [dpdk-dev] [PATCH v5 0/3] app/test: unit test to measure cycles per packet Cunming Liang
2014-10-24  5:58         ` [dpdk-dev] [PATCH v5 1/3] app/test: allow to create packets in different sizes Cunming Liang
2014-10-24  5:58         ` [dpdk-dev] [PATCH v5 2/3] app/test: measure the cost of rx/tx routines by cycle number Cunming Liang
2014-10-24  5:58         ` [dpdk-dev] [PATCH v5 3/3] ethdev: fix wrong error return refere to API definition Cunming Liang
2014-10-27  1:20         ` [dpdk-dev] [PATCH v6 0/3] app/test: unit test to measure cycles per packet Cunming Liang
2014-10-27  1:20           ` [dpdk-dev] [PATCH v6 1/3] app/test: allow to create packets in different sizes Cunming Liang
2014-10-27  1:20           ` [dpdk-dev] [PATCH v6 2/3] app/test: measure the cost of rx/tx routines by cycle number Cunming Liang
2014-11-11 23:28             ` Thomas Monjalon
2014-11-12  6:32               ` Liang, Cunming
2014-10-27  1:20           ` [dpdk-dev] [PATCH v6 3/3] ethdev: fix wrong error return refere to API definition Cunming Liang
2014-10-27 16:03             ` Ananyev, Konstantin
2014-10-27  1:45           ` [dpdk-dev] [PATCH v6 0/3] app/test: unit test to measure cycles per packet Liu, Yong
2014-10-29  5:06           ` Liang, Cunming
2014-11-12  6:24           ` [dpdk-dev] [PATCH v7 0/7] " Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 1/7] app/test: allow to create packets in different sizes Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 2/7] ixgbe:clean scattered_rx configure in dev_stop Cunming Liang
2014-11-12  7:53               ` Thomas Monjalon
2014-11-12  8:21                 ` Liang, Cunming
2014-11-12  9:24                   ` Thomas Monjalon
2014-11-12 10:29                     ` Liang, Cunming
2014-11-12 10:32                       ` Thomas Monjalon
2014-11-12 10:42                         ` Liang, Cunming
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 3/7] ether: new API to format eth_addr in string Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 4/7] app/testpmd: cleanup eth_addr print Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 5/7] examples: " Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 6/7] app/test: measure the cost of rx/tx routines by cycle number Cunming Liang
2014-11-12  6:24             ` [dpdk-dev] [PATCH v7 7/7] ethdev: fix wrong error return refere to API definition Cunming Liang
2014-11-12 23:50             ` [dpdk-dev] [PATCH v7 0/7] app/test: unit test to measure cycles per packet Thomas Monjalon
2014-10-24  5:59       ` [dpdk-dev] [PATCH v4 0/3] " Liang, Cunming
     [not found]       ` <1414130090-17910-1-git-send-email-y>
     [not found]         ` <1414130090-17910-4-git-send-email-y>
2014-10-24 11:04           ` [dpdk-dev] [PATCH v5 3/3] ethdev: fix wrong error return refere to API definition Ananyev, Konstantin
2014-10-27  0:58             ` Liang, Cunming
2014-10-28 12:21   ` [dpdk-dev] [PATCH v2 0/4] app/test: unit test to measure cycles per packet Neil Horman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=2601191342CEEE43887BDE71AB9772582139510A@IRSMSX105.ger.corp.intel.com \
    --to=konstantin.ananyev@intel.com \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=nhorman@tuxdriver.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

DPDK patches and discussions

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/dev/0 dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dev dev/ https://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.dev


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git