From: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>
To: "Shaw, Jeffrey B" <jeffrey.b.shaw@intel.com>,
Olivier MATZ <olivier.matz@6wind.com>,
"dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH RFC 05/11] mbuf: merge physaddr and buf_len in a bitfield
Date: Wed, 14 May 2014 14:07:10 +0000 [thread overview]
Message-ID: <2601191342CEEE43887BDE71AB9772580EFA666D@IRSMSX105.ger.corp.intel.com> (raw)
In-Reply-To: <4032A54B6BB5F04B8C08B6CFF08C59285542085B@FMSMSX103.amr.corp.intel.com>
Hi Oliver,
Apart from performance impact, one more concern:
As I know, theoretical limit for PA on Intel is 52 bits.
I understand that these days no-one using more than 48 bits and it probably would stay like that for next few years.
Though if we'll occupy these (MAXPHYADDR - 48) bits now, it can become a potential problem in future.
After all the savings from that changes are not that big - only 2 bytes.
As I understand you already save extra 7 bytes with other proposed modifications of mbuf.
That's enough to add TSO related information into the mbuf.
So my suggestion would be to keep phys_addr 64bit long.
Thanks
Konstantin
-----Original Message-----
From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Shaw, Jeffrey B
Sent: Friday, May 09, 2014 5:12 PM
To: Olivier MATZ; dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH RFC 05/11] mbuf: merge physaddr and buf_len in a bitfield
I agree, we should wait for comments then test the performance when the patches have settled.
-----Original Message-----
From: Olivier MATZ [mailto:olivier.matz@6wind.com]
Sent: Friday, May 09, 2014 9:06 AM
To: Shaw, Jeffrey B; dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH RFC 05/11] mbuf: merge physaddr and buf_len in a bitfield
Hi Jeff,
Thank you for your comment.
On 05/09/2014 05:39 PM, Shaw, Jeffrey B wrote:
> have you tested this patch to see if there is a negative impact to
> performance?
Yes, but not with testpmd. I passed our internal non-regression performance tests and it shows no difference (or below the error margin), even with low overhead processing like forwarding whatever the number of cores I use.
> Wouldn't the processor have to mask the high bytes of the physical
> address when it is used, for example, to populate descriptors with
> buffer addresses? When compute bound, this could steal CPU cycles
> away from packet processing. I think we should understand the
> performance trade-off in order to save these 2 bytes.
I would naively say that the cost is negligible: accessing to the length is the same as before (it's a 16 bits field) and accessing the physical address is just a mask or a shift, which should not be very long on an Intel processor (1 cycle?). This is to be compared with the number of cycles per packet in io-fwd mode, which is probably around 150 or 200.
> It would be interesting to see how throughput is impacted when the
> workload is core-bound. This could be accomplished by running testpmd
> in io-fwd mode across 4x 10G ports.
I agree, this is something we could check. If you agree, let's first wait for some other comments and see if we find a consensus on the patches.
Regards,
Olivier
next prev parent reply other threads:[~2014-05-14 14:10 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-05-09 14:50 [dpdk-dev] [PATCH RFC 00/11] ixgbe/mbuf: add TSO support Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 01/11] igb/ixgbe: fix IP checksum calculation Olivier Matz
2014-05-15 10:40 ` Ananyev, Konstantin
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 02/11] mbuf: rename RTE_MBUF_SCATTER_GATHER into RTE_MBUF_REFCNT Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 03/11] mbuf: remove rte_ctrlmbuf Olivier Matz
2014-05-25 21:39 ` Gilmore, Walter E
2014-05-26 12:23 ` Olivier MATZ
2014-05-26 16:40 ` Dumitrescu, Cristian
2014-05-26 22:43 ` Neil Horman
2014-05-27 0:17 ` Stephen Hemminger
2014-05-28 9:45 ` Ananyev, Konstantin
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 04/11] mbuf: remove the rte_pktmbuf structure Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 05/11] mbuf: merge physaddr and buf_len in a bitfield Olivier Matz
2014-05-09 15:39 ` Shaw, Jeffrey B
2014-05-09 16:06 ` Olivier MATZ
2014-05-09 16:11 ` Shaw, Jeffrey B
2014-05-14 14:07 ` Ananyev, Konstantin [this message]
2014-05-15 9:53 ` Olivier MATZ
2014-05-19 7:27 ` Olivier MATZ
2014-05-19 8:25 ` Richardson, Bruce
2014-05-19 9:30 ` Olivier MATZ
2014-05-19 9:57 ` Richardson, Bruce
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 06/11] mbuf: replace data pointer by an offset Olivier Matz
2014-05-12 14:12 ` Thomas Monjalon
2014-05-12 14:36 ` Venkatesan, Venky
2014-05-12 14:41 ` Neil Horman
2014-05-12 15:07 ` Olivier MATZ
2014-05-12 15:59 ` Stephen Hemminger
2014-05-12 16:13 ` Olivier MATZ
2014-05-12 17:13 ` Stephen Hemminger
2014-05-13 13:29 ` Olivier MATZ
2014-05-12 16:06 ` Venkatesan, Venky
2014-05-12 18:39 ` Neil Horman
2014-05-13 13:54 ` Venkatesan, Venky
2014-05-13 14:09 ` Thomas Monjalon
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 07/11] mbuf: add functions to get the name of an ol_flag Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 08/11] mbuf: change ol_flags to 32 bits Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 09/11] mbuf: rename vlan_macip_len in hw_offload and increase its size Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 10/11] testpmd: modify source address to validate checksum calculation Olivier Matz
2014-05-09 14:50 ` [dpdk-dev] [PATCH RFC 11/11] ixgbe/mbuf: add TSO support Olivier Matz
2014-05-12 14:30 ` Thomas Monjalon
2014-05-15 15:09 ` Ananyev, Konstantin
2014-05-15 15:39 ` Olivier MATZ
2014-05-15 16:30 ` Ananyev, Konstantin
2014-05-16 12:11 ` Olivier MATZ
2014-05-16 17:01 ` Ananyev, Konstantin
2014-05-19 12:32 ` Thomas Monjalon
2014-05-09 17:04 ` [dpdk-dev] [PATCH RFC 00/11] " Stephen Hemminger
2014-05-09 21:49 ` Olivier MATZ
2014-05-10 0:39 ` Stephen Hemminger
2014-05-19 12:47 ` Thomas Monjalon
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=2601191342CEEE43887BDE71AB9772580EFA666D@IRSMSX105.ger.corp.intel.com \
--to=konstantin.ananyev@intel.com \
--cc=dev@dpdk.org \
--cc=jeffrey.b.shaw@intel.com \
--cc=olivier.matz@6wind.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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).