DPDK usage discussions
 help / color / mirror / Atom feed
From: 宋捷 <songj@zctt.com>
To: "'Stephen Hemminger'" <stephen@networkplumber.org>
Cc: "'users'" <users@dpdk.org>,
	"'Van Haaren, Harry'" <harry.van.haaren@intel.com>
Subject: [dpdk-users] 答复: How can create a mempool more than 20GB?
Date: Wed, 11 Sep 2019 10:21:42 +0800	[thread overview]
Message-ID: <001c01d56847$a6ff17c0$f4fd4740$@zctt.com> (raw)
In-Reply-To: <20190910183621.50d50ddb@xps13>

Hi Stephen,

Expert,you are correct, 
I update the default hugepage size from 2M to 1G, it's worked.

Thanks you very much!


Jie



-----邮件原件-----
发件人: Stephen Hemminger [mailto:stephen@networkplumber.org] 
发送时间: 2019年9月11日 1:36
收件人: 宋捷 <songj@zctt.com>
抄送: 'users' <users@dpdk.org>; 'Van Haaren, Harry' <harry.van.haaren@intel.com>
主题: Re: How can create a mempool more than 20GB?

On Tue, 10 Sep 2019 19:02:37 +0800
宋捷 <songj@zctt.com> wrote:

> Hi All,
> 
>  
> 
>        I assigned 40GB hugepages for each node.
> 
> 
> 
>  
> 
>        I need a huge mempool buffer (20GB) for save received packets, 
> I create a mempool as follow:
> 
> rte_pktmbuf_pool_create("rx_dump_pool",  8*1000*1000, 0, 0, 
> RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id() );
> 
>  
> 
> RTE_MBUF_DEFAULT_BUF_SIZE is 2048+128 , I think 
> rte_pktmbuf_pool_create should create 8*1000*1000*(2176)≈16G buffer,
> 
> But failed with below prompt:
> 
> EAL: eal_memalloc_alloc_seg_bulk(): couldn't find suitable memseg_list
> 
> EAL: eal_memalloc_alloc_seg_bulk(): couldn't find suitable memseg_list
> 
> EAL: Error - exiting with code: 1
> 
>  
> 
> I try to reduce the size, the mempool can be created successfully as follow:
> 
> rte_pktmbuf_pool_create("rx_dump_pool", 5000000, 0, 0, 
> RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id() );
> 
>  
> 
>  
> 
> Does DPDK mempool has a limited size ?  I have no ideal for this L
> 
>  

You are going to need 1G huge pages on this, and a system with very large amount of memory. To setup 1G huge pages requires settings on kernel command line.

Also expect slow startup times with that much memory.






      reply	other threads:[~2019-09-11  2:22 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <5d7782d8.1c69fb81.26d5e.d195SMTPIN_ADDED_BROKEN@mx.google.com>
2019-09-10 17:36 ` [dpdk-users] " Stephen Hemminger
2019-09-11  2:21   ` 宋捷 [this message]

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='001c01d56847$a6ff17c0$f4fd4740$@zctt.com' \
    --to=songj@zctt.com \
    --cc=harry.van.haaren@intel.com \
    --cc=stephen@networkplumber.org \
    --cc=users@dpdk.org \
    /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).