lkml.org 
[lkml]   [2013]   [Oct]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: Got iperf regression while intel_iommu is on, how to cut the cost of cache flushing
From
Date
On Thu, 2013-10-31 at 10:44 +0800, Ethan Zhao wrote:
> Hi guys,
>
> We got network I/O performance degradation with latest stable
> kernel and the be2net driver as compared to old kernel 3.0.6. later
> we found even compared to
> the same latest stable kernel but the INTEL_IOMMU set to 'n', still
> got very noticeable performance regression:
>
> Kernel : 3.11.x with CONFIG_INTEL_IOMMU_DEFAULT_ON not set
> Network driver : be2net
>
> Average Bandwidth for :
> 1.tcp-unidirectional test : 7908 Mbits/sec
> 2.tcp-unidirectional-parallel: 9400 Mbits/sec
> 3.tcp-bidirectonal test : 5464 Mbits/sec
>
> Kernel : 3.11.x with CONFIG_INTEL_IOMMU_DEFAULT_ON set
> Network driver : be2net

You might hit a known problem for some 10Gb links.

3.12-rc6 has the fixes already, and stable has some pending backports :

commit 95bd09eb27507691520d39ee1044d6ad831c1168
tcp: TSO packets automatic sizing

commit c9eeec26e32e087359160406f96e0949b3cc6f10
tcp: TSQ can use a dynamic limit

Please check that current 3.12-rc6 is OK ?





\
 
 \ /
  Last update: 2013-10-31 17:01    [W:0.113 / U:0.228 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site