Re: RFC: m_tag pool cache

[Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost>]

Hi,

On 2016/02/19 19:05, Roy Marples wrote:
> On 19/02/2016 09:26, Kengo NAKAHARA wrote:
>> According to my measurements of using DTrace, there is a certain
>> overhead between kern_malloc() and pool_cache_get(). Here is each
>> function's average turnaround times in my measurement.
>>
>>     m_tag_get            | 124205 [ns]
>>     kern_malloc          | 108007 [ns]
>>     kmem_intr_alloc      |  89636 [ns]
>>     pool_cache_get_paddr |  65942 [ns]
> 
> ...
> 
>> Here is the patch to introduce mbuf pool cache.
>>     http://www.netbsd.org/~knakahara/mtag-pool-cache/mtag-pool-cache.patch
>>
>> Could you comment this patch? Any comments are welcome.
> 
> And the equivalent numbers with the patch?
> As I assume the patch is an optimisation, a suitable speed bump is
> expected :)

In my working case(which is ALTQ refactoring), above patch and some
optimization makes about 10% performance improvement in the *best* case.
It may be not so effective in other cases such as vlan, pf, and ipsec...


Thanks,

-- 
//////////////////////////////////////////////////////////////////////
Internet Initiative Japan Inc.

Device Engineering Section,
Core Product Development Department,
Product Division,
Technology Unit

Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost>