[PATCHv4 0/7] zswap: compressed swap caching

Seth Jennings sjenning at linux.vnet.ibm.com
Mon Feb 4 14:56:18 UTC 2013


On 02/02/2013 06:17 PM, Simon Jeons wrote:
> On Fri, 2013-02-01 at 09:13 -0600, Seth Jennings wrote:
>> On 01/31/2013 07:39 PM, Simon Jeons wrote:
>>> Hi Seth,
>>> On Tue, 2013-01-29 at 15:40 -0600, Seth Jennings wrote:
>> <snip>
>>>> Performance, Kernel Building:
>>>>
>>>> Setup
>>>> ========
>>>> Gentoo w/ kernel v3.7-rc7
>>>> Quad-core i5-2500 @ 3.3GHz
>>>> 512MB DDR3 1600MHz (limited with mem=512m on boot)
>>>> Filesystem and swap on 80GB HDD (about 58MB/s with hdparm -t)
>>>> majflt are major page faults reported by the time command
>>>> pswpin/out is the delta of pswpin/out from /proc/vmstat before and after
>>>> then make -jN
>>>>
>>>> Summary
>>>> ========
>>>> * Zswap reduces I/O and improves performance at all swap pressure levels.
>>>>
>>>> * Under heavy swaping at 24 threads, zswap reduced I/O by 76%, saving
>>>>   over 1.5GB of I/O, and cut runtime in half.
>>>
>>> How to get your benchmark?
>>
>> It's just kernel building.  So "make" :)
>>
>> I intentionally choose this workload so people wouldn't have to jump
>> through hoops to replicate the results.
> 
> Thanks for your clarify. zswap is belong to in-kernel compression,
> correct? then what's the difference between in-memory compression and
> in-kernel compression?

Dan made this distinction (possibly unintentionally) in his email.
For me "in-kernel" means "done by the kernel" and "in-memory" means
"stored in RAM".  So it is compression of RAM pages done by the kernel
and stored back in RAM.

I sum up the functionality of zswap as "compressed swap caching",
which I think better conveys what zswap does exactly.

Seth




More information about the devel mailing list