Re: [PATCH 0/4] swapcgroup(v2)

[Balbir Singh <balbir-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>]

KOSAKI Motohiro wrote:
>>> Have you seen any real world example of this? 
>> At the unsophisticated end, there are lots of (Fortran) HPC applications
>> with very large static array declarations but only "use" a small fraction
>> of that.  Those users know they only need a small fraction and are happy
>> to volunteer small physical memory limits that we (admins/queuing
>> systems) can apply.
>>
>> At the sophisticated end, the use of numerous large memory maps in
>> parallel HPC applications to gain visibility into other processes is
>> growing.  We have processes with VSZ > 400GB just because they have
>> 4GB maps into 127 other processes.  Their physical page use is of
>> the order 2GB.
> 
> Ah, agreed.
> Fujitsu HPC user said similar things ago.

OK, so this use case is HPC specific. I am not against the swap controller, but
overcommit can lead to problems if not controlled - such as OOM kill. The
virtual address space limit helps applications fail gracefully rather than swap
out excessively or OOM.

I suspect there'll be applications that swing both ways.

-- 
	Warm Regards,
	Balbir Singh
	Linux Technology Center
	IBM, ISTL