lkml.org 
[lkml]   [2006]   [Jan]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [Lhms-devel] Re: [PATCH 0/5] Reducing fragmentation using zones
Mel Gorman wrote:>
> What sort of tests would you suggest?
> The tests I have been running to date are
>
> "kbuild + aim9" for regression testing
>
> "updatedb + 7 -j1 kernel compiles + highorder allocation" for seeing how
> easy it was to reclaim contiguous blocks
>
> What tests could be run that would be representative of real-world
> workloads?
>

1. Using 1000+ processes(threads) at once
2. heavy network load.
3. running NFS
is maybe good.

>>>> And, for people who want to remove range of memory, list-based approach
>>>> will
>>>> need some other hook and its flexibility is of no use.
>>>> (If list-based approach goes, I or someone will do.)
>>>>
>>> Will do what?
>>>
>> add kernelcore= boot option and so on :)
>> As you say, "In an ideal world, we would have both".
>>
>
> List-based was frowned at for adding complexity to the main path so we may
> not get list-based built on top of zone based even though it is certinatly
> possible. One reason to do zone-based was to do a comparison between them
> in terms of complexity. Hopefully, Nick Piggin (as the first big objector
> to the list-based approach) will make some sort of comment on what he
> thinks of zone-based in comparison to list-based.
>
I think there is another point.

what I concern about is Linus's word ,this:
> My point is that regardless of what you _want_, defragmentation is
> _useless_. It's useless simply because for big areas it is so expensive as
> to be impractical.

You should make your own answer for this before posting.

From the old threads (very long!), I think one of the point was :
To use hugepages, sysadmin can specifies what he wants at boot time.
This guarantees 100% allocation of needed huge pages.
Why memhotplug cannot specifies "how much they can remove" before booting.
This will guaranntee 100% memory hotremove.

I think hugetlb and memory hotplug cannot be good reason for defragment.

Finding the reason for defragment is good.
Unfortunately, I don't know the cases of memory allocation failure
because of fragmentation with recent kernel.

-- Kame


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2006-01-20 11:43    [W:0.095 / U:0.048 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site