DragonFly BSD
DragonFly kernel List (threaded) for 2003-09
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: SLAB allocator now the default.


From: David Leimbach <leimy2k@xxxxxxx>
Date: Sun, 28 Sep 2003 11:35:22 -0500


On Sep 28, 2003, at 9:22 AM, Hiten Pandya wrote:


Matthew Dillon wrote:

The slab allocator is now the default. The NO_KMEM_MAP and
USE_SLAB_ALLOCATOR options have been removed.
USE_KMEM_MAP and NO_SLAB_ALLOCATOR options have temporarily been added,
but these too will be removed (along with the old malloc and kmem_map
code) in about a week.



Hiten,


Matt,

	I haven't looked at the Slab-alloc code recently, but I am
	wondering if you are planning to remove the power-of-2 alignment
	stuff out of malloc()?


Doesn't power of two alignment make it easier on some CPUs internally to access
data? If a person wanted to access a word overlapping two word's alignment would it
not take two bus accesses to get that data [or two cache reads... what have you]


Do the benefits of not having power of two alignment outweigh this?



	From the many papers I have been reading in the last couple of
	weeks, it seems that power-of-2 alignment seems to cause a lot
	of performance degradation in SMP cases, and cache issues.

A good example of this demonstration is in Bonwick's papers.

	I understand this would be quite a lengthy and tedious procedure
	but it is something that I have been wondering.

Regards,

--
Hiten Pandya
hmp@xxxxxxxxxxxxx





[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]