Re: [ECOS] how to prove ecos2.0 's Memory managment 's performence?

[Bart Veer <bartv at ecoscentric dot com>]

>>>>> " " == =?gb2312?B?1cUgwcE=?=  <gb2312> writes:

     > Hi,everyone.
     >    I have peeked implementation of memory management in
     > ecos2.0,eg.Doug Lea's malloc implementation.but i wonder that
     > how to prove this algorism's performance, both malloc
     > operation's speed and memory fragments? could you give some
     > crack provement?

     >  thank you very much!

Memory allocation packages involve various trade-offs: cpu cycles
consumed by malloc; ditto for free; efficient use of memory, i.e. how
much heap space gets wasted because the package allocates more than is
absolutely necessary; resistance to fragmentation; ...

For general-purpose use, it is believed that the Doug Lea
implementation is about as good as you can get. If you have special
needs, e.g. malloc() must be as fast as possible even at the expense
of efficient memory usage, than you can try some of the other
allocators provided by eCos, e.g. the fixed block allocator.

Bart

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss