Alex Rousskov wrote:
> We may need another parameter, perhaps internal, that determined chunk
> size, perhaps depending on the pool type. When the sum of all chunks
> exceeds the specified limit, we revert to malloc().
>
> Am I missing the problem you are concerned with?
Yes. We are concerned about idle memory, not memory in use. Memory in
use are supposed to be there, idle memory not.
Doing fallback to malloc() when usage is high won't solve the problem,
implementing a smart chunked allocator will except for some special
conditions (fragmentation). The larger the chunks, the bigger the
problem.
/Henrik
Received on Wed Mar 28 2001 - 15:32:49 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:13:41 MST