[zeromq-dev] Memory management

Martin Sustrik sustrik at fastmq.com
Fri May 15 10:09:51 CEST 2009


Hi Steven,

>         The slab allocates blocks of memory aligned as preferred by the
>         underlying architecture.  Using a ring buffer and arbritarily
>         slicing up blocks can make paging worse as multiple requests
>         might be needed for one buffer.  One solution is to implement a
>         hybrid design that allocates page size and alignment friendly
>         blocks of memory.
> 
>         wide stack.png
> 
> 
>         This could still be implemented in a stack or a ring buffer,
>         reference counting can be used for individual slices of a chunk.
> 
>     Page Cache Friendly.
> 
>     Sorry, I'm lost here. What do you mean by "multiple requests might
>     be needed for one buffer"?
> 
> 
> 
> If you have an 4KB page and a 4KB object but 1KB is in one page, 3KB in 
> the next page, you end up with twice the overhead of when the object 
> neatly fits in one page.

Ok, I see. However, if tx/rx are stored in a single preallocated chunk 
of memory, it ceases to be a problem as all the memory allocation is 
done in advance, no?

Martin



More information about the zeromq-dev mailing list