Hello,
I'm trying to figure out how to use Glib's faster memory
management for my Glib-styled skip list implementation. Right now I'm
calling g_malloc() and g_free() for each individual element.
I noticed that GMemChunk only allows for equally-sized elements.
This is a problem for skip lists, because your elements all have different
sizes (a size for every Node in the set 1..max_height). This is because
skip lists use a variable-sized array as the last element of the Node
struct.
Any suggestions?
One I idea I had was to have one GMemChunk for every possible
level in the skip list. max_height is usually 16 or less, but in
practical terms should probably never be more than 32. So generally
that'd be 16 different GMemChunks for a single skip list. When I allocate
a new Node, I'd just use the GMemChuck for that level. Does this make
sense? Any other suggestions?
Also, what is the point of a GAllocator?
Thank You,
Derek Simkowiak
[EMAIL PROTECTED]
--
To unsubscribe: mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
