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


	We've got an application that makes heavy use of hash tables,
but the tables sometimes get very large (> 10000 entries).  This reduces
performance dramatically because of the paging characteristics.  So, we
are wondering
	1) Would it behoove us to use btree tables, rather than hash
	2) Does anyone have a btree implementation that they are willing
to share.
	3) Does anyone have any other suggestions for improving

	The tables we make come in all sizes from 5 entries to 20000 entries,
but I'm not sure of the distributions.  There are probably ways to
estimate the potential size of the table before build it, but we haven't
worried about it much.  Right now we just rely on the growing table
mechanism that Symbolics provides.  Does anyone know if Symbolics has
any plans to include sophisticated facilities (such as btrees) in the
(already extremely useful) generalized table facility.

  -- Paul Vaughan