CLIM mail archive
[Prev][Next][Index][Thread]

Re: Performance analysis of Clim incremental redisplay

> bruce
> miller@cam.nist.gov  

writes:

> Just to clarify a minor point with the tenses:  The table displays the
> numbers for simple full display _using_ your optimized code?  Or you were
> _later_ able to optimize?

I was later able to optimize -- all the numbers for that table are
unoptimized code.  It seemed appropriate to compare apples to apples.

I've now added the optimized timings as the fourth line in the table,
so yes, the cross-over point changes:

Graph size:			100	200	400	800
-----------------------------------------------------------
Incremental redisplay:		8	26	97	351	(a)
Updating-output full display:	21	57	180	595	(b)
Simple full display:		15	32	86	234	(c)
Optimized simple full display:  5	11	23	50	(d)

Lines (a,b,c) are all with my code unoptimized.  I haven't yet
measured lines (a) or (b) with my code optimized.  My assumption is
that the difference between lines (b) and (c) represents work that
CLIM is going to do regardless of how fast my code runs.

Peter

0,,
Main Index | Thread Index