Hi Christian, I did some runtime analysis of the code and the difference between the two versions comes from the extra time libart spends in the art_svp_render_aa_iter_step and art_svp_intersector routines. My guess is that this happens because in the first example renders the horyzontal line right on a 'pixel' whereas the second example does render it between two pixles ... so what you get is the best and the worst case rendering compared ... you get a similar effect when you modify the first example to draw a LINE1.5 ... not as extreem but in the same direction ...
I'm afraid there is not much I can do about this just now ... but I plan to move to cairo for the 1.3 series and I hope that the cairo library will have better performance than libart ... cheers tobi > On Thu, 2005-09-22 at 04:20 +0200, Alex van den Bogaerdt wrote: > > > However, the following did work: > > "CDEF:val2=val1,1024,1024,*,/,FLOOR" > > About 0.6 to 0.65 seconds is not bad, certainly not when compared to 12. > > > Christian, please try that yourself. Of course you need to graph val2, not > > val1. > > You could save some extra cpu cycles by changing "1024,1024,*" into 1048576 > > or > > 1024000 (whichever is appropriate for you). > > > > Tobi, seems to me a function such as apply_gridfit() may be the cause? > > Alex, > > Far out; I get the same results you do, using my old version of rrdtool! > Graphing val2 from your CDEF works much better than graphing val1 from > the DEF. It runs in virtually the same length of time as does a graph > of one of the "fast" columns. This seems to be a fine workaround, once > I adjust my Y-axis label. I like graphing in MB (maybe even GB) better > than graphing in KB anyway. Thanks!!! > > Tobi, > > What is your take on this? > > -- ______ __ _ /_ __/_ / / (_) Oetiker @ ISG.EE, ETL F24.2, ETH, CH-8092 Zurich / // _ \/ _ \/ / System Manager, Time Lord, Coder, Designer, Coach /_/ \.__/_.__/_/ http://people.ee.ethz.ch/oetiker +41(0)44-632-5286 -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://lists.ee.ethz.ch/rrd-users WebAdmin http://lists.ee.ethz.ch/lsg2.cgi
