[<<] [<] Page 1 of 1 [>] [>>] | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
slowness of linedraw and < 100Mhz machines
From: "Aaron J. Grier" ####@####.#### Date: 17 Feb 2003 21:13:32 -0000 Message-Id: <20030217210234.GQ20557@aaron.unix.fryenet> I just merged with 0.89 and found it necessary to pull up some local line draw changes. we had replaced the per-pixel clipping with cohen-sutherland method and got over 100 times speedup. unfortunately the decided-upon dashed-line API ended up being a little different than ours, and I'm still recovering from that, but I will produce patches when that is completed. I want to stir the hornet's nest a little and ask why nano-X's performance in key areas like line and text draws is so pathetic. is everybody here running 100MHz+ CPUs where slow algorithms don't make a difference? we're saddled with a 25MHz 68331 and a SED1355 dumb frame buffer with no DMA, but I can't be the only one with such meager hardware working with nano-X. or can I? -- Aaron J. Grier | Frye Electronics, Tigard, OR | ####@####.#### | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
[<<] [<] Page 1 of 1 [>] [>>] |