Am 2014-03-24 06:28, schrieb Keith Packard:
1: intel-glamor.perf
2: intel-glamor-line.perf
1 2 Operation
------------ ------------------------- -------------------------
471000.0 231000000.0 ( 490.446) 1-pixel line
297000.0 209000000.0 ( 703.704) 1-pixel line segment
Nice speedup.
456000.0 12200000.0 ( 26.754) 10-pixel line
384000.0 10600000.0 ( 27.604) 10-pixel line segment
463000.0 134000000.0 ( 289.417) 10-pixel horizontal line
segment
460000.0 8770000.0 ( 19.065) 10-pixel vertical line
segment
371000.0 1120000.0 ( 3.019) 100-pixel line
325000.0 1060000.0 ( 3.262) 100-pixel line segment
459000.0 6550000.0 ( 14.270) 100-pixel horizontal line
segment
354000.0 851000.0 ( 2.404) 100-pixel vertical line
segment
218000.0 296000.0 ( 1.358) 500-pixel line
148000.0 289000.0 ( 1.953) 500-pixel line segment
467000.0 4020000.0 ( 8.608) 500-pixel horizontal line
segment
160000.0 177000.0 ( 1.106) 500-pixel vertical line
segment
I don't get it, why are horizontal lines so much faster than vertical
ones? This sounds like software rasterization for GL_LINES.
Should this be treated as a mesa bug or shall we write a workaround
using triangles?
_______________________________________________
[email protected]: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: http://lists.x.org/mailman/listinfo/xorg-devel
