On Sun, 10 Aug 2003 14:27:13 +1000 Carsten Haitzler (The Rasterman)
<[EMAIL PROTECTED]> babbled:
> Would I be correct in the assumption that the only accelerated path for
> xrender is the identity transform (1:1 scale)? all other transforms are done
> in software? (my initial tests here with xfree86 4.3.0 & nvidia's latest
> drivers(as of about a month ago) seem to indicate as much...) (and yes... my
> drivers are using acceleration... GL definitely is). ?
I've put together a benchmark suite so you can have a go at making the drivers
actually beat the CPU at rendering graphics via XRender :)
http://www.rasterman.com/files/render_bench.tar.gz
and
http://www.rasterman.com/files/imlib2-1.1.0.tar.gz
(which is required to load the sample images and compare against).
here is the output of running the benchmark here (24bpp, XFree86 4.3.0, Nvidia
Geforce4 4400ti, Kernel 2.4.20, 1.7Ghz Athlon, NVAGP=0 to stop random X
lockups/hangs) note: the longer something takes, the slower it is :)
As you can see from it Xrender even though (in theory) accelerated, is markedly
slower than client-side code to do the same (and admittedly common operations).
Either I've done something terribly wrong using Xrender and triggered the
slowest path imaginable... or this could help identifying parts of drivers to
optimize :)
Available XRENDER filters:
nearest
bilinear
fast
good
best
Setup...
*** ROUND 1 ***
---------------------------------------------------------------
Test: Test Xrender doing non-scaled Over blends
Time: 12.445 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing non-scaled Over blends
Time: 10.056 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing non-scaled Over blends
Time: 0.332 sec.
*** ROUND 2 ***
---------------------------------------------------------------
Test: Test Xrender doing 1/2 scaled Over blends
Time: 6.480 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing 1/2 scaled Over blends
Time: 6.462 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing 1/2 scaled Over blends
Time: 0.145 sec.
*** ROUND 3 ***
---------------------------------------------------------------
Test: Test Xrender doing 2* smooth scaled Over blends
Time: 125.867 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing 2* smooth scaled Over blends
Time: 125.898 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing 2* smooth scaled Over blends
Time: 2.810 sec.
*** ROUND 4 ***
---------------------------------------------------------------
Test: Test Xrender doing 2* nearest scaled Over blends
Time: 103.294 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing 2* nearest scaled Over blends
Time: 103.300 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing 2* nearest scaled Over blends
Time: 2.049 sec.
*** ROUND 6 ***
---------------------------------------------------------------
Test: Test Xrender doing general nearest scaled Over blends
Time: 186.092 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing general nearest scaled Over blends
Time: 186.248 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing general nearest scaled Over blends
Time: 5.273 sec.
*** ROUND 7 ***
---------------------------------------------------------------
Test: Test Xrender doing general smooth scaled Over blends
Time: 270.536 sec.
---------------------------------------------------------------
Test: Test Xrender (offscreen) doing general smooth scaled Over blends
Time: 270.789 sec.
---------------------------------------------------------------
Test: Test Imlib2 doing general smooth scaled Over blends
Time: 8.170 sec.
--
--------------- Codito, ergo sum - "I code, therefore I am" --------------------
The Rasterman (Carsten Haitzler) [EMAIL PROTECTED]
熊耳 - 車君 [EMAIL PROTECTED]
Mobile Phone: +61 (0)413 451 899 Home Phone: 02 9698 8615
_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel
