thirdparty/ppsspp
2
0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update native with framebuffer_blit_nv fix for tegra.

Browse source 
This appears to fix some but not all issues with blit on Tegra.

May affect #2135 and fix issues broken by #6115.

(also throw in a link in a comment)
This commit is contained in:
Henrik Rydgard 2014-06-05 21:29:20 +02:00
parent a64fe6324a
commit f1a225728c
2 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

3
unittest/UnitTest.cpp
View file

@ -49,6 +49,8 @@ std::string System_GetProperty(SystemProperty prop) { return ""; }
#define M_PI_2 1.57079632679489661923
// asin acos atan: https://github.com/michaldrobot/ShaderFastLibs/blob/master/ShaderFastMathLib.h
// TODO:
// Fast approximate sincos for NEON
// http://blog.julien.cayzac.name/2009/12/fast-sinecosine-for-armv7neon.html
@ -67,7 +69,6 @@ std::string System_GetProperty(SystemProperty prop) { return ""; }
// At least there are only 8 constants needed - load them into two low quads and go to town.
// For every step, VDUP the constant into a new register (out of two alternating), then VMLA or VFMA into it.
// http://www.ecse.rpi.edu/~wrf/Research/Short_Notes/arcsin/
// minimax polynomial rational approx, pretty good, get four digits consistently.
// unfortunately fastasin(1.0) / M_PI_2 != 1.0f, but it's pretty close.