OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

glsl: Use a separate div_to_mul_rcp lowering flag for integers.

Browse Source 
Using multiply and reciprocal for integer division involves potentially
lossy floating point conversions.  This is okay for older GPUs that
represent integers as floating point, but undesirable for GPUs with
native integer division instructions.

TGSI, for example, has UDIV/IDIV instructions for integer division,
so it makes sense to handle this directly.  Likewise for i965.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Signed-off-by: Bryan Cain <bryancain3@gmail.com>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
This commit is contained in:
Bryan Cain
2011-08-27 18:32:58 -05:00
committed by Kenneth Graunke
parent 87679e2ea1
commit 478034f34a
5 changed files with 79 additions and 62 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
src/glsl/ir_optimization.h
View File

@@ -29,12 +29,13 @@
*/
/* Operations for lower_instructions() */
#define SUB_TO_ADD_NEG 0x01
#define DIV_TO_MUL_RCP 0x02
#define EXP_TO_EXP2 0x04
#define POW_TO_EXP2 0x08
#define LOG_TO_LOG2 0x10
#define MOD_TO_FRACT 0x20
#define SUB_TO_ADD_NEG 0x01
#define DIV_TO_MUL_RCP 0x02
#define EXP_TO_EXP2 0x04
#define POW_TO_EXP2 0x08
#define LOG_TO_LOG2 0x10
#define MOD_TO_FRACT 0x20
#define INT_DIV_TO_MUL_RCP 0x40
bool do_common_optimization(exec_list *ir, bool linked, unsigned max_unroll_iterations);