glsl: [u/i]mulExtended optimization for GLSL

Optimize mulExtended to use 32x32->64 multiplication.

Drivers which are not based on NIR, they can set the
MUL64_TO_MUL_AND_MUL_HIGH lowering flag in order to have same old
behavior.

v2: Add missing condition check (Jason Ekstrand)

Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Suggested-by: Matt Turner <Matt Turner <mattst88@gmail.com>
Suggested-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
This commit is contained in:
Sagar Ghuge
2019-02-27 14:02:54 -08:00
parent e551040c60
commit 1d8994a63b
7 changed files with 126 additions and 4 deletions

View File

@@ -57,6 +57,7 @@ struct gl_shader_program;
#define DDIV_TO_MUL_RCP 0x100000
#define DIV_TO_MUL_RCP (FDIV_TO_MUL_RCP | DDIV_TO_MUL_RCP)
#define SQRT_TO_ABS_SQRT 0x200000
#define MUL64_TO_MUL_AND_MUL_HIGH 0x400000
/* Opertaions for lower_64bit_integer_instructions() */
#define MUL64 (1U << 0)