[Mesa-dev] [PATCH v2 0/3] Add new SPIR-V functions in NIR

[Juan A. Suarez Romero]

This short series of 3 patches extend a couple of functions SPIR-V functions to
support 64-bit, as well as it adds new functions defined in GLSL.450.std.

The first 2 commits add support for 64-bit float-point types in Step() and
SmoothStep().

The latest commit implement GLSL.std.450 NMin(), NMax() and NClamp() SPIR-V
instructions, using NIR fmax and fmin opcodes. These NIR opcodes already handle
NaN as expected by the SPIR-v instructions.


Juan A. Suarez Romero (3):
  nir: sge operation is defined for floating-point types
  nir: add support for 64-bit in SmoothStep function
  nir: implement GLSL.std.450 NMax, NMIn and NClamp operations

 src/compiler/nir/nir_opcodes.py  |  2 +-
 src/compiler/spirv/vtn_glsl450.c | 11 ++++++++---
 2 files changed, 9 insertions(+), 4 deletions(-)

-- 
2.11.0