nir+vtn: vec8+vec16 support

This introduces new vec8 and vec16 instructions (which are the only
instructions taking more than 4 sources), in order to construct 8 and 16
component vectors.

In order to avoid fixing up the non-autogenerated nir_build_alu() sites
and making them pass 16 src args for the benefit of the two instructions
that take more than 4 srcs (ie vec8 and vec16), nir_build_alu() is has
nir_build_alu_tail() split out and re-used by nir_build_alu2() (which is
used for the > 4 src args case).

v2 (Karol Herbst):
  use nir_build_alu2 for vec8 and vec16
  use python's array multiplication syntax
  add nir_op_vec helper
  simplify nir_vec
  nir_build_alu_tail -> nir_builder_alu_instr_finish_and_insert
  use nir_build_alu for opcodes with <= 4 sources
v3 (Karol Herbst):
  fix nir_serialize
v4 (Dave Airlie):
  fix serialization of glsl_type
  handle vec8/16 in lowering of bools
v5 (Karol Herbst):
  fix load store vectorizer

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
This commit is contained in:
Rob Clark
2019-03-09 17:17:55 +01:00
committed by Karol Herbst
parent b35e583c17
commit a8ec4082a4
14 changed files with 116 additions and 24 deletions
+9 -1
View File
@@ -31,14 +31,22 @@ def src_decl_list(num_srcs):
return ', '.join('nir_ssa_def *src' + str(i) for i in range(num_srcs))
def src_list(num_srcs):
return ', '.join('src' + str(i) if i < num_srcs else 'NULL' for i in range(4))
if num_srcs <= 4:
return ', '.join('src' + str(i) if i < num_srcs else 'NULL' for i in range(4))
else:
return ', '.join('src' + str(i) for i in range(num_srcs))
%>
% for name, opcode in sorted(opcodes.items()):
static inline nir_ssa_def *
nir_${name}(nir_builder *build, ${src_decl_list(opcode.num_inputs)})
{
% if opcode.num_inputs <= 4:
return nir_build_alu(build, nir_op_${name}, ${src_list(opcode.num_inputs)});
% else:
nir_ssa_def *srcs[${opcode.num_inputs}] = {${src_list(opcode.num_inputs)}};
return nir_build_alu_src_arr(build, nir_op_${name}, srcs);
% endif
}
% endfor