nir: Reduce destination size of ballot intrinsic when possible

Some hardware, like i965, doesn't support group sizes greater than 32.
In that case, we can reduce the destination size of the ballot
intrinsic, which will simplify our code generation.

Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

src/intel/compiler/brw_compiler.c

@@ -58,6 +58,7 @@ static const struct nir_shader_compiler_options scalar_nir_options = {
    .lower_unpack_unorm_2x16 = true,
    .lower_unpack_unorm_4x8 = true,
    .lower_subgroup_masks = true,
+   .max_subgroup_size = 64, /* FIXME */
    .max_unroll_iterations = 32,
 };