nir: Add lowering support for packing opcodes.

diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
index 7b39cbb4c12e5464cd377ded599205dce92a0fd5..bbd5b1af64e2d7166ecccba70b39196f0de3054a 100644 (file)
--- a/src/glsl/nir/nir.h
+++ b/src/glsl/nir/nir.h
@@ -1469,6 +1469,10 @@ typedef struct nir_shader_compiler_options {
    bool lower_ffract;
 
    bool lower_pack_half_2x16;
+   bool lower_pack_unorm_2x16;
+   bool lower_pack_snorm_2x16;
+   bool lower_pack_unorm_4x8;
+   bool lower_pack_snorm_4x8;
    bool lower_unpack_half_2x16;
 
    bool lower_extract_byte;

diff --git a/src/glsl/nir/nir_lower_alu_to_scalar.c b/src/glsl/nir/nir_lower_alu_to_scalar.c
index 5372fbeed88403c6033477e208e24e2da8fa8bf3..37cb0221e0b609cd05c757bd360a5b3bd0ea6892 100644 (file)
--- a/src/glsl/nir/nir_lower_alu_to_scalar.c
+++ b/src/glsl/nir/nir_lower_alu_to_scalar.c
@@ -134,6 +134,38 @@ lower_alu_instr_scalar(nir_alu_instr *instr, nir_builder *b)
       return;
    }
 
+   case nir_op_pack_uvec2_to_uint: {
+      assert(b->shader->options->lower_pack_snorm_2x16 ||
+             b->shader->options->lower_pack_unorm_2x16);
+
+      nir_ssa_def *word =
+         nir_extract_uword(b, instr->src[0].src.ssa, nir_imm_int(b, 0));
+      nir_ssa_def *val =
+         nir_ior(b, nir_ishl(b, nir_channel(b, word, 1), nir_imm_int(b, 16)),
+                                nir_channel(b, word, 0));
+
+      nir_ssa_def_rewrite_uses(&instr->dest.dest.ssa, nir_src_for_ssa(val));
+      nir_instr_remove(&instr->instr);
+      break;
+   }
+
+   case nir_op_pack_uvec4_to_uint: {
+      assert(b->shader->options->lower_pack_snorm_4x8 ||
+             b->shader->options->lower_pack_unorm_4x8);
+
+      nir_ssa_def *byte =
+         nir_extract_ubyte(b, instr->src[0].src.ssa, nir_imm_int(b, 0));
+      nir_ssa_def *val =
+         nir_ior(b, nir_ior(b, nir_ishl(b, nir_channel(b, byte, 3), nir_imm_int(b, 24)),
+                               nir_ishl(b, nir_channel(b, byte, 2), nir_imm_int(b, 16))),
+                    nir_ior(b, nir_ishl(b, nir_channel(b, byte, 1), nir_imm_int(b, 8)),
+                               nir_channel(b, byte, 0)));
+
+      nir_ssa_def_rewrite_uses(&instr->dest.dest.ssa, nir_src_for_ssa(val));
+      nir_instr_remove(&instr->instr);
+      break;
+   }
+
    case nir_op_fdph: {
       nir_ssa_def *sum[4];
       for (unsigned i = 0; i < 3; i++) {

diff --git a/src/glsl/nir/nir_opcodes.py b/src/glsl/nir/nir_opcodes.py
index be3cd17193fb52d3079126847111da59c35f3334..3b82c3ce1f9aaf8bd324e6f72de6fd1367cb0c55 100644 (file)
--- a/src/glsl/nir/nir_opcodes.py
+++ b/src/glsl/nir/nir_opcodes.py
@@ -237,6 +237,16 @@ unpack_2x16("unorm")
 unpack_4x8("unorm")
 unpack_2x16("half")
 
+unop_horiz("pack_uvec2_to_uint", 0, tuint, 2, tuint, """
+dst = (src0.x & 0xffff) | (src0.y >> 16);
+""")
+
+unop_horiz("pack_uvec4_to_uint", 0, tuint, 4, tuint, """
+dst = (src0.x <<  0) |
+      (src0.y <<  8) |
+      (src0.z << 16) |
+      (src0.w << 24);
+""")
 
 # Lowered floating point unpacking operations.
 

diff --git a/src/glsl/nir/nir_opt_algebraic.py b/src/glsl/nir/nir_opt_algebraic.py
index b761b54cd706199768bce268c357993aa6acc9ae..56b0f5ea7b2f76191fc8169a1b00a230e1efe2dd 100644 (file)
--- a/src/glsl/nir/nir_opt_algebraic.py
+++ b/src/glsl/nir/nir_opt_algebraic.py
@@ -258,6 +258,26 @@ optimizations = [
    (('extract_uword', a, b),
     ('iand', ('ushr', a, ('imul', b, 16)), 0xffff),
     'options->lower_extract_word'),
+
+    (('pack_unorm_2x16', 'v'),
+     ('pack_uvec2_to_uint',
+        ('f2u', ('fround_even', ('fmul', ('fsat', 'v'), 65535.0)))),
+     'options->lower_pack_unorm_2x16'),
+
+    (('pack_unorm_4x8', 'v'),
+     ('pack_uvec4_to_uint',
+        ('f2u', ('fround_even', ('fmul', ('fsat', 'v'), 255.0)))),
+     'options->lower_pack_unorm_4x8'),
+
+    (('pack_snorm_2x16', 'v'),
+     ('pack_uvec2_to_uint',
+        ('f2i', ('fround_even', ('fmul', ('fmin', 1.0, ('fmax', -1.0, 'v')), 32767.0)))),
+     'options->lower_pack_snorm_2x16'),
+
+    (('pack_snorm_4x8', 'v'),
+     ('pack_uvec4_to_uint',
+        ('f2i', ('fround_even', ('fmul', ('fmin', 1.0, ('fmax', -1.0, 'v')), 127.0)))),
+     'options->lower_pack_snorm_4x8'),
 ]
 
 # Add optimizations to handle the case where the result of a ternary is