nir: Copy propagation between blocks

[mesa.git] / src / compiler / nir / nir_opcodes.py

diff --git a/src/compiler/nir/nir_opcodes.py b/src/compiler/nir/nir_opcodes.py
index a762fdd2201b75587a33395b907f436e96abdd6e..4ef4ecc6f221ba42bf809050f0801910823b9076 100644 (file)
--- a/src/compiler/nir/nir_opcodes.py
+++ b/src/compiler/nir/nir_opcodes.py
@@ -91,6 +91,7 @@ tfloat = "float"
 tint = "int"
 tbool = "bool32"
 tuint = "uint"
+tuint16 = "uint16"
 tfloat32 = "float32"
 tint32 = "int32"
 tuint32 = "uint32"
@@ -180,9 +181,9 @@ for src_t in [tint, tuint, tfloat]:
          bit_sizes = [8, 16, 32, 64]
       for bit_size in bit_sizes:
           if bit_size == 16 and dst_t == tfloat and src_t == tfloat:
-              rnd_modes = ['rtne', 'rtz', 'undef']
+              rnd_modes = ['_rtne', '_rtz', '']
               for rnd_mode in rnd_modes:
-                  unop_convert("{0}2{1}{2}_{3}".format(src_t[0], dst_t[0],
+                  unop_convert("{0}2{1}{2}{3}".format(src_t[0], dst_t[0],
                                                        bit_size, rnd_mode),
                                dst_t + str(bit_size), src_t, "src0")
           else:
@@ -282,19 +283,34 @@ dst.x = (src0.x <<  0) |
         (src0.w << 24);
 """)
 
+unop_horiz("pack_32_2x16", 1, tuint32, 2, tuint16,
+           "dst.x = src0.x | ((uint32_t)src0.y << 16);")
+
 unop_horiz("pack_64_2x32", 1, tuint64, 2, tuint32,
            "dst.x = src0.x | ((uint64_t)src0.y << 32);")
 
+unop_horiz("pack_64_4x16", 1, tuint64, 4, tuint16,
+           "dst.x = src0.x | ((uint64_t)src0.y << 16) | ((uint64_t)src0.z << 32) | ((uint64_t)src0.w << 48);")
+
 unop_horiz("unpack_64_2x32", 2, tuint32, 1, tuint64,
            "dst.x = src0.x; dst.y = src0.x >> 32;")
 
+unop_horiz("unpack_64_4x16", 4, tuint16, 1, tuint64,
+           "dst.x = src0.x; dst.y = src0.x >> 16; dst.z = src0.x >> 32; dst.w = src0.w >> 48;")
+
+unop_horiz("unpack_32_2x16", 2, tuint16, 1, tuint32,
+           "dst.x = src0.x; dst.y = src0.x >> 16;")
+
 # Lowered floating point unpacking operations.
 
 
-unop_horiz("unpack_half_2x16_split_x", 1, tfloat32, 1, tuint32,
-           "unpack_half_1x16((uint16_t)(src0.x & 0xffff))")
-unop_horiz("unpack_half_2x16_split_y", 1, tfloat32, 1, tuint32,
-           "unpack_half_1x16((uint16_t)(src0.x >> 16))")
+unop_convert("unpack_half_2x16_split_x", tfloat32, tuint32,
+             "unpack_half_1x16((uint16_t)(src0 & 0xffff))")
+unop_convert("unpack_half_2x16_split_y", tfloat32, tuint32,
+             "unpack_half_1x16((uint16_t)(src0 >> 16))")
+
+unop_convert("unpack_32_2x16_split_x", tuint16, tuint32, "src0")
+unop_convert("unpack_32_2x16_split_y", tuint16, tuint32, "src0 >> 16")
 
 unop_convert("unpack_64_2x32_split_x", tuint32, tuint64, "src0")
 unop_convert("unpack_64_2x32_split_y", tuint32, tuint64, "src0 >> 32")
@@ -308,17 +324,17 @@ dst = 0;
 for (unsigned bit = 0; bit < 32; bit++)
    dst |= ((src0 >> bit) & 1) << (31 - bit);
 """)
-unop("bit_count", tuint32, """
+unop_convert("bit_count", tuint32, tuint, """
 dst = 0;
-for (unsigned bit = 0; bit < 32; bit++) {
+for (unsigned bit = 0; bit < bit_size; bit++) {
    if ((src0 >> bit) & 1)
       dst++;
 }
 """)
 
-unop_convert("ufind_msb", tint32, tuint32, """
+unop_convert("ufind_msb", tint32, tuint, """
 dst = -1;
-for (int bit = 31; bit >= 0; bit--) {
+for (int bit = bit_size - 1; bit >= 0; bit--) {
    if ((src0 >> bit) & 1) {
       dst = bit;
       break;
@@ -340,9 +356,9 @@ for (int bit = 31; bit >= 0; bit--) {
 }
 """)
 
-unop("find_lsb", tint32, """
+unop_convert("find_lsb", tint32, tint, """
 dst = -1;
-for (unsigned bit = 0; bit < 32; bit++) {
+for (unsigned bit = 0; bit < bit_size; bit++) {
    if ((src0 >> bit) & 1) {
       dst = bit;
       break;
@@ -351,8 +367,8 @@ for (unsigned bit = 0; bit < 32; bit++) {
 """)
 
 
-for i in xrange(1, 5):
-   for j in xrange(1, 5):
+for i in range(1, 5):
+   for j in range(1, 5):
       unop_horiz("fnoise{0}_{1}".format(i, j), i, tfloat, j, tfloat, "0.0f")
 
 
@@ -608,6 +624,9 @@ binop_horiz("pack_half_2x16_split", 1, tuint32, 1, tfloat32, 1, tfloat32,
 binop_convert("pack_64_2x32_split", tuint64, tuint32, "",
               "src0 | ((uint64_t)src1 << 32)")
 
+binop_convert("pack_32_2x16_split", tuint32, tuint16, "",
+              "src0 | ((uint32_t)src1 << 16)")
+
 # bfm implements the behavior of the first operation of the SM5 "bfi" assembly
 # and that of the "bfi1" i965 instruction. That is, it has undefined behavior
 # if either of its arguments are 32.