nir/algebraic: Add lowering for 64-bit hadd and rhadd

v2: Rebase on 272e927d0e9 ("nir/spirv: initial handling of OpenCL.std
extension opcodes")

v3: Add a new lower_hadd64 flag that only applies to the 64-bit versions
of the instructions.

v4: Also enable the lowering when nir_lower_iadd64 is set.

Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com> [v3]
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/767>

diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h
index 6f72e43cf60f490d6ebd6506c7285c47d49b9bc7..d07e688b5a886a5a63fde15b99ff9ac46692a24e 100644 (file)
--- a/src/compiler/nir/nir.h
+++ b/src/compiler/nir/nir.h
@@ -2829,7 +2829,28 @@ typedef struct nir_shader_compiler_options {
    /* Set if nir_lower_wpos_ytransform() should also invert gl_PointCoord. */
    bool lower_wpos_pntc;
 
+   /**
+    * Set if nir_op_[iu]hadd and nir_op_[iu]rhadd instructions should be
+    * lowered to simple arithmetic.
+    *
+    * If this flag is set, the lowering will be applied to all bit-sizes of
+    * these instructions.
+    *
+    * \sa ::lower_hadd64
+    */
    bool lower_hadd;
+
+   /**
+    * Set if only 64-bit nir_op_[iu]hadd and nir_op_[iu]rhadd instructions
+    * should be lowered to simple arithmetic.
+    *
+    * If this flag is set, the lowering will be applied to only 64-bit
+    * versions of these instructions.
+    *
+    * \sa ::lower_hadd
+    */
+   bool lower_hadd64;
+
    bool lower_add_sat;
 
    /**

diff --git a/src/compiler/nir/nir_opt_algebraic.py b/src/compiler/nir/nir_opt_algebraic.py
index db2107ce97464c2a1a9c989024bcfb8bb0dc9ff6..97c4777a53414b4fa1f570149577b3e423ff9362 100644 (file)
--- a/src/compiler/nir/nir_opt_algebraic.py
+++ b/src/compiler/nir/nir_opt_algebraic.py
@@ -1003,6 +1003,11 @@ optimizations.extend([
    (('uhadd', a, b), ('iadd', ('iand', a, b), ('ushr', ('ixor', a, b), 1)), 'options->lower_hadd'),
    (('irhadd', a, b), ('isub', ('ior', a, b), ('ishr', ('ixor', a, b), 1)), 'options->lower_hadd'),
    (('urhadd', a, b), ('isub', ('ior', a, b), ('ushr', ('ixor', a, b), 1)), 'options->lower_hadd'),
+   (('ihadd@64', a, b), ('iadd', ('iand', a, b), ('ishr', ('ixor', a, b), 1)), 'options->lower_hadd64 || (options->lower_int64_options & nir_lower_iadd64) != 0'),
+   (('uhadd@64', a, b), ('iadd', ('iand', a, b), ('ushr', ('ixor', a, b), 1)), 'options->lower_hadd64 || (options->lower_int64_options & nir_lower_iadd64) != 0'),
+   (('irhadd@64', a, b), ('isub', ('ior', a, b), ('ishr', ('ixor', a, b), 1)), 'options->lower_hadd64 || (options->lower_int64_options & nir_lower_iadd64) != 0'),
+   (('urhadd@64', a, b), ('isub', ('ior', a, b), ('ushr', ('ixor', a, b), 1)), 'options->lower_hadd64 || (options->lower_int64_options & nir_lower_iadd64) != 0'),
+
    (('uadd_sat', a, b), ('bcsel', ('ult', ('iadd', a, b), a), -1, ('iadd', a, b)), 'options->lower_add_sat'),
    (('usub_sat', a, b), ('bcsel', ('ult', a, b), 0, ('isub', a, b)), 'options->lower_add_sat'),