[Opt] Make merging casts int(int(x)) less aggressive

quadpixels · May 6, 2023 · ed6302b · ed6302b
1 parent f664513
commit ed6302b
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 2 deletions.
diff --git a/taichi/transforms/alg_simp.cpp b/taichi/transforms/alg_simp.cpp
@@ -60,8 +60,10 @@ class AlgSimp : public BasicStmtVisitor {
              data_type_bits(second_cast) <= data_type_bits(first_cast);
     }
     if (is_integral(first_cast)) {
-      // int(int(a))
-      return data_type_bits(second_cast) <= data_type_bits(first_cast);
+      // int(int(a)), note it's not always equivalent when signedness differ,
+      // see #7915
+      return data_type_bits(second_cast) <= data_type_bits(first_cast) &&
+             is_signed(second_cast) == is_signed(first_cast);
     }
     // int(float(a))
     if (data_type_bits(second_cast) <= data_type_bits(first_cast) * 2) {

diff --git a/tests/python/test_optimization.py b/tests/python/test_optimization.py
@@ -143,3 +143,13 @@ def func():
     for i in range(3):
         for j in range(4):
             assert mat[i, j] == i + 1
+
+
+@test_utils.test()
+def test_casts_int_uint():
+    @ti.kernel
+    def my_cast(x: ti.f32) -> ti.u32:
+        y = ti.floor(x, ti.i32)
+        return ti.cast(y, ti.u32)
+
+    assert my_cast(-1) == 4294967295