Add a test to check that quad means computes the same result with ful…

…l optimization

cubed/tests/test_core.py

@@ -1,4 +1,5 @@
 import platform
+import random
 
 import dill
 import numpy as np
@@ -11,6 +12,7 @@
 import cubed.random
 from cubed.backend_array_api import namespace as nxp
 from cubed.core.ops import merge_chunks, partial_reduce, tree_reduce
+from cubed.core.optimization import fuse_all_optimize_dag
 from cubed.tests.utils import (
     ALL_EXECUTORS,
     MAIN_EXECUTORS,
@@ -533,3 +535,37 @@ def test_plan_quad_means(tmp_path, t_length):
 
     assert m.plan.num_tasks() > 0
     m.visualize(filename=tmp_path / "quad_means")
+
+
+def quad_means(tmp_path, t_length):
+    # based on sizes from https://gist.github.com/TomNicholas/c6a28f7c22c6981f75bce280d3e28283
+    spec = cubed.Spec(tmp_path, allowed_mem="2GB", reserved_mem="100MB")
+    u = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)
+    v = cubed.random.random((t_length, 1, 987, 1920), chunks=(10, 1, -1, -1), spec=spec)
+    uv = u * v
+    m = xp.mean(uv, axis=0)
+    return m
+
+
+def test_quad_means(tmp_path, t_length=50):
+    # run twice, with and without optimization
+    # set the random seed to ensure deterministic results
+    random.seed(42)
+    m0 = quad_means(tmp_path, t_length)
+
+    random.seed(42)
+    m1 = quad_means(tmp_path, t_length)
+
+    m1.visualize(
+        filename=tmp_path / "quad_means", optimize_function=fuse_all_optimize_dag
+    )
+
+    cubed.to_zarr(m0, store=tmp_path / "result0")
+    cubed.to_zarr(
+        m1, store=tmp_path / "result1", optimize_function=fuse_all_optimize_dag
+    )
+
+    res0 = zarr.open_array(tmp_path / "result0")
+    res1 = zarr.open_array(tmp_path / "result1")
+
+    assert_array_equal(res0[:], res1[:])