diff --git a/tests/test_superresolution.py b/tests/test_superresolution.py
index 4cf20e921779bd41e9b764594e93b2818d71411c..4229c449f842e07d42362e5bb7e7150f91d645f8 100644
--- a/tests/test_superresolution.py
+++ b/tests/test_superresolution.py
@@ -8,6 +8,7 @@
 
 """
 from os.path import dirname, join
+from time import sleep
 
 import numpy as np
 import pytest
@@ -119,17 +120,50 @@ def test_warp():
 
     lenna_file = join(dirname(__file__), "test_data", "lenna.png")
     lenna = skimage.io.imread(lenna_file, as_gray=True).astype(np.float32)
+    lenna = torch.Tensor(lenna).cuda()
 
-    warp_vectors = list(perturbation * torch.randn(lenna.shape + (2,)) for _ in range(NUM_LENNAS))
+    lr_warp_vectors = list(perturbation * torch.randn(tuple(s // 10 for s in lenna.shape) + (2,)).cuda()
+                           for _ in range(NUM_LENNAS))
 
-    warped = [torch.zeros(lenna.shape) for _ in range(NUM_LENNAS)]
+    warp_vectors = list(torch.Tensor(lenna.shape + (2,)).cuda()
+                        for _ in range(NUM_LENNAS))
+
+    scale_transform(lr_warp_vectors[0], warp_vectors[0], 10).compile()().forward(input_field=lr_warp_vectors[0],
+            output_field=warp_vectors[0])
+
+    # for i in range(len(warp_vectors)):
+    # scale(lr_warp_vectors[i], warp_vectors[i], 10)
+
+    warped = [torch.zeros(lenna.shape).cuda() for _ in range(NUM_LENNAS)]
 
     warp_kernel = translate(lenna, warped[0], pystencils.autodiff.ArrayWrapper(
-        warp_vectors[0], index_dimensions=1), interpolation_mode='linear').compile()
+        warp_vectors[0], index_dimensions=1)).compile()().forward(lenna, warped[0], warp_vectors[0])
+
+    # for i in range(len(warped)):
+    # warp_kernel(lenna[i], warped[i], warp_vectors[i])
+
+    pyconrad.imshow(warp_vectors[0][..., 1])
+    pyconrad.imshow(lr_warp_vectors[0][..., 1])
+    pyconrad.imshow(warped[0])
+    while True:
+        sleep(20)
+
+
+def test_to_polar():
+    import torch
+    NUM_LENNAS = 5
+    perturbation = 0.1
+
+    lenna_file = join(dirname(__file__), "test_data", "lenna.png")
+    lenna = skimage.io.imread(lenna_file, as_gray=True).astype(np.float32)
+
+
+    hr, lr = pystencils.fields('hr, lr: float32[2d]')
 
-    for i in range(len(warped)):
-        warp_kernel(lenna[i], warped[i], warp_vectors[i])
+    hr.set_coordinate_origin_to_field_center()
+    lr.set_coordinate_origin_to_field_center()
 
+    lr.coordinate_transform = lambda x: sympy.Matrix((x.norm(), sympy.atan2(*x) / sympy.pi * 500))
 
 def test_polar_transform():
     x, y = pystencils.fields('x, y:  float32[2d]')
diff --git a/tests/test_vesselness.py b/tests/test_vesselness.py
index 3decece519cbe0e697df6d7d24af7992e73aa726..262dc499a978e711acd56cc0e7e862bd8f265f75 100644
--- a/tests/test_vesselness.py
+++ b/tests/test_vesselness.py
@@ -19,7 +19,7 @@ import pytest
 import sympy
 
 from pystencils_autodiff.field_tensor_conversion import create_field_from_array_like
-from pystencils_reco.vesselness import eigenvalues_3x3
+from pystencils_reco.vesselness import eigenvalues_3d, eigenvalues_3x3
 
 pytest.importorskip('tensorflow')
 
@@ -42,6 +42,10 @@ def test_3x3(target):
     print(assignments)
     kernel = assignments.compile(use_auto_for_assignments=True, target=target)
     eig1, eig2, eig3 = kernel(xx=xx, xy=xy, yy=yy, xz=xz, zz=zz, yz=yz)
+    assignments = eigenvalues_3d(eig1, eig2, eig3, xx, xy, xz, yy, yz, zz)
+    print(assignments)
+    kernel = assignments.compile(use_auto_for_assignments=True, target=target)
+    eig1, eig2, eig3 = kernel(xx=xx, xy=xy, yy=yy, xz=xz, zz=zz, yz=yz)
 
 
 @pytest.mark.parametrize('target', ('cpu',))
@@ -237,8 +241,11 @@ def test_check_forward_pycuda():
     e2_field = to_gpu(np.ones(shape))
     e3_field = to_gpu(np.ones(shape))
 
-    assignments = eigenvalues_3x3(e1_field, e2_field, e3_field, xx, xy, xz, yy, yz, zz)
+    assignments = eigenvalues_3d(e1_field, e2_field, e3_field, xx, xy, xz, yy, yz, zz)
     # assignments = eigenvalues_3x3(e1_field, e2_field, e3_field, xx, xy, xz, yy, yz, zz)
     kernel = assignments.compile()
 
     kernel()
+
+
+test_3x3('cpu')