diff --git a/python/tvm/topi/adreno/utils.py b/python/tvm/topi/adreno/utils.py
index 9716a62fcc7e..6be843571fb3 100644
--- a/python/tvm/topi/adreno/utils.py
+++ b/python/tvm/topi/adreno/utils.py
@@ -555,8 +555,13 @@ def bind_data_copy(stage, axis_to_vectorize=None):
             stage.vectorize(axes[-1])
         else:
             ftc = numpy.prod(shape)
-            vthread = get_div(ftc, 8)
             fused = stage.fuse(*stage.op.axis)
+            if ftc % 4 == 0:
+                ftc = ftc / 4
+                fused, vec = stage.split(fused, factor=4)
+                stage.vectorize(vec)
+
+            vthread = get_div(ftc, 8)
             ftc = ftc / vthread
             # 1024 is a maximum work group size on the most Adreno GPU
             num_thread = get_div(ftc, 1024 // vthread)
diff --git a/tests/python/relay/opencl_texture/test_injection_texture.py b/tests/python/relay/opencl_texture/test_injection_texture.py
new file mode 100644
index 000000000000..6fdc088f9ed4
--- /dev/null
+++ b/tests/python/relay/opencl_texture/test_injection_texture.py
@@ -0,0 +1,81 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+import re
+import tvm
+import numpy as np
+from tvm import relay
+from tvm.relay import testing
+from tvm.contrib import utils
+from utils.adreno_utils import gpu_preprocess, build_run_compare
+
+
+dtype = tvm.testing.parameter("float32")
+
+
+@tvm.testing.requires_opencl
+@tvm.testing.parametrize_targets("opencl -device=adreno")
+def test_layout_transform_to_block_nchw4c(remote, target, dtype):
+    """Verification of the case NCHW->NCHW4c"""
+    input_shape = (1, 32, 720, 1280)
+    A = relay.var("data", shape=input_shape, dtype=dtype)
+    lt = relay.layout_transform(A, "NCHW", "NCHW4c")
+    mod = relay.Function([A], lt)
+
+    build_run_compare(remote, mod, {}, {"data": input_shape}, {"data": dtype}, target)
+
+
+@tvm.testing.requires_opencl
+@tvm.testing.parametrize_targets("opencl -device=adreno")
+def test_layout_transform_to_block_nchw(remote, target, dtype):
+    """Verification of the case NCHW4c->NCHW"""
+    input_shape = (1, 36, 1, 1, 4)
+    A = relay.var("data", shape=input_shape, dtype=dtype)
+    lt = relay.layout_transform(A, "NCHW4c", "NCHW")
+    mod = relay.Function([A], lt)
+
+    build_run_compare(remote, mod, {}, {"data": input_shape}, {"data": dtype}, target)
+
+
+@tvm.testing.requires_opencl
+@tvm.testing.parametrize_targets("opencl -device=adreno")
+def test_layout_transform_to_block_nhwc4c(remote, target, dtype):
+    """Verification of the case NHWC->NHWC4c"""
+    input_shape = (1, 1, 1, 144)
+    A = relay.var("data", shape=input_shape, dtype=dtype)
+    lt = relay.layout_transform(A, "NHWC", "NHWC4c")
+    mod = relay.Function([A], lt)
+
+    build_run_compare(remote, mod, {}, {"data": input_shape}, {"data": dtype}, target)
+
+
+@tvm.testing.requires_opencl
+@tvm.testing.parametrize_targets("opencl -device=adreno")
+def test_layout_transform_to_block_nhwc(remote, target, dtype):
+    """Verification of the case NHWC4c->NHWC"""
+    input_shape = (1, 80, 80, 36, 4)
+    A = relay.var("data", shape=input_shape, dtype=dtype)
+    mean = relay.mean(A, axis=[1, 2], keepdims=True)
+    cast = relay.cast(mean, "float16")
+    lt = relay.layout_transform(cast, "NHWC4c", "NHWC")
+    mod = relay.Function([A], lt)
+
+    build_run_compare(remote, mod, {}, {"data": input_shape}, {"data": dtype}, target)
+
+
+if __name__ == "__main__":
+    test_layout_transform_to_block_nhwc(None, "opencl -device=adreno", "float16")