diff --git a/cscs-checks/apps/tensorflow/tf2_horovod_check.py b/cscs-checks/apps/tensorflow/tf2_horovod_check.py
new file mode 100644
index 0000000000..e09a54bffd
--- /dev/null
+++ b/cscs-checks/apps/tensorflow/tf2_horovod_check.py
@@ -0,0 +1,68 @@
+import reframe as rfm
+import reframe.utility.sanity as sn
+
+
+@rfm.required_version('>=2.16')
+@rfm.parameterized_test(['small'], ['large'])
+class TensorFlow2HorovodTest(rfm.RunOnlyRegressionTest):
+    def __init__(self, variant):
+        self.descr = 'Distributed training with TensorFlow2 and Horovod'
+        self.valid_systems = ['daint:gpu']
+        self.valid_prog_environs = ['PrgEnv-gnu']
+        self.modules = ['Horovod/0.18.1-CrayGNU-19.10-tf-2.0.0']
+        self.sourcesdir = None
+        self.num_tasks_per_node = 1
+        self.num_cpus_per_task = 12
+        if variant == 'small':
+            self.valid_systems += ['dom:gpu']
+            self.num_tasks = 8
+            self.reference = {
+                'dom:gpu': {
+                    'throughput': (2031.6, -0.05, None, 'images/s'),
+                    'throughput_per_gpu': (253.9, -0.05, None, 'images/s'),
+                },
+                'daint:gpu': {
+                    'throughput': (2031.6, -0.05, None, 'images/s'),
+                    'throughput_per_gpu': (253.9, -0.05, None, 'images/s')
+                },
+            }
+        else:
+            self.num_tasks = 32
+            self.reference = {
+                'daint:gpu': {
+                    'throughput': (7976.6, -0.05, None, 'images/s'),
+                    'throughput_per_gpu': (253.9, -0.05, None, 'images/s')
+                },
+            }
+        self.perf_patterns = {
+            'throughput': sn.extractsingle(
+                r'Total img/sec on %s GPU\(s\): '
+                r'(?P<throughput>\S+) \S+' % self.num_tasks,
+                self.stdout, 'throughput', float),
+            'throughput_per_gpu': sn.extractsingle(
+                r'Img/sec per GPU: (?P<throughput_per_gpu>\S+) \S+',
+                self.stdout, 'throughput_per_gpu', float)
+        }
+        model = 'InceptionV3'
+        batch_size = 64
+        self.sanity_patterns = sn.all([
+            sn.assert_found(r'Model: %s' % model, self.stdout),
+            sn.assert_found(r'Batch size: %s' % batch_size, self.stdout)
+        ])
+        self.variables = {
+            'NCCL_DEBUG': 'INFO',
+            'NCCL_IB_HCA': 'ipogif0',
+            'NCCL_IB_CUDA_SUPPORT': '1',
+            'OMP_NUM_THREADS': '$SLURM_CPUS_PER_TASK',
+        }
+        self.pre_run = ['wget https://raw.githubusercontent.com/horovod/'
+                        'horovod/26b55a7890f6923ca58cdb68a765ed0ec436ab0f/'
+                        'examples/tensorflow2_synthetic_benchmark.py']
+        self.executable = 'python'
+        self.executable_opts = [
+            'tensorflow2_synthetic_benchmark.py',
+            '--model %s' % model,
+            '--batch-size %s' % batch_size,
+        ]
+        self.tags = {'production'}
+        self.maintainers = ['VK', 'RS']