Dell PowerEdge C4140 Deep Learning Performance Comparison - Scale-up vs. Scale - Page 3

Deep Learning Performance: Scale-up vs Scale-out Contents 1 Overview ...5 1.1 Definition ...5 2 Introduction ...6 2.1 Deep Learning ...7 3 Background ...7 3.1 Criteria...10 3.2 Why TensorFlow as the framework of choice 10 4 Test Methodology ...11 4.1 Testing Methodology ...12 4.1.1 Short Test ...12 4.1.2 Long Test ...13 4.2 Throughput Testing...13 4.3 Neural Models & parameters ...14 5 PowerEdge Server Details ...15 5.1 PowerEdge C4140 ...15 5.1.1 Why is C4140 Configuration-M better 15 5.2 PowerEdge R740/R740xd ...18 6 Framework Setup Details...19 6.1 Distributed Horovod-TensorFlow Setup 19 6.2 Evaluation Platform Setup ...21 7 Performance Results ...22 7.1 Single Node - Throughput (images/sec 22 7.1.1 PowerEdge R740xd ...22 7.1.2 PowerEdge C4140-V100-16GB-PCle [Config B] - Single Node 23 7.1.3 PowerEdge C4140-K-V100-16GB SXM2 Single Node 23 7.1.4 PowerEdge C4140-K- V100-32GB SXM2 Single Node 24 7.1.5 PowerEdge C4140-M- V100-16GB SXM2 Single Node 25 7.1.6 ...25 7.1.7 PowerEdge C4140- V100-SXM2 Configuration K versus Configuration M - Single Node ... 25 7.1.8 What role does CPU play in Deep learning 26 7.1.9 Conclusion...28 Architectures & Technologies Dell EMC | Infrastructure Solutions Group 2