Results: Runtime Performance of Graph Generations

Performance results of large-scale transaction generation

Execution Environments

GPU: NVIDIA Tesla V100-PCIE, 8 devices
CPU: Intel(R) Xeon(R) CPU E5-2697 v4 @ 2.30GHz, total 72 cores (18 cores, 4 sockets)
RAM: 1TB DDR4 2400MHz
OS: Ubuntu 16.04.5 LTS
Software
- GCC: version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.10)
- CUDA: V9.0.176
- Python: 2.7.12
  - NetworkX: 1.11
  - NumPy: 1.15.2
- Java: OpenJDK version "1.8.0_191" (64 bit)
- TensorFlow: 1.11.0