TuringBench Examples
Each of these examples follows the step by step procedure described on the homepage of this site.
Example 1: A Simple TuringBench Example
Topics covered:
- Getting started with the TuringBench workflow and Docker
- Creating benchmarks that can be run in a Docker container
- Benchmarking multiple versions of software with differently tagged versions of a Docker container
The GitHub repository for this example can be found at https://github.com/edwardchalstrey1/scikit-learn-classifier
Example 2: Benchmarking on GPU with NVIDIA-Docker
Topics covered:
- Running benchmark containers on different computing platforms
- Using nvidia-docker for running containers on GPU
- Creating a benchmark script with command line arguments to run alternate benchmarks with a single container
The GitHub repository for this example can be found at https://github.com/edwardchalstrey1/neural-network-julia-gpu
Example 3: High Performance Linpack Benchmark
Topics covered: Comparing the performance of different computing platforms with the Linpack Benchmark, by running this in a Docker container
The GitHub repository for this example can be found at https://github.com/edwardchalstrey1/high-performance-linpack
Example 4: Platform Agnostic Development with Docker and Singularity
Topics covered:
- Running benchmarks with Docker containers
- Building a Singularity image from a Docker image
- Running benchmark containers on HPCs
- Good practices for container use
Example 5: Running code that requires CUDA enabled GPUs on multiple platforms
Topics covered:
- Creating a CUDA enabled image for containers that run code on GPUs with Docker
- Building a CUDA enabled Singularity image from that Docker image with Singularity Hub
- Running these containers on different platforms, including HPCs