https://hgpu.org/?p=6524
Exploring the Optimization Space of Multi-Core Architectures with OpenCL Benchmarks