GPU Lynx: a Dynamic Instrumentation System for Data-Parallel Applications on GPGPU Architectures
GPU Lynx is a dynamic instrumentation framework intended to facilitate the creation and insertion of PTX instrumentation procedures in CUDA applications. This project was developed by Naila Farooqui and is based on concepts described in her paper presented at ISPASS 2012. GPU Lynx utilizes GPU Ocelot’s PTX internal representation and provides a custom interposer layer for the CUDA Runtime API as well as a compilation path from C to PTX.
Visit the project site for GPU Lynx: http://code.google.com/p/gpulynx/.