GPGPU-Sim Correlator

This website is devoted to plotting and correlating the performance model for GPGPU-Sim for perforamnce and other key system parameters (obtained through nvprof) against a variety of cards for a variety of applications. This is an open-source project, and we encourage those interested to help improve both the analysis (by contributing new card configurations and/or applications) and the relative accuracy of the simulator itself by contibuting to the codebase. If you use this data or our infrastructure, please cite our SIGMETRICS 2018 paper:

Akshay Jain, Mahmoud Khairy, Timothy G. Rogers, A Quantitative Evaluation of Contemporary GPU Simulation Methodology.
In the 2018 ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Science (SIGMETRICS-2018), Irvine, California.

The Infrastrucutre used to generate these plots will be available on Github https://github.com/purdue-aalp/gpgpu-sim_simulations
The raw hardware data for all these applications can be found at https://engineering.purdue.edu/tgrogers/gpgpu-sim/hw_data/quadro.v100.cycle.tgz

The correlation data has been migrated to the Accel-Sim project

http://accel-sim.github.io/qv100-cycles.html