Reverse Engineering Power Management on NVIDIA GPUs - Anatomy of an Autonomic-ready System Martin Peres To cite this version: Martin Peres. Reverse Engineering Power Management on NVIDIA GPUs - Anatomy of an Autonomic-ready System. ECRTS, Operating Systems Platforms for Embedded Real-Time appli- cations 2013, Jul 2013, Paris, France. hal-00853849 HAL Id: hal-00853849 https://hal.archives-ouvertes.fr/hal-00853849 Submitted on 23 Aug 2013 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Reverse engineering power management on NVIDIA GPUs - Anatomy of an autonomic-ready system Martin Peres Ph.D. student at LaBRI University of Bordeaux Hobbyist Linux/Nouveau Developer Email:
[email protected] Abstract—Research in power management is currently limited supported nor documented by NVIDIA. As GPUs are leading by the fact that companies do not release enough documentation the market in terms of performance-per-Watt [3], they are or interfaces to fully exploit the potential found in modern a good candidate for a reverse engineering effort of their processors. This problem is even more present in GPUs despite power management features. The choice of reverse engineering having the highest performance-per-Watt ratio found in today’s NVIDIA’s power management features makes sense as they processors.