COMputing Platform for Adaptive optics SystemS : a high-performance heterogeneous platform for the development of the E-ELT adaptive optics – COMPASS

Full length E-ELT simulations are compute-intensive applications and as such good candidates for considering the use hardware accelerators like manycore processors. Among those accelerators, the CUDA hardware was designed to provide graphics processors (GPUs) equipped with HPC-compatible features. The proposed platform will rely on a scalable heterogeneous architecture, based on GPUs as accelerators and using commodity components, able to provide sufficient computing power at a reasonable cost. The main objective of the COMPASS project is to provide a full scale end-to-end AO development platform to the PHASE community, able to address the E-ELT scale and including a real-time core that can be directly integrated on a real system. Additionally, one of the key topics of this project is the development of a prototype for a high speed, low latency, image acquisition and processing system dedicated to AO systems and fully integrated in the simulation framework. The goal of the COMPASS project is to lead developments along four main axis: AO modeling, real-time control for AO, low-latency image acquisition and E-ELT instruments design. This project will federate the efforts of various teams with complementary expertise from high performance computing to adaptive optics systems to astrophysics around a high performance development platform. Spin-offs in each of these domains are expected from such a multi-disciplinary collaboration.

Soon to come

While these developments are mainly driven by the E-ELT instrumentation needs, it could have other applications like the real time processing of image streams for detection, recognition and identification in the surveillance and decision-help contexts as in defense, industry, security or medical surgery.

Soon to come

The European Southern Observatory is leading the design phases for the European-Extremely Large Telescope (E-ELT), a 39m diameter telescope, to provide Europe with the biggest eye on the universe ever built, with a first light foreseen in 2022. The E-ELT will be the first telescope that will entirely depend, for routine operations, on adaptive optics (AO), an instrumental technique for the correction of dynamically evolving aberrations in an optical system, used on astronomical telescopes to compensate, in real-time, for the effect of atmospheric turbulence. The two first light instruments: ELT-CAM (a wide-field imager) and ELT-IFU (an integral field spectrograph) are both designed to be coupled to AO modules. The PHASE partnership, gathering most of French AO community is one of the core contributors to the AO modules, being strongly involved in both consortia selected to lead the final design studies for the two first light instruments. The proposed COMputing Platform for Adaptive optics SystemS (COMPASS) shall provide the PHASE community with powerful means to lead the development of both these AO modules as the final design phases should begin in 2012. Based on a total integration of software with hardware and relying on a high performance heterogeneous architecture, the COMPASS platform will be used to perform end-to-end simulations of the AO system behavior and performance as well as to design and test new concepts for the Real-Time Computer (RTC), a core component of any AO system. It will also provide critical decision tools for optimizing the opto-mechanical design of the instruments that will be developed for the E-ELT. The simulation of an AO system involves multiple physics from atmospheric turbulence models to tomographic reconstruction to control theory. Moreover, full length E-ELT simulations are compute-intensive applications and as such good candidates for considering the use hardware accelerators like manycore processors. Among those accelerators, the CUDA hardware was designed to provide graphics processors (GPUs) equipped with HPC-compatible features. The proposed platform will rely on a scalable heterogeneous architecture, based on GPUs as accelerators and using commodity components, able to provide sufficient computing power at a reasonable cost. The main objective of the COMPASS project is to provide a full scale end-to-end AO development platform to the PHASE community, able to address the E-ELT scale and including a real-time core that can be directly integrated on a real system. Additionally, one of the key topics of this project is the development of a prototype for a high speed, low latency, image acquisition and processing system dedicated to AO systems and fully integrated in the simulation framework. The goal of the COMPASS project is to lead developments along four main axis: AO modeling, real-time control for AO, low-latency image acquisition and E-ELT instruments design. While these developments are mainly driven by the E-ELT instrumentation needs, it could have other applications like the real time processing of image streams for detection, recognition and identification in the surveillance and decision-help contexts as in defense, industry, security or medical surgery. This project will federate the efforts of various teams with complementary expertise from high performance computing to adaptive optics systems to astrophysics around a high performance development platform. Spin-offs in each of these domains are expected from such a multi-disciplinary collaboration.