The rapid increase of data rates and volumes (peta-operations/second, 15 petabytes/year) from physics and astronomy simulations, observations, experiments and analyses are reaching critical computational impasse. To meet the demands of petascale to exascale computing challenges, fundamentally new energy efficient supercomputing architectures and solutions will have to be researched and developed. The uniqueness of this work is in adopting science/application-driven approach where the key application-drivers are identified up-front to assess the efficacy of this new approach. With wide appeal, the new period of energy efficient science with new performance metrics such as operations-per-watt presents a great opportunity to lead the progress of scientific research.
We will build a cyberinfrastructure of comprehensive software libraries, tools, frameworks, easy-assembly common hardware modules and complete turnkey solutions by leveraging emerging many-core architectures with emphasis on Graphics Processing Units (GPUs). The three-year research plan will develop algorithms, and hardware infrastructures for efficient scalable solutions directly applicable to a broad range of compute-intensive scientific problems. The set of applications are categorized into three separate domains - simulation, instrumentation and data processing - covering specific real-case challenges in cosmology, astronomy, optics, and image/data processing with potential of interdisciplinary relevance. The developed cyberinfrastructure will be released to the broader scientific community with methodologies for easy implementation.
Successfully harnessing the power of the parallel architectures such as GPUs for compute-intensive scientific problems via the planned cyberinfrastructure will open doors for new discovery and revolutionize the growth of science. The infrastructure will actively identify interdisciplinary acceleration overlaps and will alleviate adoption. Extremely high-speed massive simulations will cut the overall execution times by several orders of magnitudes, thereby reducing monthly time cycles, prone to malfunctions and delays, to hours and minutes. Remote on-site handling of high data rates will make real-time imaging in radio astronomy possible for the first time.
Partnerships have been established with international groups in National Astronomical Observatories of China, and University of Heidelberg, Germany. The proposed research will engage and enable students. The combination of low cost devices and cyberinfrastructure will supply affordable high performance computing for young researchers and students. The infrastructure will be released to the broader community in yearly cycles with open source license punctuated with workshops to widen the scope of the research.
There were two major goals in the final year of this NSF research grant: GOAL 1: To develop and demonstrate a multi-purpose flexible architecture and open source hardware, software, libraries, and tools for high performance real time in-situ signal processing and computing. To demonstrate the tools/libraries and open source hardware/software for high performance in-situ computing and signal processing, we designed, implemented, deployed and tested a multipurpose real time instrument for NSF's Arecibo Observatory (the world's largest radio telescope, in Arecibo, Puerto Rico). The "SERENDIP VI/ALFABURST" instrumentation is a combination of general purpose FPGA boards, high speed analog to digital converter Boards, CPU/GPU Servers, connected together with 10Gbit ethernet switches, this general purpose instrument currently does two scientific observing tasks simultaneously: the instrument searches conducts a powerful search for radio signals from extraterrestrial civilizations (SETI), and it simultaneously searches for Fast Radio Bursts (FRB's) and alerts other telescopes for real time follow-up observations. The SETI surveys enabled by this new computing instrumentation are the most powerful radio searches for advanced extraterrestrial life ever undertaken; they explore regions of the sky and frequency bands never before observed, with high sensitivity. There are two main reasons that this SETI sky survey is so powerful: 1) Because the instrument observes and analyzes data whenever astronomers are using the arecibo telescope ("commensal" or "piggyback" observations) we are able to get an enormous amount of telescope time on the world's best telescope for SETI. 2) This real time instrument has an enormous amount of computing power - it simultaneously analyzes 5 billion spectral bins, covering 7 beams, and two polarizations, a total of 5 GHz of bandwidth. The fast radio burst (FRB) search is also very powerful, also because the instrument observes in commenal/piggyback mode on Arecibo. If we can identify a FRB in near real time, we can alert lower frequency telescope arrays to quickly follow up on our detection - these arrays will hopefully be able to localize the source of the FRB to much better angular resolution on the sky; that could allow astronomers to finally figure out where these mysterious extremely high energy bursts are coming from, and what is causing them. The attached photos is the SERENDIP VI - ALFABURST real time signal processor at the Arecibo Observatory. The rack on the left digitizes 14 signals at 1 Gigasample per second, then forms course spectra, and uses ethernet to transmit the 80 Gbit/sec of data to the graphics processing (GPU) based servers in the right rack. All the hardware, software, gateware, tools, and libraries are open source and available to the community, available at http://casper.berkeley.edu GOAL 2: To add features and complete the development and testing of the "InfraStructure for Astrophysics Applications Computing" (ISAAC) software for the astronomy community. The goal of the The ISAAC project, led by Dr. Hemant Shukla, is to develop infrastructure to accelerate Physics and Astronomy applications using multi-core and GPU architectures.