Six Laser Guide Stars (LGS) are included in the design of the European Extremely Large Telescope (ELT), with all of its current instruments taking advantage of them using Shack-Hartmann (SH) wavefront sensors (WFS). However, this implementation raises new issues related to the unprecedented elongation that results from the perspective effect combined to the thickness of the sodium layer. In order to investigate wavefront sensing with an elongated LGS on a SH WFS, we are taking advantage of the presence of the multi-object adaptive optics demonstrator CANARY on the William Herschel Telescope (WHT), in La Palma island, that was upgraded with a sodium LGS WFS for our experiment. The LGS is generated by ESO’s transportable Wendelstein LGS unit and the elongation is obtained by positioning the laser launch telescope 40 meters away from the WHT. With this experiment we are able to measure wavefronts using an elongated LGS WFS. In this paper, we present results obtained during the latest run of observations in September 2017. In these results is comprised an error breakdown of wavefront measurement on elongated LGS. The performances of several centroiding methods are compared thanks to this error breakdown. Additionally, we take advantage of varying observation conditions with respect to seeing and sodium profile to establish the robustness of the different centroiding methods. Finally, these performances are evaluated for different SH designs, to explore which compromises can be reached with respect to pixel scale and sub-aperture field of view.
With the next-generation of Extremely Large Telescopes (ELTs), the demands of adaptive optics real-time control (AO RTC) increase massively compared to the most complex AO systems in use today. Green Flash, an ongoing EU funded project, is investigating the optimal architecture for ELT scale AO RTC, with an emphasis on GPU and many core CPU solutions. The Intel Xeon Phi range of x86 CPUs is our current focus of investigation into CPU technologies to solve the ELT-scale AO RTC problem. Built using Intels Many Integrated Core (MIC) architecture incorporating 64 general purpose x86 CPU cores into a single CPU package paired with a large pool of on-chip high bandwidth MCDRAM, the Xeon Phi includes many of the advantages of current technologies. The current generation Xeon Phi is readily compatible with standard Linux operating systems and all of the tools and libraries, and as a standard socketed CPU it eliminates the latency introduced by the extra data transfers required for previous Xeon Phis and other accelerator devices. The Durham Adaptive Optics Real-time Controller (DARC) is a freely available, on-sky tested, fully modular, x86 CPU based AO RTC which which is ideally suited to be a basis for our investigation into ELT scale AO RTC performance. We present a proof of concept AO RTC system, in collaboration with the Green Flash project, for ELT scale MCAO, with the requirements of the MAORY AO system in mind, using an optimised DARC on Xeon Phi hardware to achieve the required performance.
CANARY is a wide-field AO on-sky test facility which has been operated annually on the 4.2m William Herschel Telescope since 2010. CANARY has the stated goal of testing and demonstrating AO technologies that are critical for ELT AO performance. It has seen four distinct phases where new AO technologies have been developed and demonstrated, including NGS MOAO in 2010 (phase A), Rayleigh LGS and NGS MOAO in 2012 and 2013 (phase B, with LGS commissioning in 2011), LTAO operation in 2014 and 2015, and finally operation with a single Sodium laser guide star launched far off axis in 2016 and 2017 (phase D). By launching this laser guide star 40m off axis, extremely elongated laser guide star spots are created in the CANARY LGS Shack-Hartmann wavefront sensor. Therefore, the 7×7 sub-apertures of CANARY can be used to test wavefront sensing performance of a sub-pupil of the ELT located furthest from the laser launch axis. We present an overview of CANARY in its phase D configuration. Depending on where in the sky the LGS is pointing, the projected baseline between the on-axis LGS wavefront sensor and the laser launch location, as seen by the wavefront sensor, will vary from about 20-40m, allowing us to artificially generate different degrees of elongation. Additionally, the well sampled CANARY sub-apertures have 30×30 pixels each and a 20 arcsecond field of view, using an OCAM2S EMCCD camera. This means that by shrinking sub-apertures, and optionally by binning pixels, we are able to investigate different pixel scales and fields of view for the ELT systems, thus determining the optimal design parameters. Here we discuss the closed loop tests that were performed to investigate the effect of spot truncation and extreme elongation. We include different correlation techniques, including standard FFT-based correlation, brute force correlation and correlation by difference squared. We also mention dynamic and automatic updates of the correlation reference images while the AO loop is engaged that have previously been reported. The matched filter algorithm is also mentioned, with a pointer to our prior on-sky investigations. We give our recommendation for the ELT wavefront sensing algorithm of choice, and our evidence based reasons for this recommendation, which may come as a surprise to some. Finally we also present the future experiments to be performed with CANARY, give details of the OPTICON funded programme which enables the hosting of AO experiments on CANARY, allowing the AO community to get involved.
The Green Flash initiative responds to a critical challenge in the astronomical community. Scaling up the real-time control solutions of AO instruments in operation to the specifications of the AO modules at the core of the next generation of extremely large telescopes is not a viable option. The main goal of this project is to design and build a prototype for an AO RTC targeting the E-ELT first-light AO instrumentation. We have proposed innovative technical solutions based on emerging technologies in High Performance Computing, assessed this enabling technologies through prototyping and are now assembling a full scale demonstrator to be validated with a simulator and eventually tested on sky. In this paper, we report on downselection process that led us to the final prototype architecture and the performance of our full scale prototype obtained with a real-time simulator.
Adaptive Optics (AO) is a necessary technology for ensuring the success of the next generation of extremely large telescopes (ELTs). It’s used to help mitigate the perturbing effects of Earth’s atmosphere on the incoming light from astronomical objects and will be an integral part of ELTs for obtaining close to diffraction limited images. To maintain a correction of the incoming wavefront under dynamic atmospheric conditions, which can change significantly on the order of milliseconds, the frame-by-frame reconstruction must be operated in real-time, with hard limits on the time interval between measuring the disturbance and applying a correction. The main problem size for AO RTC increases with the 4th power of telescope diameter and so the computational demands of AO RTCs for ELTs, with primary mirror diameters between 20-40m, increase significantly compared to the current generation of 10m class telescopes. This makes the investigation into and the development of real-time controllers (RTCs) for ELT scale AO systems critical for ensuring the effectiveness of these instruments for first light. Green Flash, which is an ongoing EU funded project, has the aim of investigating the optimal hardware architecture for ELT scale AO RTC, with an emphasis on GPU and Xeon Phi solutions. The Intel Xeon Phi, built using Intel’s Many Integrated Core (MIC) architecture, incorporates ≥64 general purpose x86 CPU cores into a single CPU package paired with a large pool of on chip high bandwidth MCDRAM, it has many of the advantages of current technologies without some of the more significant drawbacks. The most computationally intensive aspects of most AO RTC pipelines are large matrix-vector multiplications mainly used to compute the reconstructed wavefronts which are highly parallelizable and are generally memory bandwidth bound. This makes the Xeon Phi with it’s large CPU count and high bandwidth memory ideally suited for acceleration of the reconstruction task and therefore for ELT scale AO RTC. The most recent incarnation of the Xeon Phi platform is available as a standard socketed x86 CPU allowing previous efforts made in developing CPU based RTC software to be used as a basis for a Xeon Phi based RTCs with the added advantage that any optimisations made for the MIC architecture can be carried forward to future x86 CPU based systems. The Durham Adaptive Optics Real-time Controller (DARC) is an example of a freely available, on-sky tested, fully modular, x86 CPU based AO RTC which which is ideally suited to be a basis for our investigation into ELT scale AO RTC performance. We present a proof of concept AO RTC system, in collaboration with the Green Flash project, using an optimised DARC on a multi-node homogeneous Xeon Phi cluster to demonstrate the potential of the MIC platform for AO RTC. We will present our methods of optimisation for the C based DARC for the Xeon Phi, including BIOS, kernel and OS tuning as well as considerations for multi-threading and massively parallel algorithm development.
We present the outcomes of an evaluation of middleware technologies for adaptive optics real-time control against the requirements of the Green Flash project, which are derived from the most demanding requirements of proposed first generation E-ELT instruments. The technology down-selection process applied in Green Flash is described, and measured performance of the selected middlewares on the hardware of a Green Flash prototype RTC are presented.
The main goal of Green Flash is to design and build a prototype for a Real-Time Controller (RTC) targeting the European Extremely Large Telescope (E-ELT) Adaptive Optics (AO) instrumentation. The E-ELT is a 39m diameter telescope to see first light in the early 2020s. To build this critical component of the telescope operations, the astronomical community is facing technical challenges, emerging from the combination of high data transfer bandwidth, low latency and high throughput requirements, similar to the identified critical barriers on the road to Exascale. With Green Flash, we will propose technical solutions, assess these enabling technologies through prototyping and assemble a full scale demonstrator to be validated with a simulator and tested on sky. With this R&D program we aim at feeding the E-ELT AO systems preliminary design studies, led by the selected first-light instruments consortia, with technological validations supporting the designs of their RTC modules. Our strategy is based on a strong interaction between academic and industrial partners. Components specifications and system requirements are derived from the AO application. Industrial partners lead the development of enabling technologies aiming at innovative tailored solutions with potential wide application range. The academic partners provide the missing links in the ecosystem, targeting their application with mainstream solutions. This increases both the value and market opportunities of the developed products. A prototype harboring all the features is used to assess the performance. It also provides the proof of concept for a resilient modular solution to equip a large scale European scientific facility, while containing the development cost by providing opportunities for return on investment.
CANARY is a Multi-Object Adaptive Optics (MOAO) system designed to demonstrate the AO aspects of proposed EELT
instruments such as the multi-object spectrograph EAGLE. The first phase of Canary will be executed on the 4.2m
William Herschel Telescope in 2010. We describe here the AO Real-time Control System (RTCS) for Canary. This is
based on a distributed architecture of components interconnected by a fast serial fabric (sFPDP). The hardware used is a
hybrid of FPGA and CPU technology. The middleware used for system data telemetry and control is based on CORBA
and the publish/subscribe pattern. The system is designed to be easily modified and extended for the later, higher order,
phases of CANARY. In order to provide the increase in computational power required in higher order systems, the
current CPU technology can be readily replaced by acceleration hardware based on FPGA or GPU technologies. The
Canary RTCS thus provides a test-bed for these new technologies that will be required for E-ELT instruments. These
design concepts can be developed to provide an RTCS for E-ELT instruments and are in line with those under
consideration by ESO for the E-ELT AO systems to which instruments such as EAGLE will be required to interface.
The CANARY on-sky MOAO demonstrator is being integrated in the laboratory and a status update about its
various components is presented here. We also discuss the alignment and calibration procedures used to improve
system performance and overall stability. CANARY will be commissioned at the William Herschel Telescope at
the end of September 2010.
EAGLE is a multi-object 3D spectroscopy instrument currently under design for the 42-metre European Extremely Large
Telescope (E-ELT). Precise requirements are still being developed, but it is clear that EAGLE will require (~100 x 100
actuator) adaptive optics correction of ~20 - 60 spectroscopic subfields distributed across a ~5 arcminute diameter field
of view. It is very likely that LGS will be required to provide wavefront sensing with the necessary sky coverage. Two
alternative adaptive optics implementations are being considered, one of which is Multi-Object Adaptive Optics
(MOAO). In this scheme, wavefront tomography is performed using a set of LGS and NGS in either a completely open-loop
manner, or in a configuration that is only closed-loop with respect to only one DM, probably the adaptive M4 of the
E-ELT. The fine wavefront correction required for each subfield is then applied in a completely open-loop fashion by
independent DMs within each separate optical relay. The novelty of this scheme is such that on-sky demonstration is
required prior to final construction of an E-ELT instrument. The CANARY project will implement a single channel of an
MOAO system on the 4.2m William Herschel Telescope. This will be a comprehensive demonstration, which will be
phased to include pure NGS, low-order NGS-LGS and high-order woofer-tweeter NGS-LGS configurations. The LGSs
used for these demonstrations will be Rayleigh systems, where the variable range-gate height and extension can be used
to simulate many of the LGS effects on the E-ELT. We describe the requirements for the various phases of MOAO
demonstration, the corresponding CANARY configurations and capabilities and the current conceptual designs of the
TThe European Southern Observatory (ESO) and Durham University's Centre for Advanced Instrumentation (CfAI) have
progressed the 'Standard Platform for Adaptive optics Real-Time Applications' (SPARTA) to a preliminary design
which successfully passed review in July 07.
SPARTA's Real-Time Control box contains a mixture of processors, Digital Signal Processors and Field Programmable
Gate Arrays (FPGA). The card selected in the design to host SPARTA's FPGA based Wavefront Processing Unit
(WPU) is the VMETRO dual processor, dual FPGA VPF1 board. CfAI benchmarked this card along with the
VMETRO Zero Latency Switch Card to assess their suitability for the SPARTA design.
The VPF1 benchmark tests summarised in this paper includes: transfering data between FPGAs, between the FPGAs and
PPCs (with and without the TransComm controller), from PPC to an external host via Ethernet and finally from the
FPGA optical transceiver using CfAI's Serial Front Panel Data Port (sFPDP) core.
Results show that the VPF1 is suitable to use as a WPU host card as the measured 280 MB/s FGPA to PPC TransComm
data rates easily cope with SPARTA's maximum requirement of 30 MB/s. Using the VPF1 as a "back-end" controller
introduces a latency of 17μs, which is greater than the original requirements of 12μs but still acceptable. The conclusion
of the benchmarking tests and the design review is that the VPF1 and Zero Latency Switch have been selected for use in
Numerical Simulation is an essential part of the design and optimisation of astronomical adaptive optics systems. Simulations of adaptive optics are computationally expensive and the problem scales rapidly with telescope aperture size, as the required spatial order of the correcting system increases. Practical realistic simulations of AO systems for extremely large telescopes are beyond the capabilities of all but the largest of modern parallel supercomputers. Here we describe a more cost effective approach through the use of hardware acceleration using field programmable gate arrays. By transferring key parts of the simulation into programmable logic, large increases in computational bandwidth can be expected. We show that the calculation of wavefront sensor images (involving a 2D FFT, photon shot noise addition, background and readout noise), and centroid calculation can be accelerated by factor of 400 times when the algorithms are transferred into hardware. We also provide details about the simulation platform and framework that we have developed at Durham.
The European Southern Observatory (ESO) and Durham University's Centre for Advanced Instrumentation (CfAI) continue to progress the design of a next generation Adaptive Optics (AO) Real-Time Control System (RTCS). This common flexible platform, labelled SPARTA 'Standard Platform for Adaptive optics Real-Time Applications' will control the AO systems for a set of 2nd generation VLT instrumentation, and will scale to implement the initial AO systems for the European Extremely Large Telescope (E-ELT).
Durham has used Field Programmable Gate Arrays (FPGA) to design a front-end Wavefront Sensor (WFS) Processing Unit (WPU) for SPARTA. FPGA devices have been used to alleviate the highly parallel computationally intensive WPS processing task from system processors to increase the obtainable control loop frequency and reduce the computational latency in the control system. The FPGA device reduces WFS frames to gradient vectors before passing the data to the system processors. The FPGA allows the processors to deal with other tasks such as wavefront reconstruction, telemetry and real-time data recording, allowing for more complex adaptive control algorithms to be executed.
Durham has design, coded, implemented and tested a FPGA core incorporating the VITA 17.1 standard serial Front Panel Data Port (sFPDP) protocol to allow a data transfer rate of 2.5Gbps-1 from the WFS Controller to the SPARTA platform.
This paper overviews the SPARTA WPU requirements and design, the sFPDP FPGA Core and a description of the platform's implementation phase.
KEYWORDS: Digital signal processing, Logic, Calibration, Computing systems, Adaptive optics, Field programmable gate arrays, Control systems, Telecommunications, System integration, Real-time computing
Whilst the high throughput and low latency requirements for the next generation AO real-time control systems have posed a significant challenge to von Neumann architecture processor systems, the Field Programmable Gate Array (FPGA) has emerged as a long term solution with high performance on throughput and excellent predictability on latency. Moreover, FPGA devices have highly capable programmable interfacing, which lead to more highly integrated system.
Nevertheless, a single FPGA is still not enough: multiple FPGA devices need to be clustered to perform the required subaperture processing and the reconstruction computation. In an AO real-time control system, the memory bandwidth is often the bottleneck of the system, simply because a vast amount of supporting data, e.g. pixel calibration maps and the reconstruction matrix, need to be accessed within a short period. The cluster, as a general computing architecture, has excellent scalability in processing throughput, memory bandwidth, memory capacity, and communication bandwidth. Problems, such as task distribution, node communication, system verification, are discussed.
ESO is starting a number of new projects collectively called Second Generation VLT instrumentation. Several of them will use Adaptive Optics (AO). In comparison with today's ESO AO systems, the 2nd Generation VLT AO systems will be much bigger (in terms of degrees of freedom) and faster (in terms of loop frequency). Consequently the Real-Time Computer controlling these AO systems will be significantly bigger and more challenging to build compared with today's AO systems in operation. To support the new requirements ESO started the development of a common flexible platform called SPARTA for Standard Platform for Adaptive optics Real Time Applications. The guidelines along which SPARTA is developed recognize the importance of industry standards over custom development to lower the development costs, ease the maintenance and make the system upgradeable thus delivering the performance required. SPARTA is based on a hybrid architecture that comprises all the major computing architectures available today: the high computational throughput is achieved through the combination of FPGA and DSP usage, where DSP are used as fast coprocessors and FPGA are used as front and as communication infrastructure, thus guaranteeing also the low latency. The flexibility is spread between the usage of both high-end CPUs and again the DSPs. All three technologies are organized in a parallel system interconnected by fast serial fabrics based on standard protocols. External input / output interfaces are also based on industry standard protocols, thus enabling the usage of commercially available tools for development and testing.
The European Southern Observatory (ESO) and Durham University's Centre for Advanced Instrumentation (CfAI) are currently designing a standard next generation Adaptive Optics (AO) Real-Time Control System. This platform, labelled SPARTA 'Standard Platform for Adaptive optics Real-Time Applications' will initially control the AO systems for ESO's 2nd generation VLT instruments, and will scale to implement the initial AO systems for ESO's future 100m telescope OWL. Durham's main task is to develop the Wavefront Sensor (WFS) front end and Statistical Machinery for the SPARTA platform using Field Programmable Gate Arrays (FPGA). SPARTA takes advantage of a FPGA device to alleviate the highly parallel computationally intensive tasks from the system processors, increasing the obtainable control loop frequency and reducing the computational latency in the control system. The WFS pixel stream enters a PMC hosted FPGA card contained within the SPARTA platform via optical fibres carrying the VITA 17.18/10 standard 2.5Gbps-1 serial Front Panel Data Port (sFPDP) protocol. Each FPGA board can receive a maximum of 10Gbs-1 of data via on-board optical transceivers. The FPGA device reduces WFS frames to gradient vectors before passing the data to the system processors. The FPGA allows the processors to deal with other tasks such as wavefront reconstruction, telemetry and real-time data recording, allowing for more complex adaptive control algorithms to be executed. This paper overviews the SPARTA requirements and current platform architecture, Durham's Wavefront Processor FPGA design and it concludes with a future plan of work.
Durham University's Centre for Advanced Instrumentation (CfAI) are currently producing a generic high-performance low-cost real-time control system (RTCS) for adaptive optics (AO) based on Field Programmable Gate Array (FPGA) technology. This platform, labelled DARTS, 'Durham Adaptive optics Real Time System', will primarily be used as the controller for Durham's enhanced Rayleigh Technical Demonstrator (RTD) system. However, DARTS could be used as a low latency control system for existing AO instruments or could be used for future 'budget' AO Natural Guide Star (NGS) and/or Laser Guide Star (LGS) RTCS. DARTS uses an FPGA device to host an end-to-end modular real-time AO pipeline connected to a Wishbone control bus. The FPGA takes advantage of the pipeline's highly parallel computationally intensive tasks which usually are calculated in series by a system processor. DARTS hopes to increase the obtainable control loop frequency and reduce the computational latency of the RTD's RTCS. DARTS is capable of high bandwidth I/O due to the implementation of the serial Front Panel Data Port (sFPDP) industrial protocol. The hardware's I/O design is modular, allowing for the future connection of various WFSs and DMs via signal converters. Various communications architectures are suggested to allow non real-time configuration and visualisation data to flow between the wishbone control bus and a processing device, either externally or internally to the FPGA device. This paper reveals the current status of the project.