SPARTA: the ESO standard platform for adaptive optics real time applications

Abstract

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.

SPHERE instrumentation software: a progress report

“…The SPARTA architecture is more complicated, being composed of hard real-time VME-based nodes and soft real-time nodes based on x86 servers. A thorough description of the SPARTA architecture can be found in [3]. Finally, all the nodes in the network are connected through a gigabit Ethernet LAN.…”

FPGA cluster for high-performance AO real-time control system

“…The following table [1] summarizes the requirement of the control matrix and its MVM operation for the 2 nd generation AO RTCS for ESO VLT. The requirements for the European ELT AO systems are much higher.…”

“…The size of the control matrix for the ELT XAO system can be 160, 000 x 160, 000 with a frame rate of 1kHz [7], which means 95.4GB matrix storage and 25.6TFMAC for the single precision MVM operation. 1 Latency is defined as the time when the last pixel is received by the RTC and the time when the last control voltage arrives at the deformable mirror controller (CODE) It is very hard for a computing system based on generic CPUs to deliver such performance. The computing power may not be a problem for a multi-processor computer in terms of the accumulated pure CPU speed.…”

Investigation of POWER8 processors for astronomical adaptive optics real-time control

“…doesn't require a particular hardware set to operate, is clearly advantageous. Previous system designs have frequently relied on specific hardware, typically digital signal processors (DSPs) and field programmable gate arrays (FPGAs) (for example the ESO SPARTA system, Fedrigo et al 2006), which, due to long periods spent in design, are often close to obsolescence even during commissioning, with availability of spare parts becoming problematic, and specific programming knowledge required. Hardware failure of these systems then poses the risk that an entire new system will require designing, with the original software not being portable to new hardware.…”

“…Although porting to new hardware is typically limited to other PC-like systems that have an operating system running on a central processing unit (CPU), this is not always the case. In particular, the DARC system has a modular design which allows parts of the real-time pipeline to be placed in alternative hardware, including for example: (i) pixel processing and slope calculation in FPGA using a customised version of the SPARTA system (Fedrigo et al 2006) (ii) wavefront reconstruction using graphics processing units (GPUs) (Basden et al 2010) (iii) a full GPU pipeline, from raw WFS images to DM demands.…”