Bluehive - A field-programable custom computing machine for extreme-scale real-time neural network simulation

Moore, Simon W.; Fox, Paul J.; Marsh, S. J. T.; Markettos, A. Theodore; Mujumdar, Alan

doi:10.1109/fccm.2012.32

Cited by 94 publications

(73 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…FPGAs and GPUs are more readily available, configurable hardware platforms which have been used for neural simulations [1] [16]. Neural simulation is however often bounded by communication, while GPUs and FPGAs are typically better suited for tasks where the ratio communication/computation favours the latter; additionally their power requirements hamper the possibility to use them in autonomous agents.…”

Section: Related Workmentioning

confidence: 99%

Event-based neural computing on an autonomous mobile platform

Galluppi

Denk

Meiner

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Abstract-Living organisms are capable of autonomously adapting to dynamically changing environments by receiving inputs from highly specialized sensory organs and elaborating them on the same parallel, power-efficient neural substrate. In this paper we present a prototype for a comprehensive integrated platform that allows replicating principles of neural information processing in real-time. Our system consists of (a) an autonomous mobile robotic platform, (b) on-board actuators and multiple (neuromorphic) sensors, and (c) the SpiNNaker computing system, a configurable neural architecture for exploration of parallel, brain-inspired models. The simulation of neurally inspired perception and reasoning algorithms is performed in real-time by distributed, low-power, low-latency event-driven computing nodes, which can be flexibly configured using C or specialized neural languages such as PyNN and Nengo. We conclude by demonstrating the platform in two experimental scenarios, exhibiting real-world closed loop behavior consisting of environmental perception, reasoning and execution of adequate motor actions.

show abstract

Section: Related Workmentioning

confidence: 99%

Event-based neural computing on an autonomous mobile platform

Galluppi

Denk

Meiner

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

show abstract

“…Our initial prototyping platform was the Bluehive [17], shown in Figure 5. Bluehive is a custom FPGA cluster containing 16 Altera Stratix IV GX 230 FPGAs.…”

Section: Platformsmentioning

confidence: 99%

“…In [17] we presented a full-custom system for neural network simulation. The whole system was implemented in Bluespec SystemVerilog (BSV), which was interfaced to our interconnect via simple FIFO bridges.…”

Section: Example Applicationsmentioning

confidence: 99%

Reliably prototyping large SoCs using FPGA clusters

Fox

Markettos

Moore

2014

2014 9th International Symposium on Reconfigurable and Communication-Centric Systems-on-Chip (ReCoSoC)

View full text Add to dashboard Cite

Abstract-Prototyping large SoCs (Systems on Chip) using multiple FPGAs introduces a risk of errors on inter-FPGA links. This raises the question of how we can prove the correctness of a SoC prototyped using multiple FPGAs. We propose using high-speed serial interconnect between FPGAs, with a transparent error detection and correction protocol working on a link-by-link basis. Our inter-FPGA interconnect has an interface that resembles that of a network-on-chip, providing a consistent interface to a prototype SoC and masking the difference between on-chip and off-chip interconnect. Lowlatency communication and low area usage are favoured at the expense of a little bandwidth inefficiency, a trade-off we believe is appropriate given the high bandwidth of inter-FPGA links.

show abstract

“…Our previous work on custom neural computation pipelines for FPGAs resulted in a high performance design described using Bluespec HDL [3], capable of efficiently streaming neuron and synapse parameters from DDR2 memory to achieve real-time performance with 64k neurons and 64M synapses per Altera Stratix IV 230 FPGA [4].…”

Section: Introductionmentioning

confidence: 99%

Managing the FPGA memory wall: Custom computing or vector processing?

Naylor

Fox

Markettos

et al. 2013

2013 23rd International Conference on Field Programmable Logic and Applications

View full text Add to dashboard Cite

Managing the memory wall is critical for massively parallel FPGA applications where data-sets are large and external memory must be used. We demonstrate that a soft vector processor can efficiently stream data from external memory whilst running computation in parallel. A non-trivial neural computation case study illustrates that multi-core vector processing coupled with careful layout of data structures performs similarly to an elaborate full-custom memory controller and execution pipeline. The vector processing version was far simpler to code so we encourage others to consider vector machines before contemplating a full-custom architecture on FPGA.

show abstract

Bluehive - A field-programable custom computing machine for extreme-scale real-time neural network simulation

Cited by 94 publications

References 11 publications

Event-based neural computing on an autonomous mobile platform

Event-based neural computing on an autonomous mobile platform

Reliably prototyping large SoCs using FPGA clusters

Managing the FPGA memory wall: Custom computing or vector processing?

Contact Info

Product

Resources

About