A transprecision floating-point cluster for efficient near-sensor data analytics

Abstract

Recent applications in the domain of near-sensor computing require the adoption of floating-point arithmetic to reconcile high precision results with a wide dynamic range. In this paper, we propose a multi-core computing cluster that leverages the fined-grained tunable principles of transprecision computing to provide support to near-sensor applications at a minimum power budget. Our design -based on the open-source RISC-V architecture -combines parallelization and sub-word vectorization with near-threshold operation, leading to a highly scalable and versatile system. We perform an exhaustive exploration of the design space of the transprecision cluster on a cycle-accurate FPGA emulator, with the aim to identify the most efficient configurations in terms of performance, energy efficiency, and area efficiency. We also provide a full-fledged software stack support, including a parallel runtime and a compilation toolchain, to enable the development of end-to-end applications. We perform an experimental assessment of our design on a set of benchmarks representative of the near-sensor processing domain, complementing the timing results with a post place-&-route analysis of the power consumption. Finally, a comparison with the state-of-the-art shows that our solution outperforms the competitors in energy efficiency, reaching a peak of 97 Gflop/s/W on single-precision scalars and 162 Gflop/s/W on half-precision vectors.

Efficient Transform Algorithms for Parallel Ultra-Low-Power IoT End Nodes

“…We performed an experimental evaluation using a cycleaccurate PULP emulator implemented on a Xilinx UltraScale+ VCU118 FPGA board 1 . We used a configuration with 8 cores and 4 FPUs, which is the most energy-efficient for near-sensor applications [17]. We measured execution cycles, instructions, and stalls for each code region employing hardware performance counters.…”

Source Code Classification for Energy Efficiency in Parallel Ultra Low-Power Microcontrollers

“…The FPU architecture is pipelined with a single stage. This architecture, introduced in [10] as 8c4f1p (8 cores, 4 floating-point units, 1 pipeline stage), is the most energy-efficient configuration of PULP; experimental results show that this solution outperforms its main competitors in the domain of embedded processing systems.…”