Endpoint devices for Internet-of-Things not only need to work under extremely tight power envelope of a few milliwatts, but also need to be flexible in their computing capabilities, from a few kOPS to GOPS. Near-threshold (NT) operation can achieve higher energy efficiency, and the performance scalability can be gained through parallelism. In this paper, we describe the design of an open-source RISC-V processor core specifically designed for NT operation in tightly coupled multicore clusters. We introduce instruction extensions and microarchitectural optimizations to increase the computational density and to minimize the pressure toward the shared-memory hierarchy. For typical data-intensive sensor processing workloads, the proposed core is, on average, $3.5\times $ faster and $3.2\times $ more energy efficient, thanks to a smart L0 buffer to reduce cache access contentions and support for compressed instructions. Single Instruction Multiple Data extensions, such as dot products, and a built-in L0 storage further reduce the shared-memory accesses by $8\times $ reducing contentions by $3.2\times $ . With four NT-optimized cores, the cluster is operational from 0.6 to 1.2 V, achieving a peak efficiency of 67 MOPS/mW in a low-cost 65-nm bulk CMOS technology. In a low-power 28-nm FD-SOI process, a peak efficiency of 193 MOPS/mW (40 MHz and 1 mW) can be achieved.

Near-Threshold RISC-V Core With DSP Extensions for Scalable IoT Endpoint Devices

P2012 is an area- and power-efficient many-core computing accelerator based on multiple globally asynchronous, locally synchronous processor clusters. Each cluster features up to 16 processors with independent instruction streams sharing a multi-banked one-cycle access L1 data memory, a multi-channel DMA engine and specialized hardware for synchronization and aggressive power management. P2012 is 3D stacking ready and can be customized to achieve extreme area and energy efficiency by adding domain-specific HW IPs to the cluster. The first P2012 SoC prototype in 28nm CMOS will sample in Q3, featuring four 16-processor clusters, a 1MB L2 memory and delivering 80GOPS (with 32 bit single precision floating point support) in 18mm2 with 2W power consumption (worst-case). P2012 can run standard OpenCL™ and proprietary Native Programming Model SW components to achieve the highest level of control on application-to-resource mapping. A dedicated version of the OpenCV vision library is provided in the P2012 SW Development Kit to enable visual analytics acceleration. This paper will discuss preliminary performance measurements of common feature extraction and tracking algorithms, parallelized on P2012, versus sequential execution on ARM CPUs.

Platform 2012, a many-core computing accelerator for embedded SoCs: performance evaluation of visual analytics applications

P2012 is an area- and power-efficient many-core computing fabric based on multiple globally asynchronous, locally synchronous (GALS) clusters supporting aggressive fine-grained power, reliability and variability management. Clusters feature up to 16 processors and one control processor with independent instruction streams sharing a multi-banked L1 data memory, a multi-channel DMA engine, and specialized hardware for synchronization and scheduling. P2012 achieves extreme area and energy efficiency by supporting domain-specific acceleration at the processor and cluster level through the addition of dedicated HW IPs. P2012 can run standard OpenCL and OpenMP parallel codes well as proprietary Native Programming Model (NPM) SW components that provide the highest level of control on application-to-resource mapping. In Q3 2011 the P2012 SW Development Kit (SDK) has been made available to a community of RD it includes full OpenCL and NPM development environments. The first P2012 SoC prototype in 28nm CMOS will sample in Q4 2012, featuring four clusters and delivering 80GOPS (with single precision floating point support) in 15.2mm2 with 2W power consumption.

P2012: building an ecosystem for a scalable, modular and high-efficiency embedded computing accelerator

We develop a compositional behavioural model that integrates a variation of probabilistic automata into a conservative extension of interactive Markov chains. The model is rich enough to embody the semantics of generalised stochastic Petri nets. We define strong and weak bisimulations and discuss their compositionality properties. Weak bisimulation is partly oblivious to the probabilistic branching structure, in order to reflect some natural equalities in this spectrum of models. As a result, the standard way to associate a stochastic process to a generalised stochastic Petri net can be proven sound with respect to weak bisimulation.

On Probabilistic Automata in Continuous Time

DRAM latency continues to be a critical bottleneck for system performance. In this work, we develop a low-cost mechanism, called Charge Cache, that enables faster access to recently-accessed rows in DRAM, with no modifications to DRAM chips. Our mechanism is based on the key observation that a recently-accessed row has more charge and thus the following access to the same row can be performed faster. To exploit this observation, we propose to track the addresses of recently-accessed rows in a table in the memory controller. If a later DRAM request hits in that table, the memory controller uses lower timing parameters, leading to reduced DRAM latency. Row addresses are removed from the table after a specified duration to ensure rows that have leaked too much charge are not accessed with lower latency. We evaluate ChargeCache on a wide variety of workloads and show that it provides significant performance and energy benefits for both single-core and multi-core systems.

/pdf/chargecache-reducing-dram-latency-by-exploiting-row-access-8b1eccbqb6.pdf

ChargeCache: Reducing DRAM latency by exploiting row access locality

Requiring more bandwidth at reasonable power consumption, new communication infrastructures must provide adequate solutions to guarantee performance during physical integration. In this paper, we propose the design of a low-power asynchronous Network-on-Chip which is implemented in a bottom-up approach using optimized hard-macros. This architecture is fully testable and a new design flow is proposed to overcome CAD tools limitations regarding asynchronous logic. The proposed architecture has been successfully implemented in CMOS 65nm in a complete circuit. It achieves a 550Mflit/s throughput on silicon, and exhibits 86% power reduction compared to an equivalent synchronous NoC version.

A fully-asynchronous low-power framework for GALS NoC integration

In complex embedded applications, optimisation and adaptation of both dynamic and leakage power have become an issue at SoC grain. A fully power-aware globally-asynchronous locally-synchronous network-on-chip (NoC) circuit is presented in this paper. Network-on-chip architecture combined with a globally-asynchronous locally-synchronous paradigm is a natural enabler for DVFS mechanisms. The circuit is arranged around an asynchronous network-on-chip providing scalable communication and a 17 Gb/s throughput while automatically reducing its power consumption by activity detection. Both dynamic and static power consumptions are globally reduced using adaptive design techniques applied locally for each synchronous NoC units. No fine control software is required during voltage and frequency scaling. Power control is localized and a minimal latency cost is observed.

An Asynchronous Power Aware and Adaptive NoC Based Circuit

Baseband processing for advanced Telecom applications have to face two contradictory issues [1]. The first one is the flexibility required, with the exploding number of modes for a single protocol (e.g. 63 for 3GPP-LTE), the number of protocols to be supported by a single chip (≫10 in 2010) and new applications requiring a handover between protocols. The second concern is related to performance and power consumption: performance demands are exploding (up to 100GOPS are now required) with decreasing power consumption constraints (roughly 500mW).

A 477mW NoC-based digital baseband for MIMO 4G SDR

A fully power aware globally asynchronous locally synchronous network-on-chip circuit is presented in this paper. The circuit is arranged around an asynchronous network-on-chip providing a 17 Gbits/s throughput and automatically reducing its power consumption by activity detection. Both dynamic and static power consumptions are globally reduced using adaptive design techniques applied locally for each NoC units. The dynamic power consumption can be reduced up to a factor of 8 while the static power consumption is reduced by 2 decades in stand-by mode.

An asynchronous power aware and adaptive NoC based circuit

Wide voltage range operation for DSPs brings more versatility to achieve high energy efficiency in mobile applications. This paper describes a 32 bits DSP fabricated in 28 nm Ultra Thin Body and Box FDSOI technology. Body Biasing Voltage (VBB) scaling from 0 V up to ±2 V decreases the core VDDMIN to 397 mV and increases clock frequency by +400%@500 mV and +114%@1.3 V. The DSP frequency measurements show 2.6 GHz@1.3 V(VDD)@2 V(VBB) and 460 MHz@397 mV(VDD)@2 V(VBB). The lowest peak energy efficiency is measured at 62 pJ/op at 0.53 V. In addition to technological gains, maximum frequency tracking design techniques are proposed for wide voltage range operation. On silicon, at 0.6 V, those techniques allow high energy gain of 40.6% w.r.t. a worst case corner approach.

Yvain Thonnart

Papers

A fully-asynchronous low-power framework for GALS NoC integration

An Asynchronous Power Aware and Adaptive NoC Based Circuit

A 477mW NoC-based digital baseband for MIMO 4G SDR

An asynchronous power aware and adaptive NoC based circuit

A 460 MHz at 397 mV, 2.6 GHz at 1.3 V, 32 bits VLIW DSP Embedding F MAX Tracking