Many-core approach to 2D-DCT calculation using an FPGA

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia de Electrónica e Telecomunicações

Algorithm-oriented design of efficient many-core architectures applied to dense matrix multiplication

Recent integrated circuit technologies have opened the possibility to design parallel architectures with hundreds of cores on a single chip. The design space of these parallel architectures is huge with many architectural options. Exploring the design space gets even more difficult if, beyond performance and area, we also consider extra metrics like performance and area efficiency, where the designer tries to design the architecture with the best performance per chip area and the best sustainable performance. In this paper we present an algorithm-oriented approach to design a many-core architecture. Instead of doing the design space exploration of the many core architecture based on the experimental execution results of a particular benchmark of algorithms, our approach is to make a formal analysis of the algorithms considering the main architectural aspects and to determine how each particular architectural aspect is related to the performance of the architecture when running an algorithm or set of algorithms. The architectural aspects considered include the number of cores, the local memory available in each core, the communication bandwidth between the many-core architecture and the external memory and the memory hierarchy. To exemplify the approach we did a theoretical analysis of a dense matrix multiplication algorithm and determined an equation that relates the number of execution cycles with the architectural parameters. Based on this equation a many-core architecture has been designed. The results obtained indicate that a 100 mm(2) integrated circuit design of the proposed architecture, using a 65 nm technology, is able to achieve 464 GFLOPs (double precision floating-point) for a memory bandwidth of 16 GB/s. This corresponds to a performance efficiency of 71 %. Considering a 45 nm technology, a 100 mm(2) chip attains 833 GFLOPs which corresponds to 84 % of peak performance These figures are better than those obtained by previous many-core architectures, except for the area efficiency which is limited by the lower memory bandwidth considered. The results achieved are also better than those of previous state-of-the-art many-cores architectures designed specifically to achieve high performance for matrix multiplication.

A many-core co-processor for embedded parallel computing on FPGA

Single processor architectures are unable to provide the required performance of high performance embedded systems. Parallel processing based on general-purpose processors can achieve these performances with a considerable increase of required resources. However, in many cases, simplified optimized parallel cores can be used instead of general-purpose processors achieving better performance at lower resource utilization. In this paper, we propose a configurable many-core architecture to serve as a co-processor for high-performance embedded computing on Field-Programmable Gate Arrays. The architecture consists of an array of configurable simple cores with support for floating-point operations interconnected with a configurable interconnection network. For each core it is possible to configure the size of the internal memory, the supported operations and number of interfacing ports. The architecture was tested in a ZYNQ-7020 FPGA in the execution of several parallel algorithms. The results show that the proposed many-core architecture achieves better performance than that achieved with a parallel generalpurpose processor and that up to 32 floating-point cores can be implemented in a ZYNQ-7020 SoC FPGA.

Sparse matrix multiplication on a reconfigurable many-core architecture

Sparse matrix-vector multiplication (SMVM) is a fundamental operation in many scientific and engineering applications. In many cases sparse matrices have thousands of rows and columns where most of the entries are zero, while non-zero data is spread over the matrix. This sparsity of data locality reduces the effectiveness of data cache in general-purpose processors quite reducing their performance efficiency when compared to what is achieved with dense matrix multiplication. In this paper, we propose a parallel processing solution for SMVM in a many-core architecture. The architecture is tested with known benchmarks using a ZYNQ-7020 FPGA. The architecture is scalable in the number of core elements and limited only by the available memory bandwidth. It achieves performance efficiencies up to almost 70% and better performances than previous FPGA designs.

Investigation on AUTOSAR-Compliant Solutions for Many-Core Architectures

As of today, AUTOSAR is the de facto standard in the automotive industry, providing a common software architec- ture and development process for automotive applications. While this standard is originally written for singlecore operated Elec- tronic Control Units (ECU), new guidelines and recommendations have been added recently to provide support for multicore archi- tectures. This update came as a response to the steady increase of the number and complexity of the software functions embedded in modern vehicles, which call for the computing power of multicore execution environments. In this paper, we enumerate and analyze the design options and the challenges of porting AUTOSAR-based automotive applications onto multicore platforms. In particular, we investigate those options when considering the emerging many- core architectures that provide a more scalable environment than the traditional multicore systems. Such platforms are suitable to enable massive parallel execution, and their design is more suitable for partitioning and isolating the software components.

Methodologies for the WCET Analysis of Parallel Applications on Many-core Architectures

Euromicro Conference on Digital System Design (DSD 2015), Funchal, Portugal.

Partitioning the Network-on-Chip to Enable Virtualization on Many-Core Processors

6th International Real-Time Scheduling Open Problems Seminar (RTSOPS 2015), Lund, Sweden.

A system model and stack for the parallelization of time-critical applications on many-core architectures

3rd Workshop on High-performance and Real-time Embedded Systems (HIRES 2015). 21, Jan, 2015. Amsterdam, Netherlands.

P-SOCRATES: A parallel software framework for time-critical many-core systems

Article in Press, Corrected Proof

Real-time Parallel Applications on Many-core Architectures

Presented at INForum - Simpósio de Informática (INFORUM 2015). 7 to 8, Sep, 2015. Portugal.

Parallel Software framework for time-critical many-core systems

The recent technological advancements and market trends are causing an interesting phenomenon towards the convergence of High-Performance Computing (HPC) and Embedded Computing (EC) domains. On one side, new kinds of HPC applications are being required by markets needing huge amounts of information to be processed within a bounded amount of time. On the other side, EC systems are increasingly concerned with providing higher performance in real-time, challenging the performance capabilities of current architectures. The advent of next-generation many-core embedded platforms has the chance of intercepting this converging need for predictable high-performance, allowing HPC and EC applications to be executed on efficient and powerful heterogeneous architectures integrating general-purpose processors with many-core computing fabrics. To this end, it is of paramount importance to develop new techniques for exploiting the massively parallel computation capabilities of such platforms in a predictable way. P-SOCRATES will tackle this important challenge by merging leading research groups from the HPC and EC communities. The time-criticality and parallelisation challenges common to both areas will be addressed by proposing an integrated framework for executing workload-intensive applications with real-time requirements on top of next-generation commercial-off-the-shelf (COTS) platforms based on many-core accelerated architectures. The project will investigate new HPC techniques that fulfil real-time requirements. The main sources of indeterminism will be identified, proposing efficient mapping and scheduling algorithms, along with the associated timing and schedulability analysis, to guarantee the real-time and performance requirements of the applications.

Optimització d'una aplicació bioinformàtica d'aliniament de seqüències executada en processadors many-core (GPUs)

Las herramientas de análisis de secuencias genómicas permiten a los biólogos identificar y entender regiones fundamentales que tienen implicación en enfermedades genéticas. Actualmente existe una necesidad de dotar al ámbito científico de herramientas de análisis eficientes. Este proyecto lleva a cabo una caracterización y análisis del rendimiento de algoritmos utilizados en la comparación de secuencias genómicas completas, y ejecutadas en arquitecturas MultiCore y ManyCore. A partir del análisis se evalúa la idoneidad de este tipo de arquitecturas para resolver el problema de comparar secuencias genómicas. Finalmente se propone una serie de modificaciones en las implementaciones de estos algoritmos con el objetivo de mejorar el rendimiento.

Optimització d'una aplicacio bioinformàtica d'alineament de seqüències executada en processadors multi-core i many-core (GPUs)

Las aplicaciones de alineamiento de secuencias son una herramienta importante para la comunidad científica. Estas aplicaciones bioinformáticas son usadas en muchos campos distintos como pueden ser la medicina, la biología, la farmacología, la genética, etc. A día de hoy los algoritmos de alineamiento de secuencias tienen una complejidad elevada y cada día tienen que manejar un volumen de datos más grande. Por esta razón se deben buscar alternativas para que estas aplicaciones sean capaces de manejar el aumento de tamaño que los bancos de secuencias están sufriendo día a día. En este proyecto se estudian y se investigan mejoras en este tipo de aplicaciones como puede ser el uso de sistemas paralelos que pueden mejorar el rendimiento notablemente.

Exploration and Design of Power-Efficient Networked Many-Core Systems

Multiprocessing is a promising solution to meet the requirements of near future applications. To get full benefit from parallel processing, a manycore system needs efficient, on-chip communication architecture. Networkon- Chip (NoC) is a general purpose communication concept that offers highthroughput, reduced power consumption, and keeps complexity in check by a regular composition of basic building blocks. This thesis presents power efficient communication approaches for networked many-core systems. We address a range of issues being important for designing power-efficient manycore systems at two different levels: the network-level and the router-level. From the network-level point of view, exploiting state-of-the-art concepts such as Globally Asynchronous Locally Synchronous (GALS), Voltage/ Frequency Island (VFI), and 3D Networks-on-Chip approaches may be a solution to the excessive power consumption demanded by today’s and future many-core systems. To this end, a low-cost 3D NoC architecture, based on high-speed GALS-based vertical channels, is proposed to mitigate high peak temperatures, power densities, and area footprints of vertical interconnects in 3D ICs. To further exploit the beneficial feature of a negligible inter-layer distance of 3D ICs, we propose a novel hybridization scheme for inter-layer communication. In addition, an efficient adaptive routing algorithm is presented which enables congestion-aware and reliable communication for the hybridized NoC architecture. An integrated monitoring and management platform on top of this architecture is also developed in order to implement more scalable power optimization techniques. From the router-level perspective, four design styles for implementing power-efficient reconfigurable interfaces in VFI-based NoC systems are proposed. To enhance the utilization of virtual channel buffers and to manage their power consumption, a partial virtual channel sharing method for NoC routers is devised and implemented. Extensive experiments with synthetic and real benchmarks show significant power savings and mitigated hotspots with similar performance compared to latest NoC architectures. The thesis concludes that careful codesigned elements from different network levels enable considerable power savings for many-core systems.

Adaptive Routing Approaches for Networked Many-Core Systems

Through advances in technology, System-on-Chip design is moving towards integrating tens to hundreds of intellectual property blocks into a single chip. In such a many-core system, on-chip communication becomes a performance bottleneck for high performance designs. Network-on-Chip (NoC) has emerged as a viable solution for the communication challenges in highly complex chips. The NoC architecture paradigm, based on a modular packet-switched mechanism, can address many of the on-chip communication challenges such as wiring complexity, communication latency, and bandwidth. Furthermore, the combined benefits of 3D IC and NoC schemes provide the possibility of designing a high performance system in a limited chip area. The major advantages of 3D NoCs are the considerable reductions in average latency and power consumption. There are several factors degrading the performance of NoCs. In this thesis, we investigate three main performance-limiting factors: network congestion, faults, and the lack of efficient multicast support. We address these issues by the means of routing algorithms. Congestion of data packets may lead to increased network latency and power consumption. Thus, we propose three different approaches for alleviating such congestion in the network. The first approach is based on measuring the congestion information in different regions of the network, distributing the information over the network, and utilizing this information when making a routing decision. The second approach employs a learning method to dynamically find the less congested routes according to the underlying traffic. The third approach is based on a fuzzy-logic technique to perform better routing decisions when traffic information of different routes is available. Faults affect performance significantly, as then packets should take longer paths in order to be routed around the faults, which in turn increases congestion around the faulty regions. We propose four methods to tolerate faults at the link and switch level by using only the shortest paths as long as such path exists. The unique characteristic among these methods is the toleration of faults while also maintaining the performance of NoCs. To the best of our knowledge, these algorithms are the first approaches to bypassing faults prior to reaching them while avoiding unnecessary misrouting of packets. Current implementations of multicast communication result in a significant performance loss for unicast traffic. This is due to the fact that the routing rules of multicast packets limit the adaptivity of unicast packets. We present an approach in which both unicast and multicast packets can be efficiently routed within the network. While suggesting a more efficient multicast support, the proposed approach does not affect the performance of unicast routing at all. In addition, in order to reduce the overall path length of multicast packets, we present several partitioning methods along with their analytical models for latency measurement. This approach is discussed in the context of 3D mesh networks.