Run - time Adaptable VLIW Processors
نویسنده
چکیده
To my father and all other members of my family Summary In this dissertation, we propose to combine programmability with reconfig-urability by implementing an adaptable programmable VLIW processor in a reconfigurable hardware. The approach allows applications to be developed at high-level (C language level), while at the same time, the processor organization can be adapted to the specific requirements (both static and dynamic) of different applications. Our proposed customizable VLIW processor called ρ-VEX can be adapted at design-time as well as at run-time. Its instruction set architecture (ISA) is based on the VEX ISA and a toolchain (parametrized C compiler and sim-ulator) is publicly available from Hewlett Packard (HP) for architectural exploration and code generation. The design-time parameters include the pro-cessor's issue-width, the type of different functional units (FUs) and their la-tencies, the type and size of multiported register files, degree of pipelining, size of instruction and data memories, type of interrupt and exception systems, selection of default custom operations, datapath sharing. If the behavior of applications is not known at design-time or an application has different phases with distinct requirements, a fixed processor may not perform efficiently for all the applications/phases. To this end, we propose a run-time reconfigurable processor that can adapt its organization dynamically during execution. The run-time parameters include the processor's issue-width, the type and number of different FUs, and the register file size. Additionally, we propose config-urable fault tolerance techniques for the ρ-VEX processor. The designer can choose to include or exclude the fault tolerance in the processor at design-time. When the fault tolerance is included, it can be made permanently enabled or enabled/disabled at run-time. All these options enable users to trade-off between hardware area/resources, performance, power/energy consumption, and reliability. The processor is available as open-source. i Samenvatting In dit proefschrift stellen we voor om programmeerbaarheid te combineren met reconfigureerbaarheid door het implementeren van een aanpasbare pro-grammeerbare VLIW processor in herconfigureerbare hardware. De aanpak staat het ontwikkelen van toepassingen op hoog niveau (C programmeer taal-niveau) toe, terwijl op hetzelfde moment de processor organisatie kan worden aangepast aan de specifieke eisen (zowel statisch als dynamisch) van verschil-lende toepassingen. Onze voorgestelde aanpasbare VLIW processor, genaamd ρ-VEX, kan tijdens design-time evenals tijdens run-time aangepast worden. De instructie set archi-tectuur (ISA) is gebaseerd op de VEX ISA en een toolchain (geparametriseerde C compiler en simulator) is publiek beschikbaar gesteld door Hewlett Packard (HP) voor architectuur exploratie en code …
منابع مشابه
Aligned Scheduling: Cache-Efficient Instruction Scheduling for VLIW Processors
The performance of statically scheduled VLIW processors is highly sensitive to the instruction scheduling performed by the compiler. In this work we identify a major deficiency in existing instruction scheduling for VLIW processors. Unlike most dynamically scheduled processors, a VLIW processor with no load-use hardware interlocks will completely stall upon a cache-miss of any of the operations...
متن کاملWeld: A Multithreading Technique Towards Latency-Tolerant VLIW Processors
This paper presents a new architecture model, named Weld, for VLIW processors. Weld integrates multithreading support into a VLIW processor to hide run-time latency effects that cannot be determined by the compiler. It does this through a novel hardware technique called operation welding that merges operations from different threads to utilize the hardware resources more efficiently. Hardware c...
متن کاملA DAG-Based Design Approach for Recon gurable VLIW Processors
for Recon gurable VLIW Processors Cesare Alippi William Fornaciari Laura Pozzi Mariagiovanna Sami Dip. di Elettronica e Informazione, Politecnico di Milano, ITALY falippi,fornacia,lpozzi,[email protected] Abstract This paper explores the possibility of enabling a partial customisability of the Instruction Set of Very Long Instruction Wold processors for embedded applications, by exploiting F...
متن کاملA Vector-µSIMD-VLIW Architecture for Multimedia Applications
Media processing has motivated strong changes in the focus and design of processors. These applications are composed of heterogeneous regions of code, some of them with high levels of DLP and other ones with only modest amounts of ILP. A common approach to deal with these applications are μSIMD-VLIW processors. However, the ILP regions fail to scale when we increase the width of the machine, wh...
متن کاملEnergy-aware compilation and hardware design for VLIW embedded systems
Tomorrow’s embedded devices need to run multimedia applications demanding high computational power with low energy consumption constraints. In this context, the register file is a key source of power consumption and its inappropriate design and management severely affects system power. In this paper, we present a new approach to reduce the energy of shared register files in forthcoming embedded...
متن کامل