Search engine for discovering works of Art, research articles, and books related to Art and Culture
ShareThis
Javascript must be enabled to continue!

HAD:A Prototype Of Dataflow Compute Architecture

View through CrossRef
Abstract To investigate the features, implementation, and applications of data flow architecture, a novel dataflow computing system, HAD (Hardware Accelerated Dataflow), is proposed. This system enhances the traditional control flow processor by integrating a dataflow component.HAD introduces two specialized instructions to the RISC ISA and provides OS system calls for applications. Users can execute programs that adhere to the dataflow execution model on the system, enabling synchronization and mutual exclusion between processes through the inherent characteristics of dataflow computing.We further analyze how our dataflow computing can reduce the overhead of OS system calls for synchronization and mutual exclusion between processes.Lastly, we perform simulations using Verilator, implement the system on an FPGA device based on the RISC-V prototype architecture, and transplant a miniature operating system for it. We evaluate its performance using EPCC Microbenchmarks. The simulation and physical experiment results indicate that when the number of processes is substantial, the system exhibits a time reduction of up to 29\% compared to the traditional approach. The simulation results closely correspond to the actual outcomes.
Title: HAD:A Prototype Of Dataflow Compute Architecture
Description:
Abstract To investigate the features, implementation, and applications of data flow architecture, a novel dataflow computing system, HAD (Hardware Accelerated Dataflow), is proposed.
This system enhances the traditional control flow processor by integrating a dataflow component.
HAD introduces two specialized instructions to the RISC ISA and provides OS system calls for applications.
Users can execute programs that adhere to the dataflow execution model on the system, enabling synchronization and mutual exclusion between processes through the inherent characteristics of dataflow computing.
We further analyze how our dataflow computing can reduce the overhead of OS system calls for synchronization and mutual exclusion between processes.
Lastly, we perform simulations using Verilator, implement the system on an FPGA device based on the RISC-V prototype architecture, and transplant a miniature operating system for it.
We evaluate its performance using EPCC Microbenchmarks.
The simulation and physical experiment results indicate that when the number of processes is substantial, the system exhibits a time reduction of up to 29\% compared to the traditional approach.
The simulation results closely correspond to the actual outcomes.

Related Results

The architecture of differences
The architecture of differences
Following in the footsteps of the protagonists of the Italian architectural debate is a mark of culture and proactivity. The synthesis deriving from the artistic-humanistic factors...
Software synthesis from dataflow schedule graphs
Software synthesis from dataflow schedule graphs
AbstractThe dataflow-model of computation is widely used in design and implementation of signal processing systems. In dataflow-based design processes, scheduling—the assignment an...
Fine Grain Algorithm Parallelization on a Hybrid Control-flow and Dataflow Processor
Fine Grain Algorithm Parallelization on a Hybrid Control-flow and Dataflow Processor
Abstract The execution time of a high performance computing algorithm depends on multiple factors: the algorithm scalability, the chosen hardware, the communication speed b...
Architecture between heteronomy and self-generation
Architecture between heteronomy and self-generation
Introduction «I have never worked in the technocratic exaltation, solving a constructive problem and that’s it. I’ve always tried to interpret the space of human life» (Vitto...
DataFlow based- Automatic parallelization of MATLAB/Simulink models for Fitting Modern Multicore Architectures
DataFlow based- Automatic parallelization of MATLAB/Simulink models for Fitting Modern Multicore Architectures
Abstract In many fields including aerospace, automotive, and telecommunications, Math-Works’ MATLAB/Simulink is the current de facto standard for model-based design. The st...
Kievan Rus’
Kievan Rus’
Robert Ousterhout, the author of a magnificent book “Eastern Medieval Architecture. The Building Traditions of Bizantium and Neighboring Lands”, published by Oxford University Pres...
Dataflow Processing in Memory Achieves Significant Energy Efficiency
Dataflow Processing in Memory Achieves Significant Energy Efficiency
The large difference between processor CPU cycle time and memory access time, often referred to as the memory wall, severely limits the performance of streaming applications. Some ...
A Compute-In-Memory Architecture and System-Technology Codesign Simulator Based 3D NAND Flash
A Compute-In-Memory Architecture and System-Technology Codesign Simulator Based 3D NAND Flash
The rapid advancement of large language models (LLM) such as ChatGPT has imposed unprecedented demands on hardware in terms of computational power, memory capacity, and energy effi...

Back to Top