Title :
Shared memory heterogeneous computation on PCIe-supported platforms
Author :
Shukla, Sandeep K. ; Yang Yang ; Bhuyan, Laxmi N. ; Brisk, Philip
Abstract :
Domain-disparity between CPU and Hardware Accelerators(HA) leads to CPU under-utilization and inter-domain data copy overheads. By exposing HA memory to OS and host MMU, these overheads can be eliminated. In this paper, we present a shared virtual memory real system design for PCIe-based HAs to enable parallel heterogeneous execution in CPU and HAs without driver overheads. We extend Linux with a custom memory manager and scheduler to manage HA memory and application-cores respectively. Our FPGA-based multi-application logic design supports simultaneous execution of multiple heterogeneous applications. We show the advantages of heterogeneous execution and analyze how our design reduces OS overhead.
Keywords :
field programmable gate arrays; logic design; shared memory systems; CPU; FPGA-based multiapplication logic design; Linux; PCIe-supported platform; custom memory manager; domain-disparity; hardware accelerator; interdomain data copy overhead; parallel heterogeneous execution; shared memory heterogeneous computation; shared virtual memory real system; Engines; Field programmable gate arrays; Hardware; Instruction sets; Kernel; Memory management; Support vector machines;
Conference_Titel :
Field Programmable Logic and Applications (FPL), 2013 23rd International Conference on
Conference_Location :
Porto
DOI :
10.1109/FPL.2013.6645580