Advanced Pattern based Memory Controller for FPGA based HPC applications

Sönmez

2014 International Conference on ReConFigurable Computing and FPGAs (ReConFig14)

et al. 2014

Self Cite

In this paper, we propose a hardware mechanism for embedded multi-core memory system called Pattern Aware Memory System (PAMS). The PAMS supports static and dynamic data structures using descriptors and specialized memory and reduces area, cost, energy consumption and hit latency. When compared with a Baseline Memory System, the PAMS consumes between 3 and 9 times and 1.13 and 2.66 times less program memory for static and dynamic data structures respectively. The benchmarking applications (having static and dynamic data structures) results show that PAMS consumes 20% less hardware resources, 32% less on chip power and achieves a maximum speedup of 52x and 2.9x for static and dynamic data structures respectively. The results show that the PAMS multi-core system transfers data structures up to 4.65x faster than the MicroBlaze baseline system.

Section: A Memory Organizationmentioning

confidence: 99%

Section: B Data Structures and Access Descriptionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

PAMS: Pattern Aware Memory System for embedded systems

Sönmez

2014 International Conference on ReConFigurable Computing and FPGAs (ReConFig14)

et al. 2014

Self Cite

“…MAPC uses the Pattern Descriptor Unit (PDU) to hold the information of complex memory access patterns in descriptors [7] [8]. The set of parameters for a descriptor block includes local address, base address, size, stride and offset (shown in Figure 2).…”

Section: B Pattern Descriptor Unitmentioning

confidence: 99%

MAPC: Memory access pattern based controller

2014 24th International Conference on Field Programmable Logic and Applications (FPL)

Unsal

et al. 2014

Self Cite

Traditionally, system designers have attempted to improve system performance by scheduling the processing cores and by exploring different memory system configurations and there is comparatively less work done scheduling the accesses at the memory system level and exploring data accesses on the memory systems. In this paper, we propose a memory access pattern based controller (MAPC). MAPC organizes data accesses in descriptors, prioritizes them with respect to the number and size of transfer requests. When compared to the baseline multicore system, the MAPC based system achieves between 2.41x to 5.34x of speedup for different applications, consumes 28% less hardware resources and 13% less dynamic power.

“…2) Specialized Memory: The ViPS Specialized Memory [14] (SM) is directly connected to the Processing Unit and provides single cycle data access. Like a cache, the SM temporarily holds data to speed up later accesses.…”

Section: B Memory Unitmentioning

confidence: 99%

ViPS: Visual processing system for medical imaging

2015 8th International Conference on Biomedical Engineering and Informatics (BMEI)

Cristal

et al. 2015

Self Cite

Abstract-Imaging has become an indispensable tool in modern medicine. Various powerful and expensive platforms to study medical imaging applications appear in recent years. In this article, we design and propose a Visual Processing System (ViPS) that processes medical imaging applications efficiently. ViPS provides a user-friendly programming environment and highperformance architecture to perform image analysis, features extraction and object recognition for complex real-time images or videos. The data structure of image or video is described in the program memory using pattern descriptors; ViPS uses specialized 3D memory structure to handle complex images or videos and processes them on microprocessors or application specific hardware accelerators. The proposed system is highly reliable in terms of cost, performance, and power. ViPS based system is implemented and tested on a Xilinx Virtex-7 FPGA VC707 Evaluation Kit. The performance of ViPS is compared with the Intel i7 multi-core, GPU Jetson TK1 Embedded Development Kit with 192 CUDA cores based graphic systems. When compared with the Intel and GPU-based systems, the results show that ViPS performs real-time video reconstruction at 2x and 1.45x of higher frame rate, achieves 14.6x to 4.8x of speedup while executing different image processing applications and 20.3% and 12.6% of speedup for video processing algorithms respectively.