A deadlock detection and recovery algorithm using the formalism of a directed graph matrix

International Journal of Parallel, Emergent and Distributed Sys

2015

This article presents a GPU-based single-unit deadlock detection methodology and its algorithm, GPU-OSDDA. Our GPU-based design utilizes parallel hardware of GPU to perform computations and thus is able to overcome the major limitation of prior hardware-based approaches by having the capability of handling thousands of processes and resources, whilst achieving real-world run-times. By utilizing a bit-vector technique for storing algorithm matrices and designing novel, efficient algorithmic methods, we not only reduce memory usage dramatically but also achieve two orders of magnitude speedup over CPU equivalents. Additionally, GPU-OSDDA acts as an interactive service to the CPU, because all of the aforementioned computations and matrix management techniques take place on the GPU, requiring minimal interaction with the CPU. GPU-OSDDA is implemented on three GPU cards: Tesla C2050, Tesla K20c, and Titan X. Our design shows overall speedups of 6-595X over CPU equivalents.

Section: Related Workmentioning

confidence: 99%

“…In the past, many software-based deadlock detection algorithms [2][3][4][5][6] were written, but they lacked the speed necessary to make them viable in real world systems. As a result, researchers have developed hardware-based algorithms [7][8][9][10][11] that expanded upon the findings of these software algorithms.…”

Section: Introductionmentioning

confidence: 99%

GPU-OSDDA: a bit-vector GPU-based deadlock detection algorithm for single-unit resource systems

Abell

International Journal of Parallel, Emergent and Distributed Sys

2015

“…Resource Allocation Graph (RAG) representation was utilized to perform deadlock detection in both [12] and [4]. An adjacency matrix representation was leveraged with matrix multiplication in an algorithm proposed by Leibfried [7], which has a runtime complexity of Oðm 3 Þ. In 1991, an algorithm that detects deadlock in Oð1Þ runtime was proposed by Kim and Koh [6].…”

Section: Related Workmentioning

confidence: 99%

A True O(1) Parallel Deadlock Detection Algorithm for Single-Unit Resource Systems and Its Hardware Implementation

Xiao

IEEE Trans. Parallel Distrib. Syst.

2010

Due to rapid technology advance, Multiprocessor System-on-Chips (MPSoCs) are likely to become commodity computing platforms for embedded applications. In the future, it is possible that an MPSoC is equipped with a large number of processing elements as well as on-chip resources. The management of these faces many challenges, among which deadlock is one of the most crucial issues. This paper presents a novel hardware-oriented deadlock detection algorithm suitable for current and future MPSoCs. Unlike previously published methods whose runtime complexities are often affected by the number of processing elements and resources in the system, the proposed algorithm leverages specialized hardware to guarantee O(1) overall runtime complexity. Such complexity is achieved by: 1) classifying resource allocation events; 2) for each type of events, using hardware to perform a set of specific detection and/or preparation operations that only takes constant runtime; and 3) updating necessary information for multiple resources in parallel in hardware. We implement the algorithm in Verilog HDL and demonstrate through simulation that each algorithm invocation takes at most four clock cycles.

“…Based on the same RAG representation, Holt [3] devised a deadlock detection algorithm with a reduced O(m × n) run-time complexity for multi-unit resource systems. Leibfried [7] introduced an adjacency matrix representation for resource allocations and performed deadlock detection in the means of matrix multiplication, which has an O(m 3 ) run-time complexity. A decade ago, having extended his previous work [5], Kim presented an algorithm that detects deadlock in O(1) runtime for multi-unit resource systems [4], but its overall runtime remains O(m×n), which has not been improved since Holt's work [3].…”

Section: Related Workmentioning

confidence: 99%

A novel O(1) parallel deadlock detection algorithm and architecture for multi-unit resource systems

Xiao

2007 25th International Conference on Computer Design

2007

This paper introduces a novel O(1) parallel deadlock detection approach for multi-unit resource System-on-aChips (SoCs), inspired by Kim's method in O(1) detection as well as Shiu's method in parallel processing. Our contributions are (i) the first O(1) hardware deadlock detection and (ii) O(min(m, n)) preparation, both for multiunit resource systems, where m and n are the number of processes and resources, respectively. O(min(m, n)), previously O(m × n), is achieved by performing all the searches for sink nodes for each and every resource in parallel in hardware over a matrix representing resource allocations as well as other auxiliary matrices. Our experiments demonstrate that deadlock detection always takes two clock cycles.