A major overhead in software DSM (Distributed Shared Memory) is the cost of remote memory accesses necessitated by the protocol as well as induced by false sharing. This paper introduces a dynamic prefetching method i...A major overhead in software DSM (Distributed Shared Memory) is the cost of remote memory accesses necessitated by the protocol as well as induced by false sharing. This paper introduces a dynamic prefetching method implemented in the JIAJIA software DSM to reduce system overhead caused by remote accesses. The prefetching method records the interleaving string of INV (invalidation) and GETP (getting a remote page) operations for each cached page and analyzes the periodicity of the string when a page is invalidated on a lock or barrier. A prefetching request is issued after the lock or barrier if the periodicity analysis indicates that GETP will be the next operation in the string. Multiple prefetching requests are merged into the same message if they are to the same host. Performance evaluation with eight well-accepted benchmarks in a cluster of sixteen PowerPC workstations shows that the prefetching scheme can significantly reduce the page fault overhead and as a result achieves a performance increase of 15%-20% in three benchmarks and around 8%-10% in another three. The average extra traffic caused by useless prefetches is only 7%-13% in the evaluation.展开更多
The performance gap between software DSM systems and message passing platforms prevents the prevalence of software DSM system greatly, though great efforts have been delivered in this area in the past decade. In this ...The performance gap between software DSM systems and message passing platforms prevents the prevalence of software DSM system greatly, though great efforts have been delivered in this area in the past decade. In this paper, we take the challenge to find where we should focus our efforts in the future design. The components of total system overhead of software DSM systems are analyzed in detail firstly. Based on a state-of-the-art software DSM system JIAJIA, we measure these components on Dawning parallel system and draw five important conclusions which are different from some traditional viewpoints. (1) The performance of the JIAJIA software DSM system is acceptable. For four of eight applications, the parallel ef ficiency achieved by JIAJIA is about 80%, while for two others, 70% efficiency can be obtained. (2) 40.94% interrupt service time is overlapped with waiting time. (3) Encoding and decoding diffs do not cost much time (<1%), so using hardware sup port to encode/decode diffs and send/receive messages is not worthwhile. (4) Great endeavours should be put to reduce data miss penalty and optimize synchronization operations, which occupy 11.75% and 13.65% of total execution time respectively.(5) Communication hardware overhead occupies 66.76% of the whole communication time in the experimental environment, and communication software overhead does not take much time as expected. Moreover, by studying the effect of CPU speed to system overhead, we find that the common speedup formula for distributed memory systems does not work under software DSM systems. Therefore, we design a new speedup formula special to software DSM systems, and point out that when the CPU speed increases the speedup can be increased too even if the network speed is fixed, which is impossible in message passing systems. Finally, we argue that JIAJIA system has desired scalability.展开更多
基金the National Natural Science Foundation of China (No.60073018).
文摘A major overhead in software DSM (Distributed Shared Memory) is the cost of remote memory accesses necessitated by the protocol as well as induced by false sharing. This paper introduces a dynamic prefetching method implemented in the JIAJIA software DSM to reduce system overhead caused by remote accesses. The prefetching method records the interleaving string of INV (invalidation) and GETP (getting a remote page) operations for each cached page and analyzes the periodicity of the string when a page is invalidated on a lock or barrier. A prefetching request is issued after the lock or barrier if the periodicity analysis indicates that GETP will be the next operation in the string. Multiple prefetching requests are merged into the same message if they are to the same host. Performance evaluation with eight well-accepted benchmarks in a cluster of sixteen PowerPC workstations shows that the prefetching scheme can significantly reduce the page fault overhead and as a result achieves a performance increase of 15%-20% in three benchmarks and around 8%-10% in another three. The average extra traffic caused by useless prefetches is only 7%-13% in the evaluation.
文摘The performance gap between software DSM systems and message passing platforms prevents the prevalence of software DSM system greatly, though great efforts have been delivered in this area in the past decade. In this paper, we take the challenge to find where we should focus our efforts in the future design. The components of total system overhead of software DSM systems are analyzed in detail firstly. Based on a state-of-the-art software DSM system JIAJIA, we measure these components on Dawning parallel system and draw five important conclusions which are different from some traditional viewpoints. (1) The performance of the JIAJIA software DSM system is acceptable. For four of eight applications, the parallel ef ficiency achieved by JIAJIA is about 80%, while for two others, 70% efficiency can be obtained. (2) 40.94% interrupt service time is overlapped with waiting time. (3) Encoding and decoding diffs do not cost much time (<1%), so using hardware sup port to encode/decode diffs and send/receive messages is not worthwhile. (4) Great endeavours should be put to reduce data miss penalty and optimize synchronization operations, which occupy 11.75% and 13.65% of total execution time respectively.(5) Communication hardware overhead occupies 66.76% of the whole communication time in the experimental environment, and communication software overhead does not take much time as expected. Moreover, by studying the effect of CPU speed to system overhead, we find that the common speedup formula for distributed memory systems does not work under software DSM systems. Therefore, we design a new speedup formula special to software DSM systems, and point out that when the CPU speed increases the speedup can be increased too even if the network speed is fixed, which is impossible in message passing systems. Finally, we argue that JIAJIA system has desired scalability.
