Improving The Communication Performance Of I/O Intensive And Communication Intensive Application In Cluster Computer SystemsKumar, V Santhosh 10 1900 (has links)
Cluster computer systems assembled from commodity off-the-shelf components have emerged as a viable and cost-effective alternative to high-end custom parallel computer systems.In this thesis, we investigate how scalable performance can be achieved for database systems on clusters. In this context we specﬁcally considered database query processing for evaluation of botlenecks and suggest optimization techniques for obtaining scalable application performance. First we systematically demonstrated that in a large cluster with high disk bandwidth, the processing capability and the I/O bus bandwidth are the two major performance bottlenecks in database systems. To identify and assess bottlenecks, we developed a Petri net model of parallel query execution on a cluster. Once identiﬁed and assessed,we address the above two performance bottlenecks by offoading certain application related tasks to the processor in the network interface card. Offoading application tasks to the processor in the network interface cards shifts the bottleneck from cluster processor to I/O bus. Further, we propose a hardware scheme,network attached disk ,and a software scheme to achieve a balanced utilization of re-sources like host processor, I/O bus, and processor in the network interface card. The proposed schemes result in a speedup of upto 1.47 compared to the base scheme, and ensures scalable performance upto 64 processors. Encouraged by the beneﬁts of ofﬂoading application tasks to network processors, we explore the possibilities of performing the bloom ﬁlter operations in network processors. We combine ofﬂoading bloom ﬁlter operations with the proposed hardware schemes to achieve upto 50% reduction in execution time. The later part of the thesis provides introductory experiments conducted in Community At-mospheric Model(CAM), a large scale parallel application used for global weather and climate prediction. CAM is a communication intensive application that involves collective communication of large messages. In our limited experiment, we identiﬁed CAM to see the effect of compression techniques and ofﬂoading techniques (as formulated for database) on the performance of communication intensive applications. Due to time constraint, we considered only the possibility of compression technique for improving the application performance. However, ofﬂoading technique could be taken as a full-ﬂedged research problem for further investigation In our experiment, we found compression of messages reduces the message latencies, and hence improves the execution time and scalability of the application. Without using compression techniques, performance measured on 64 processor cluster resulted in a speed up of only 15.6. While lossless compression retains the accuracy and correctness of the program, it does not result in high compression. We therefore propose lossy compression technique which can achieve a higher compression, yet retain the accuracy and numerical stability of the application while achieving a scalable performance. This leads to speedup of 31.7 on 64 processors compared to a speedup of 15.6 without message compression. We establish that the accuracy within prescribed limit of variation and numerical stability of CAM is retained under lossy compression.
Role Of Sea Surface Temperature Gradient In Intraseasonal Oscillation Of Convection In An Aquaplanet ModelDas, Surajit 09 1900 (has links) (PDF)
In this thesis we examine intra-seasonal oscillations (ISO) in the aqua-planet setup of the Community Atmospheric Model (CAM) version 5.1, mainly based on July and January climatological sea surface temperature (SST). We investigate mainly two questions -what should be the SST distribution for the existence of (a) northward moving ISO in summer, and (b) eastward moving MJO-like modes in winter. In the first part of the thesis we discuss the northward propagation. A series of experiments were performed with zonally symmetric and asymmetric SST distributions. The basic lower boundary condition is specified from zonally averaged observed July and January SST. The zonally symmetric July SST experiment produced an inter tropical convergence zone (ITCZ) on both sides of the equator. Poleward movement is not clear, and it is confined to the region between the double ITCZ. In July, the Bay of Bengal (BOB) and West Pacific SST is high compared to the rest of the northern tropics. When we impose a zonally asymmetric SST structure with warm SST spanning about 80 of longitude, the model shows a monsoon-like circulation, and some northward propagating convective events. Analysis of these events shows that two adjacent cells with cyclonic and anticyclonic vorticity are created over the warm SST anomaly and to the west. The propagation occurs due to the convective region drawn north in the convergence zone between these vortices. Zonally propagating Madden-Julian oscillations (MJO) are discussed in the second part of the thesis. All the experiments in this part are based on the zonally symmetric SST. The zonally symmetric January SST configuration gives an MJO-like mode, with zonal wave number 1 and a period of 40-90 days. The SST structure has a nearly meridionally symmetric structure, with local SST maxima on either side of the equator, and a small dip in the equatorial region. If we replace this dip with an SST maximum, the time-scale of MJO becomes significantly smaller (20-40 days). The implication is that an SST maximum in the equatorial region reduces the strength of MJO, and a flat SST profile in the equatorial region is required for more energetic of MJO. This result was tested and found to be valid in a series of further experiments.
Page generated in 3.9561 seconds