Download presentation
Presentation is loading. Please wait.
Published byBarbara Bates Modified over 9 years ago
1
07/14/08
2
2 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
3
3 Introduction
4
4 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
5
5 Cluster Definition Group of computers and servers ( loosely connected computers ) that act like a single system. computers Each system called a Node. Node contain one or more Processor, Ram,Hard disk and LAN card. Nodes work in Parallel. We can increase performance by adding more Nodes. A c
6
07/14/08 These cluster computers are referred to by many names, from a poor-man's supercomputer to COWs (clusters of workstations), and NOWs (networks of workstations). cheaper than traditional MPP systems, and often use the same processors, but are more difficult to use since the network capabilities are currently much lower. Cluster computers are also usually much smaller, most often involving fewer than 100 computers
7
07/14/08 Figure 1. A Cluster Architecture.
8
07/14/08 large Linux cluster at the Chemnitz University of Technology, GermanyLinuxChemnitz University of Technology
9
9
10
07/14/08 A simple, home-built Beowulf clusterBeowulf cluster
11
11
12
Motivation for Clusters/Grids Many science and engineering problems today require large amounts of computational resources and cannot be executed in a single machine. Large commercial supercomputers are very expensive… A lot of computational power is underutilized around the world in machines sitting idle.
13
13 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
14
14 Cluster types Load Balancing Cluster (Parallel BLAST). Computing Cluster(Parallel sequence alignment). High-availability (HA) clusters.
15
15 Cluster types:Load Balancing Cluster Task
16
07/14/08 A load balancing cluster with two servers and 4 user stations
17
07/14/08 Load-balancingLoad-balancing clusters are configurations in which cluster-nodes share computational workload to provide better overall performance. For example, a web server cluster may assign different queries to different nodes, so the overall response time will be optimized
18
18 Cluster types:Computing Cluster Task
19
07/14/08 "Computer clusters" are used for computation-intensive purposes, rather than handling IO-oriented operations such as web service or databases.IO-oriented
20
20 Cluster type:High-availability Clusters
21
07/14/08 "High-availability clusters“ improve the availability of the cluster approach. They operate by having redundant nodes, which are then used to provide service when system components failHigh-availability clustersnodes
22
22 Cluster advantages Performance. Scalability. Maintenance. Cost.
23
23 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
24
24 Node 1 switch Node 4 Node 2 Node 3 Internet Cluster
25
25 Communication : Switch 5-Port 10/100Mbps. Processor and Ram: -Master Node Duo core Processor 1.86 GHZ. Ram 1GB. -Node 1 Pentium 4 Ram 1GB. -Node 2 Pentium 4 Ram 1GB -Node 3 Pentium 4 Ram 512 MB Cluster specification
26
26 Operating System OPEN SUSE 10.3 http://software.opensuse.org/ MPICH2 http://www.mcs.anl.gov/research/projects/mpich2/ Cluster specification (cont’)
27
27 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
28
28 Performance of the Cluster is affected by 1-Node speed. 2-Running Program.
29
29 Working… Running Program(sequential)
30
30 Working… Running Program(sequential)
31
31 Working… Running Program(sequential)
32
32 Running Program(sequential)
33
33 Data sent Running Program(Parallel)
34
34 Working… Running Program(Parallel)
35
35 Finished… Results Get results… Running Program(Parallel)
36
36 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
37
37 Sequence Alignment
38
38 Sequence Alignment Used to : 1-Compare between sequences. 2-Search databases.
39
39 How to Align two Sequences. if we have two sequences A A A C G A A A T G A Let match=1, gap=-1, miss-match=0. they can be aligned as: 1- A A A C G A | | | | | | Score=3 A A T _ G A 2- A A A C _ G A | | | | | | | Score=1 A A _ _ T G A
40
40 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance Cluster Computer for Basic Problems.. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
41
41 BLAST (Basic Local Alignment Search Tool) Searching DataBases
42
42 BLAST Algorithm (High scoring pairs)
43
43 Blast search types. BLASTN - Compares a nucleotide query sequence against a nucleotide sequence database. BLASTP- Compares an amino acid query sequence against a protein sequence database. TBLASTN- Compares a protein query sequence against a nucleotide sequence Database. BLASTX- Compares nucleotide query sequence against a protein sequence database.
44
44 Why We need BLAST to be parallelized ?
45
45 Our Program:Parallel BLAST
46
46 Parallel BLAST(cont’) Formatdb.c Nucleotide sequence database “formatdb -i DATABASE -p F “. Protein sequence database “formatdb -i DATABASE -p T “.
47
47 Linux_Cluster_BLASTALL.c “blastall -p BLAST Search Type -d DATABASE -i QUERY FILE -o out. Txt” Parallel BLAST(cont’)
48
48 Results Average of running 1000 Query, 1000 times.
49
49 Results(cont’) Average of running 1000 Query, 1000 times.
50
50 Results(cont’) Average of running 1000 Query, 1000 times.
51
51 Conclusion about Parallel BLAST. Performane: Batter by using CLUSTER. Scalability:More Nodes time decrease.
52
52 Points Introduction. Cluster and Supercomputers. Cluster Types and Advantages. Our Cluster. Cluster Performance. Cluster Computer for Basic Problems. General Idea about Sequence Alignment. BLAST and Parallel BLAST Algorithm. Sequence Alignment and Parallel Sequence Alignment. Learned Skills.
53
53 Sequence Alignment Compare between sequences
54
54 Sequence Alignment Introduction. Sequence Alignment Benefits. Sequence Alignment Types.
55
55 Needleman-Wunsch Algorithm
56
56 Why We need Sequence Alignment to be parallelized ?
57
57 Parallel Sequence Alignment algorithm
58
58 Our Sequence Alignment Program Pairwise Alignment. Built Using Needleman-Wunsch algorithm.
59
59 Learned Skills. Using Linux (Suse 10.3) operating system. Programming using C language. Cluster computers and how to build one. MPICH2 for message passing interfaces between nodes. Latex. Team working, and helping each other. Presentation skills.
60
07/14/08 Message passing and communication Two widely used approaches for communication between cluster nodes are PVM, Parallel Virtual Machine :(1989)PVM provides a run- time environment for message-passing, task and resource management, and fault notification. PVM can be used by user programs written in C, C++, or Fortran, etcParallel Virtual Machine MPI, Message Passing Interface (1990)MPI implementations typically use TCP/IP and socket connections. [ MPI is now a widely available communications model that enables parallel programs to be written in languages such as C, Fortran, Python, etcMessage Passing InterfaceTCP/IP [CFortranPython
61
07/14/08 Cluster management: Task scheduling When a large multi-user cluster needs to access very large amounts of data, task scheduling becomes a challenge.task scheduling Node failure management When a node in a cluster fails, strategies such as "fencing" may be employed to keep the rest of the system operational.fencing Fencing is the process of isolating a node or protecting shared resources when a node appears to be malfunctioning.
62
Cluster Computing - challenges Software to create a single system image Fault tolerance Debugging tools Job scheduling All these have been/are being addressed since then and are leading towards a successful era for cluster computing
63
63 Thank you
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.