Upload
dina-morrison
View
216
Download
0
Tags:
Embed Size (px)
Citation preview
High Performance ComputingPresented To Mam Saman Iftikhar
Presented BY Siara Nosheen MSCS 2nd sem 2514
2
MotivationWhy high amount of computation is
needed?
Genetic engineering: Searching for matching DNA patterns in large DNA banks.
Cosmology: Simulations on very complex systems, such as simulating the formation of a galaxy • Climate: Solving very high
precision floating point calculations, simulating
chaotic systems.
• Financial modeling and commerce: Simulating
chaotic systems, like climate modeling problem.
• Cryptography: Searching very large state spaces, to find out the cryptographic key; factoring very large
numbers.
Software: Searching large state spaces for evaluating and verifying the software.
3
• Ways to improve performance?
Work harder
Work smarter
Get help
Hardware improvements
Better algorithms
Parallelism
High Performance Computing(HPC)
Types of Parallel Architecture
Forms of HPC
• The commodity HPC cluster • Dedicated Supercomputer• HPC in Cloud Computing• Grid computing
Who Uses HPC Today's?
Successful HPC applications span many industrial, government, and academic sectors. The following is a list of major areas where HPC has a significant presence:
• Bio-sciences and the human genome: Drug discovery, disease detection/prevention
• Computer aided engineering (CAE): Automotive design and testing, transportation, structural, mechanical design
• Chemical engineering: Process and molecular design
• Digital content creation (DCC) and distribution: Computer aided graphics in film and media Economics/financial: Wall Street risk analysis, portfolio management, automated trading
Who Uses HPC Today's?
• Electronic design and automation (EDA): Electronic component design and verification
• Geosciences and geo-engineering: Oil and gas exploration and reservoir modeling
• Mechanical design and drafting: 2D and 3D design and verification, mechanical modeling
• Defense and energy: Nuclear stewardship, basic and applied research
• Government labs: Basic and applied research
• University/academic: Basic and applied research
• Weather forecasting: Near term and climate/earth modeling
9
ClusterA cluster is a type of parallel and distributed processing system, which consists of a collection of interconnected stand-alone computers working together as a single, integrated computing resource.
10
A cluster architecture
High speed interconnect
Computing nodes
•Myricom—1.28 Gbps in each direction•IEEE SCI latency under 2.5 microseconds, 3.2 Gbps each direction (ring or torus topology)•Ethernet-star topology
• In most cases limitation is the server’s internal PCI bus system.
Cluster Middleware
• To support Single System Image (SSI)
• Resource management and scheduling software
• -Initial installation• -Administration• -Scheduling• -Allocation of
hardware • -Allocation software
components
Parallel programming environments and tools
• Compilers• Parallel Virtual
Machine (PVM) • Message
Passing Interface (MPI)
Parallel and sequential applications
Master nodes
11
• High performance
• High availability
High computing capability
•Consider the fail possibility of each hardware of software •Includes redundancy•A subset of this type is the load balancing clusters •Typically used for business applications—web servers
Types of Clusters
12
• Homogeneous clusters:
• Heterogeneous clusters:
•In homogeneous clusters all nodes have similar properties. Each node is much like any other. Amount of memory and interconnects are similar.
Nodes have different characteristics, in the sense of memory and interconnect performance.
Types of Clusters
13
• Single-tier clusters:
• Multi-tier clusters:
•There is no hierarchy of nodes is defined. Any node may be used for any purpose. The main advantage of the single tier cluster is its simplicity. The main disadvantage is its limit to be expanded.
There is a hierarchy between nodes. There are node sets, where each set has a specialized function
Types of Clusters
14
• Multiple Instruction Multiple Data (MIMD)
• Distributed Memory Processing (DMP)
•Each of the nodes has its own instruction memory and data memory.
•Programs can not directly access the memory of remote systems in the cluster. They have to use a kind of message passing between nodes.
Clusters in the Flynn’s Taxanomy
15
Benefits of Clusters
• Ease of building: – No expensive and long development projects.
• Price performance benefit: – Highly available COTS products are used.
• Flexibility of configuration: – Number of nodes, nodes’ performance, inter-
connection topology can be upgraded. System can be modified without loss of prior work. •Scale up: Increasing the throughput
of each computing node.
Scale out: Increase the number of computing nodes. Requires efficient i/o between nodes and cost effective management of large number of nodes.
16
Efficiency of a Cluster
• Cluster throughout is a function of the following• CPUs: Total number and speed of cpus• Efficiency of the parallel algorithms• Inter-Process Communication: Efficiency of the
inter-process communication between the computing nodes
• Storage I/O: Frequency and size of input data reads and output data writes
• Job Scheduling: Efficiency of the scheduling
Figure: Ranger provides computational capabilities to the national research community and is built using the Sun Constellation System architecture from Sun Microsystems.
Figure: Typical 1U Cluster Node — the Sun Fire X2200 M2 server(1U indicates that the server is 1.75 inches high)
Clusters categories
Clusters are often broken into two Categories • Capability • Capacity
Clusters categories
A capability cluster is designed to handle (or be capable of handling) large compute jobs that may employ every node in the cluster.
Clusters categories
• Capacity clusters, on the other hand, are those that are used to deliver a certain amount of computing capacity to the end users. For instance, a capacity cluster may support hundreds of users running any number of jobs that require a smaller number of nodes. Most clusters are used in a capacity fashion.
Figure : The Sun Blade 6048 chassis holds up to 48 blade server modules,1,152 cores, delivering up to 12 TFLOPS in a single rack.
HPC Building Blocks
• Choosing Cluster Hardware• Finding Cluster Software
HPC Building Blocks
• Choosing Cluster Hardware– Crunching numbers: processors and nodes– A little help: Co-processors– The communication: interconnects– Remembering the Storage– Racking and Stacking– Power and cooling
HPC Building Blocks• Finding Cluster Software
– Operating Systems – HPC glue: Cluster software– File systems– Sharing is caring: HPC recourse schedulers– Ready to run Application Software – Provisioning: Creating the Cluster
• Local• Remote
– Cluster Tool Kits• Sun HPC Software, Linux Edition• Rocks Clusters• Oscar
Work Hard is the key of sccuess.