Found 312 Articles for Computer Architecture

Myrinet is a high-performance, packet-communication, and switching technology. It was produced by Myricom as a high-performance alternative to conventional Ethernet networks. Myrinet switches are multiple-port components that route a packet entering on an input channel of a port to the output channel of the port selected by the packet.Myrinet switches have 4, 8, 12, 16 ports. For an n-port switch, the ports are addressed 0, 1, 2... n - 1. For any switching permutation, there may be as many packets traversing a switch concurrently as the switch has ports. These switches are implemented using two types of VLSI chips such ... Read More

An n1 x n2 switch consists of n1 input ports, n2 output ports, links connecting each input to every output, control logic to select a specific connection, and internal buffers.A switch is used to establish connections from the input ports to the output ports. These connections may be one-to-one, which represents point-to-point connections, or one-to-many, which represents multicast or broadcast. The case of many-to-one should cause conflicts at the output ports and therefore needs arbitration to resolve conflicts if allowed. When only one-to-one connections are allowed, the switch is called a crossbar. An n x n crossbar switch can establish ... Read More

Ethernet is a packet-switched LAN technology introduced by Xerox PARC in the early 1970s. Ethernet was designed to be a shared bus technology where multiple hosts are connected to a shared communication medium. All hosts connected to an Ethernet receive every transmission, making it possible to broadcast a packet to all hosts at the same time.Ethernet uses a distributed access control scheme called Carrier Sense Multiple Access with Collision Detect (CSMA/CD). Multiple machines can access an Ethernet at the same time. Each machine senses whether a carrier wave is present to determine whether the network is idle before it sends ... Read More

A cluster is a collection of stand-alone computers connected using some interconnection network. Each node in a cluster could be a workstation, personal computer, or even a multiprocessor system.A node is an autonomous computer that may be engaged in its private activities while at the same time cooperating with other units in the context of some computational task. Each node has its input/output systems and its operating system.When all nodes in a cluster have the same architecture and run the same operating system, the cluster is called homogeneous, otherwise, it is heterogeneous. The interconnection network could be a fast LAN ... Read More

Parallel applications can be designed using the client/server model. A client may divide a big application into several smaller problems that can be processed by multiple servers simultaneously. All the servers compute the solution to their respective problems and send their results to the client.The client assembles the results from each server and outputs the final result to the user. The client acts as the master (supervisor) while the servers act as the slaves (workers) in the master-slave (supervisor-workers) model as shown in the figure. The steps are taken at the client and each server is summarized as follows.Client (Supervisor)Client ... Read More

Sockets are used to provide the capability of making connections from one application running on one machine to another running on a different machine. A socket abstraction consists of the data structure that holds the information needed for communication, and the system calls that manipulate the socket structure. Once a socket is created, it can be used to wait for an incoming connection (passive socket) or can be used to initiate a connection (active socket).A client can establish an active connection to a remote server by creating an instance of a socket. To establish a server connection and bind it ... Read More

Several possible problems can result from the use of certain routing mechanisms in message-passing systems. These include deadlock, livelock, and starvation, which are as follows −Deadlock − When two messages each hold the resources required by the other to move, both messages will be blocked. This is called a deadlock. It is a phenomenon that occurs whenever there exists a cyclic dependency on resources. Management of resources in a network is the responsibility of the flow control mechanism used. Resources must be allocated in a manner that avoids deadlock.Livelock − Livelock describes a situation in which a message keeps going ... Read More

Shared memory systems can be designed using bus-based or switch-based interconnection networks. The simplest network for shared memory systems is the bus. The bus/cache architecture alleviates the requirement for expensive multiport memories and interface circuitry and the need to adopt a message-passing paradigm when developing application software.The bus may get saturated if multiple processors are trying to access the shared memory (via the bus) simultaneously. A typical bus-based design uses caches to solve the bus contention problem. High-speed caches connected to each processor on one side and the bus on the other side mean that local copies of instructions and ... Read More

Benchmark performance refers to the use of a set of integer and floating-point programs (known collectively as a benchmark) that are designed to test different performance aspects of the computing system(s) under test. Benchmark programs should be designed to provide fair and effective comparisons among high-performance computing systems. For a benchmark to be meaningful, it should evaluate faithfully the performance for the intended use of the system.The examples of benchmarks are the Dhrystone and Whetstone benchmarks. These are synthetic (not real) benchmarks intended to measure the performance of real machines.The Dhrystone benchmark addresses integer performance. It consists of 100 statements ... Read More

It is a point-to-point connection to other nodes or devices. All the network nodes are connected. Mesh has n (n-1)/2 physical channels to link n devices. Mesh architecture with wrap-around connections forms a torus. Several routing mechanisms have been used to route messages around meshes. One such routing mechanism is known as dimension-ordering routing. Using this technique, a message is routed in one given dimension at a time, arriving at the proper coordinate in each dimension before proceeding to the next dimension.There are two techniques to transmit data over the Mesh topology that are as follows −RoutingIn routing, the nodes have a routing logic, ... Read More