Advanced Computer Architecture and Parallel Processing (Wiley Series on Parallel and Distributed Computing) (v. 2)
Format: PDF / Kindle (mobi) / ePub
Computer architecture deals with the physical configuration, logical structure, formats, protocols, and operational sequences for processing data, controlling the configuration, and controlling the operations over a computer. It also encompasses word lengths, instruction codes, and the interrelationships among the main parts of a computer or group of computers. This two-volume set offers a comprehensive coverage of the field of computer organization and architecture.
together. There are two major factors used to categorize such systems: the processing units themselves, and the interconnection network that ties them together. A number of communication styles exist for multiprocessing networks. These can be broadly classiﬁed according to the communication model as shared memory (single address space) versus message passing (multiple address spaces). Communication in shared memory systems is performed by writing to and reading from the global memory, while
multiprocessor systems. Figure 2.2 shows an illustration of a single bus system. In its general form, such a system consists of N processors, each having its own cache, connected by a Figure 2.2 Example single bus system. TEAM LinG - Live, Informative, Non-cost and Genuine ! 2.2 BUS-BASED DYNAMIC INTERCONNECTION NETWORKS 21 shared bus. The use of local caches reduces the processor – memory trafﬁc. All processors communicate with a single shared memory. The typical size of such a system
varies between 2 and 50 processors. The actual size is determined by the trafﬁc per processor and the bus bandwidth (deﬁned as the maximum rate at which the bus can propagate data once transmission has started). The single bus network complexity, measured in terms of the number of buses used, is O(1), while the time complexity, measured in terms of the amount of input to output delay is O(N). Although simple and easy to expand, single bus multiprocessors are inherently limited by the bandwidth of
contention will lead to the inability to satisfy the connection 100 to 001, that is, blocking. Notice however that while connection 101 to 011 is established, the arrival of a request for a connection such as 100 to 110 can be satisﬁed. Rearrangeable Networks Rearrangeable networks are characterized by the property that it is always possible to rearrange already established connections in order to make allowance for other connections to be established simultaneously. The Benes is a well-known
unnecessary buffering in front of an idle channel is avoided. In order to reduce the size of the required buffers and decrease the incurred network latency, a technique called wormhole routing has been introduced. Here, a packet is divided into smaller units called ﬂits (ﬂow control bits). These ﬂits move in a pipeline fashion with a header ﬂit leading the way to the destination node. When the header ﬂit is blocked due to network congestion, the remaining ﬂits are also blocked. Only a buffer that