184 results about "Massively parallel" patented technology

A novel massively parallel supercomputer of hundreds of teraOPS-scale includes node architectures based upon System-On-a-Chip technology, i.e., each processing node comprises a single Application Specific Integrated Circuit (ASIC). Within each ASIC node is a plurality of processing elements each of which consists of a central processing unit (CPU) and plurality of floating point processors to enable optimal balance of computational performance, packaging density, low cost, and power and cooling requirements. The plurality of processors within a single node may be used individually or simultaneously to work on any combination of computation or communication as required by the particular algorithm being solved or executed at any point in time. The system-on-a-chip ASIC nodes are interconnected by multiple independent networks that optimally maximizes packet communications throughput and minimizes latency. In the preferred embodiment, the multiple networks include three high-speed networks for parallel algorithm message passing including a Torus, Global Tree, and a Global Asynchronous network that provides global barrier and notification functions. These multiple independent networks may be collaboratively or independently utilized according to the needs or phases of an algorithm for optimizing algorithm processing performance. For particular classes of parallel algorithms, or parts of parallel calculations, this architecture exhibits exceptional computational performance, and may be enabled to perform calculations for new classes of parallel algorithms. Additional networks are provided for external connectivity and used for Input / Output, System Management and Configuration, and Debug and Monitoring functions. Special node packaging techniques implementing midplane and other hardware devices facilitates partitioning of the supercomputer in multiple networks for optimizing supercomputing resources.

Systems and methods of building massively parallel computing systems using low power computing complexes in accordance with embodiments of the invention are disclosed. A massively parallel computing system in accordance with one embodiment of the invention includes at least one Solid State Blade configured to communicate via a high performance network fabric. In addition, each Solid State Blade includes a processor configured to communicate with a plurality of low power computing complexes interconnected by a router, and each low power computing complex includes at least one general processing core, an accelerator, an I / O interface, and cache memory and is configured to communicate with non-volatile solid state memory.

A massively parallel supercomputer of petaOPS-scale includes node architectures based upon System-On-a-Chip technology, where each processing node comprises a single Application Specific Integrated Circuit (ASIC) having up to four processing elements. The ASIC nodes are interconnected by multiple independent networks that optimally maximize the throughput of packet communications between nodes with minimal latency. The multiple networks may include three high-speed networks for parallel algorithm message passing including a Torus, collective network, and a Global Asynchronous network that provides global barrier and notification functions. These multiple independent networks may be collaboratively or independently utilized according to the needs or phases of an algorithm for optimizing algorithm processing performance. The use of a DMA engine is provided to facilitate message passing among the nodes without the expenditure of processing resources at the node.

Multi-gate NOR flash thin-film transistor (TFT) string arrays are organized as three dimensional stacks of active strips. Each active strip includes a shared source sublayer and a shared drain sublayer that is connected to substrate circuits. Data storage in the active strip is provided by charge-storage elements between the active strip and a multiplicity of control gates provided by adjacent local word-lines. The parasitic capacitance of each active strip is used to eliminate hard-wire ground connection to the shared source making it a semi-floating, or virtual source. Pre-charge voltages temporarily supplied from the substrate through a single port per active strip provide the appropriate voltages on the source and drain required during read, program, program-inhibit and erase operations. TFTs on multiple active strips can be pre-charged separately and then read, programmed or erased together in a massively parallel operation.

A router for interconnecting external devices coupled to the router. The router comprises a switch fabric and a plurality of routing nodes coupled to the switch fabric. Each routing node is capable of transmitting data packets to, and receiving data packets from, the external devices and is further capable of transmitting data packets to, and receiving data packets from, other routing nodes via the switch fabric. The router also comprises a control processor for comparing the N most significant bits of a first subnet address associated with a first external port of a first routing node with the N most significant bits of a second subnet address associated with a second external port of the first routing node. The router determines a P-bit prefix of similar leading bits in the first and second subnet addresses and transmits the P-bit prefix to other routing nodes.

The invention discloses a large-scale parallel program property-predication realizing method based on representative replay, belonging to the technical field of parallel program property prediction. The method comprises the following steps: collecting communication sequences and calculation vectors of parallel programs, analyzing calculation similarity of the processes, selecting a representativeprocess, recording the communication content of the representative process, using a calculation node of a target platform to replay the representative process, obtaining the sequence calculation timeof the representative process, and using the calculation time to replace the calculation time of other processes; obtaining the communication records of the parallel programs; and using a network simulator to automatically predict the final program property. By using the invention, only few hardware sources can be used, and the accurate parallel program property prediction can be obtained.