Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

Today's data centers may contain tens of thousands of computers with significant aggregate bandwidth requirements. The network architecture typically consists of a tree of routing and switching elements with progressively more specialized and expensive equipment moving up the network hierarchy. Unfortunately, even when deploying the highest-end IP switches/routers, resulting topologies may only support 50% of the aggregate bandwidth available at the edge of the network, while still incurring tremendous cost. Non-uniform bandwidth among data center nodes complicates application design and limits overall system performance.In this paper, we show how to leverage largely commodity Ethernet switches to support the full aggregate bandwidth of clusters consisting of tens of thousands of elements. Similar to how clusters of commodity computers have largely replaced more specialized SMPs and MPPs, we argue that appropriately architected and interconnected commodity switches may deliver more performance at less cost than available from today's higher-end solutions. Our approach requires no modifications to the end host network interface, operating system, or applications; critically, it is fully backward compatible with Ethernet, IP, and TCP.

/pdf/a-scalable-commodity-data-center-network-architecture-5evikm9tu4.pdf

A scalable, commodity data center network architecture

Power dissipation is a fundamental problem for nanoelectronic circuits. Scaling the supply voltage reduces the energy needed for switching, but the field-effect transistors (FETs) in today's integrated circuits require at least 60 mV of gate voltage to increase the current by one order of magnitude at room temperature. Tunnel FETs avoid this limit by using quantum-mechanical band-to-band tunnelling, rather than thermal injection, to inject charge carriers into the device channel. Tunnel FETs based on ultrathin semiconducting films or nanowires could achieve a 100-fold power reduction over complementary metal-oxide-semiconductor (CMOS) transistors, so integrating tunnel FETs with CMOS technology could improve low-power integrated circuits.

Tunnel field-effect transistors as energy-efficient electronic switches

Concern about the performance of wires wires in scaled technologies has led to research exploring other communication methods. This paper examines wire and gate delays as technologies migrate from 0.18-/spl mu/m to 0.035-/spl mu/m feature sizes to better understand the magnitude of the the wiring problem. Wires that shorten in length as technologies scale have delays that either track gate delays or grow slowly relative to gate delays. This result is good news since these "local" wires dominate chip wiring. Despite this scaling of local wire performance, computer-aided design (CAD) tools must still become move sophisticated in dealing with these wires. Under scaling, the total number of wires grows exponentially, so CAD tools will need to handle an ever-growing percentage of all the wires in order to keep designer workloads constant. Global wires present a more serious problem to designers. These are wires that do not scale in length since they communicate signals across the chip. The delay of these wives will remain constant if repeaters are used meaning that relative to gate delays, their delays scale upwards. These increased delays for global communication will drive architectures toward modular designs with explicit global latency mechanisms.

The future of wires

Steep subthreshold swing transistors based on interband tunneling are examined toward extending the performance of electronics systems. In particular, this review introduces and summarizes progress in the development of the tunnel field-effect transistors (TFETs) including its origin, current experimental and theoretical performance relative to the metal-oxide-semiconductor field-effect transistor (MOSFET), basic current-transport theory, design tradeoffs, and fundamental challenges. The promise of the TFET is in its ability to provide higher drive current than the MOSFET as supply voltages approach 0.1 V.

/pdf/low-voltage-tunnel-transistors-for-beyond-cmos-logic-3w5c8knrc5.pdf

Low-Voltage Tunnel Transistors for Beyond CMOS Logic

According to one embodiment of the present invention, a method for refreshing memory includes receiving a synchronization command at a memory device. An internal refresh timer is reset within the memory device based on receiving the synchronization command. An internal refresh trigger is generated within the memory device based on the internal refresh timer reaching a predetermined value. A refresh of a memory array is performed within the memory device based on the internal refresh trigger.

Memory device refresh

A system and method for providing a memory device having a shared error feedback pin. The system includes a memory device having a data interface configured to receive data bits and CRC bits, CRC receiving circuitry, CRC creation circuitry, a memory device pad, and driver circuitry. The CRC receiving circuitry utilizes a CRC equation for the detection of errors in one or more of the received data and the received CRC bits. The CRC creation circuitry utilizes the CRC equation for the creation of CRC bits consistent with data to be transmitted to a separate device bits. The memory device pad is configured for reporting of any errors detected in the received data and the received CRC bits. The driver circuitry is connected to the memory device pad and merged with one or more other driver circuitries resident on one or more other memory devices into an error reporting line.

Providing a memory device having a shared error feedback pin

A method and apparatus are provided which are implemented in a chip I/O buffer-multiplexor circuit or I/O buffer cell 201. The I/O buffer portion includes a receiver circuit 205 for receiving bus input signals to the buffered chip, and a driver circuit 203 for driving output signals from the buffered chip to a data bus. An integrated multiplexor or MUX circuit 207 selectively gates one of three possible signals to chip internal logic. The three signals applied to the MUX circuit include a boundary scan test signal BS MUX for testing scan points in an integrated circuit, a bypass Data In signal DI which is generated by chip internal drive logic, and a DQ signal received by the I/O buffer receiver circuit from a data bus. The data input node of the I/O buffer is wired directly to the new multiplexor data input. Additional control signals are provided for orthogonal selection of the three multiplexor data inputs.

Bidirectional off-chip driver with receiver bypass

The IBM Blue Gene®/Q supercomputer is designed for highly efficient computing for problems dominated by floating-point computation. Its target mean time between failures for a 96-rack, 98, 304-node system is three days, allowing tasks requiring computation for many days to run at scale, with little time wasted on checkpoint-restart operations. This paper describes various elements of the compute application-specific integrated circuit and the system package, and how they contribute to low power consumption and high reliability.

Packaging the IBM Blue Gene/Q supercomputer

The providing of an array interface of conductive joint members for use in forming interconnections between mating surfaces such as a pad on a surface mount electronic device and contacts on a circuit card where one portion of the conductive joint members are of a relatively elongated or oval outline and are oriented with the longer dimension in one direction to accommodate wiring spacing and another portion oriented in a different direction for accommodating expansion stress. In manufacturing when the relatively elongated shape is oriented with the longer dimension along the wiping motion direction in a screen type forming of the conductive joint members the slurry of material that is to be the conductive joint members fills the openings in the screen more reliably and the areas of the conductive members are more uniform. The invention provides the advantages of an increase in the number of wiring lines, an increase in uniformity of wiped screen deposition conductive joint member formation, ability to employ more than one out of chip and wiring levels in expansion mismatch stress relief, and ability by conductive joint member dimensional alignment to improve reliability and flexibility.

Paul W. Coteus

Papers

Memory device refresh

Providing a memory device having a shared error feedback pin

Bidirectional off-chip driver with receiver bypass

Packaging the IBM Blue Gene/Q supercomputer

Stress accommodation in electronic device interconnect technology for millimeter contact locations