SC|09 Disruptive Technologies Presentations
For last half century, we’ve witnessed exponential growth in the scale and capability of semiconductor electronics and related technologies, a phenomenon commonly referred to as Moore’s Law. This institutionalized revolution in technology has enabled extraordinary changes in almost all facets of modern society. Steady and predictable improvements in Silicon VLSI have largely squeezed out other, potentially disruptive competing technologies. The ITRS roadmap suggests that this will continue to be true for another decade.
This is not to suggest that another decade of Moore’s Law will not bring disruptive change. Power density has constrained processor clock rates, and increased performance now comes from an exponential growth in the number of cores. The density of logic is increasing faster than that of DRAM, and the familiar ratio of one Byte per Flop/s is increasingly unaffordable. Technologies such as NAND-Flash are appearing in systems, and we may soon enjoy persistent main memory.
Therefore, the first objective of this year’s Disruptive Technologies talks will be present a vision of how another decade of Moore’s Law will change the technology underlying today’s systems. Wilfried Haensch from IBM Research will present how VLSI will likely evolve out to the end of the next decade, and the impact that will have on the logic that underpins computing and networking systems. Dean Klien from Micron will describe how DRAM will evolve, and what new technologies might soon compliment it. Finally, Keren Bergman from Columbia University will discuss how similar rapid changes in optics will impact the networks which tie our systems together.
The second objective of Disruptive Technologies at SC|09 will be to understand how another decade of Moore’s Law will impact system software and applications. Thomas Sterling from Louisiana State University will describe how the exponential growth in the number of cores could lead to a new model of execution. Vivek Sarkar from Rice will discuss how systems software, such as the operating system, will likely evolve. Finally, Tim Mattson from Intel will discuss changes in how the developers of applications, ranging games to large-scale scientific codes, are already adapting to the challenge of programming extreme parallel systems.
SC|09 Disruptive Technologies Exhibits
There will once again be a Disruptive Technology exhibit at SC|09. This will feature innovative new technology, often a preview of what will be presented in larger SC|09 exhibits. The technology presented will address a number of concerns ranging from maximizing the PUE of servers to running Matlab scripts on GPUs.
Green Revolution Cooling will demonstrate immersion of a server node in mineral (e.g. transformer) oil, which is commonly used to cool electrical transformers. Mineral oil is less conductive that rubber and has 1,200x the heat capacity of air. It could dramatically increase density and reduce cooling costs.
The Ethernet Alliance will demonstrate the ability to converge LAN, SAN and IPC traffic over a 10 Gigabit Ethernet fabric. The converged Ethernet fabric demonstration will simultaneously highlight multi-vendor interoperability while supporting various application traffic types such as FCoE, DCB, iSCSI, iWARP, LAN, and HPC. The emerging technology should save power, energy and money.
Intel will demonstrate that an alternatve approach to achieving and HPC Datacenter with PUE approaching 1 is possible. It relies on 85C Processors to run in servers at a higher datacenter ambient, and eliminate the need for mechanical refrigeration (CRAC). They also propose 400Vdc power distribution to gain up to another 28% efficiency vs. typical datacenter practices. Intel will also demonstrate a new Universal Interconnect Technology that provides links with Coherence, I/O and Clustering capabilities.
RNA networks will demonstrate how memory virtualization can make memory a shared network resource available to all nodes in the data center. This includes the ability to dynamically construct or provision a shared memory environment within a cluster, based on application requirements. This will enable one to support either a shared or distributed memory model, dynamically, and without sacrificing performance and scale.
AccelerEyes will demonstrate its Jacket technology, which provides developers with a familiar, high-level MATLAB interface for exploiting data parallelism. Jacket automatically translates the high-level code into low-level optimized GPU kernels and efficient memory transfers. The only code modification required for this is the addition of MATLAB keywords. The high degree of programmability will lower-barrier to entry and facilitate a rise in SIMD coding over the next five years.
The University of Illinois, Chicago, will demonstrate the disruptive capability of large data cloud for high performance analytics. To simplify: the problem has been that high performance computers, which have the power to build sophisticated models, do not have access to the data, while databases and archives, which do have access to the data, do not have the power to build sophisticated models. Large data clouds combine a data management function with a high performance computing function. Today, for the first time, a one hundred node commodity cluster running Hadoop can manage and archive a 100 TB dataset and perform high performance analytics as needed.