NumaConnect Adapter cards deliver true plug-and-play connectivity between standard servers running a single pristine OS image. The hardware-controlled ccNuma environment allows applications to have fast access to the aggregate memory of all servers in the system. NumaChip is fabricated by IBM and was seen operating with IBM x3755 at SC12 in Salt Lake City, Utah.

NumaConnect enables very large shared memory systems built from commodity servers at the price point of high-end clusters. This combination sets a new level of price/capability/performance that has never before been demonstrated with x86 processors. Systems built with NumaConnect run standard operating systems and all x86 applications.

NumaConnect provides a record-breaking main memory capacity of up to 256TBytes in a single commodity-based system. “Customers with Big Data problems are excited by the ability to directly address any record anywhere in their entire data set in a microsecond or less,” says Einar Rustad, co-founder and VP of Numascale.

“This is orders of magnitude faster than clusters or systems based on any form of existing mass-storage devices and will enable data analysis and decision support applications to be applied in new and innovative ways,” says Kåre Løchsen, Numascale founder and CEO. He adds, “It is also worth mentioning that pilot customers have been surprised by the ease of use through statements like: ‘Using Numaconnect shared memory clusters is just like running programs on my desktop, just much more powerful!’”

The technology is being tested for oil exploration applications by Statoil and for general HPC datacenter use by the University of Oslo. Some of the main applications planned in Oslo include bioscience and computational chemistry. The EU project PRACE has financed a 72-node cluster of IBM x3755s for proving the technology.

NumaConnect works with AMD Opteron-based servers connecting to HyperTransport and directly addressable, shared memory can be up to 256TBytes.

NumaConnect provides system-wide cache coherency logic with a directory-based protocol that scales to 4096 nodes where each node can have up to three multicore processors. The cache coherency logic is implemented in an ASIC together with interconnect fabric circuitry with routing tables for multi-dimensional Torus topologies. This type of fabric is very scalable and the same topology is used in the world’s largest supercomputers.

NumaConnect starts shipping in December for use with servers from IBM, Supermicro, and AIC.

About Numascale

Numascale has offices in Europe, Asia, and USA, and is supported by: Statoil, ProVenture, Investinor, Innovation Norway, Norges forskningsråd, and Eurostars.

NumaConnect allows high-volume manufactured server boards to be used as building blocks for systems with features that are only found in the high-end enterprise servers. NumaConnect includes full support for virtualization of processing, memory, and I/O resources, and can be used with standard operating systems.