Maximum performance demands a new paradigm

As microprocessors reach the limits of attainable clock frequencies and acceptable power consumption, Intel and AMD are scaling up the number of cores per chip, thereby enabling higher degrees of parallelism. To take advantage of potential and only-modest speed improvements, existing software must be modified. This slow down in growth of computer power presents a challenge to the business community.

Alternative computing technology, and dataflow computing in particular, are emerging as a powerful solution to overcome the slowdown of scaling with traditional processors.

Maxeler Technologies' core competence is in delivering substantially increased performance for HPC applications through the use of dataflow technology. Dataflow compute engines implement applications as dataflow graphs directly in hardware, compiling from a high-level application program description written in Java.

By mapping compute-intensive algorithms directly into parallel dataflow pipelines, tightly coupled to a conventional CPU through a high-speed I/O bus, complete applications can be accelerated by orders of magnitude over conventional CPU implementations. By exploiting massive parallelism at the bit-level, Maxeler supercomputing solutions deliver performance far in excess of CPUs at approximately a tenth of the clock frequency. Thus, our solutions offer substantial improvements in cost/performance and power/performance ratios.