As far as we know, Micron is the only company that has implemented a technology like this on this scale.
Micron has developed a graphically-based design tool, the AP Workbench, an SDK, and a PCIe board populated with Automata Processor DIMMs that will allow end users to develop and test applications on the AP.
Micron has taped out the first silicon and has prototypes in-house at the Boise facility; samples will be available in 2014.
It liberates computer scientists from the constraints of the Von Neumann architecture, allowing them to implement a level of parallelism previously unattainable. The AP will allow unprecedented computing power to be deployed throughout the industry, which can power innovation in ways we might not have even thought of.
The AP uses a DDR3-like memory interface chosen to simplify the physical design-in process for system integrators. The AP will be made available as single components or as DIMM modules, enhancing the integration process. A PCIe board populated with AP DIMMs will be available to early access application developers to jump-start plug-in development of AP applications.
Micron is working with initial partners and early adopters. We intend to make the SDK available the first part of 2014.
Many of today’s most challenging computer science problems require highly parallel methods to solve. In conventional computing, parallelism involves the same operation being computed on many chunks of data at once and can be cumbersome and complex, often requiring significant effort on the part of programmers and system designers. The Automata Processor’s parallelism exploits the very high and natural parallelism found in Micron’s semiconductor devices—a different kind of parallelism that is more appropriate than conventional CPUs for the class of problems the AP targets. By answering thousands or even millions of different questions about data as it is being streamed across the chip, the AP provides an architecture that delivers the parallelism required to address problems in an efficient, manageable method.
The AP is not a memory device, but it is memory based. Unlike a conventional CPU, the AP is a scalable two-dimensional fabric comprised of thousands of processing elements each programmed to perform a targeted task or operation, ultimately delivering unprecedented performance. Additionally, the AP is massively parallel. Whereas conventional CPU architectures can have anywhere from 2 to 64 processors, an AP can encompass hundreds of thousands or even millions of tiny processors.
Applications with large, unstructured sets of data, or applications that require real time-results, such as cyber security, bioinformatics, big data analytics, and video/image analysis, are examples where the AP could deliver significant value.
The sequential instruction processing nature of conventional CPU/GPU architectures is not well aligned to the class of problems addressed by the AP.
Micron is developing a full suite of tools and materials for developers to design, compile, test, and deploy applications using the AP. Visit micron.com for more information.