Flavors Technology Incorporated |
Slide 3 of 24
Briefly, the Parallel Inference Machine, or PIM, is a VME-based parallel processing platform. The PIM is scalable from hundred to many thousands of processing cells, all running in parallel. The PIM features a shared memory available to all processing cells, and each processing cell has its own local memory. The PIM runs at a fixed rate determined by the user according to the requirements of the application. Further, the PIM runs "systolically" in a read-execute-write cycle that is repeated at a user-determined rate.
Paracell is the programming language of the PIM. The language is English-like, understandable by both application and domain engineers. Agents are developed in simple "if-then-else" statements.
The Paracell Navigator is the programming environment for writing Paracell code. The Navigator is a graphical environment, providing a 3x3 matrix that extends ten levels deep so that applications can be easily broken down and are manageable when completed. The Navigator includes tools for debugging, tracking variables, editing configurations, developing MMIs, et cetera, and handles all other interactions with the software and hardware.