PROCESSOR FOR LARGE GRAPH ALGORITHM COMPUTATIONS AND MATRIX OPERATIONS

Patent №

US 8,751,556

Granted

2014-06-10

Filed 2011

Owner

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

AI components

2

evo · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

13153490

A multiprocessor system and method for performing matrix operations includes multiple processors cooperatively performing a sparse matrix operation. Distributed among the processors are non-zero matrix elements of first and second sparse matrices. Mapped across the processors are the matrix elements of a results matrix. Each processor receives, from the other processors, non-zero matrix elements of the first matrix that had been distributed to those other processors and generates partial results based on the received non-zero matrix elements of the first matrix and on the non-zero matrix elements of the second matrix distributed to that processor. Each processor receives those partial results generated by other processors and associated with the matrix elements of the results matrix mapped to that processor. Each processor generates a final value for each matrix element of the results matrix mapped to that processor based on the partial results generated by that processor and on the partial results received from the other processors associated with that matrix element of the results matrix.

AI classification

AI hardware1.00
Evolutionary computation0.71
Machine learning0.47
Knowledge representation0.40
Natural language0.00
Vision0.00
Speech0.00
Planning0.00

Ownership

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

assignment · 263990421

Assignors

SONG, WILLIAM S.

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC