This technology is an error correcting code algorithm that identifies errors in data with a simplified approach, conserving resources without compromising performance.
Current error correcting codes (ECCs) consume substantial amounts of memory in the form of redundant information known as metadata. While increased metadata can provide improved performance, the tradeoff between productivity and computational resource consumption remains a critical key attribute of ECCs. ECCs that can maintain similar performance while decreasing demand would prove highly advantageous for multiple applications including memory systems, commercial servers, data encryption and security, and supercomputers.
This error correcting code known as MUSE relies on simple multiplication and division operations to encode, decode, and detect both single and multi-bit errors. By implementing shuffling and aliasing to allow for continuous optimization of its algorithm, MUSE offers a high performance but resource-efficient solution to error corrections in various memory and data storage/transmission systems. Having been demonstrated to achieve 100% successful single error correction with multi-bit correction rates near 80%, MUSE has significant extra memory states, negligible performance overhead, and reasonable area overhead when compared to traditional ECCs. As a result, MUSE can be used for a wide range of electronic applications including memory systems, commercial servers, and data security.
Lakshminarasimhan Sethumadhavan, Ph.D.
IR CU21290
Licensing Contact: Greg Maskel