Company Filing History:
Years Active: 2016
Title: Innovations by Charles Tsen in Fused Multiply Add Operations
Introduction
Charles Tsen is an accomplished inventor based in Portland, OR (US). He has made significant contributions to the field of computing, particularly in the area of floating-point arithmetic. His innovative approach has led to the development of a unique method for correcting errors in fused multiply add operations.
Latest Patents
Charles Tsen holds a patent for an invention titled "Efficient correction of normalizer shift amount errors in fused multiply add operations." This method addresses the challenge of correcting shift errors in fused multiply add operations. The process involves adjusting a normalized floating-point number prior to performing a shift error correction. This results in an adjusted normalized floating-point number, which is then corrected by shifting the mantissa in one direction. The invention includes a fused multiply add module that consists of a normalizer module, compensation logic, and a rounder. The normalizer module is designed to normalize a floating-point number based on an estimated quantity of leading zeros, while the compensation logic manages the correction of shift errors.
Career Highlights
Charles Tsen is currently employed at Nvidia Corporation, a leading technology company known for its advancements in graphics processing and computing technologies. His work at Nvidia has allowed him to apply his expertise in floating-point arithmetic to enhance the performance of computing systems.
Collaborations
One of his notable coworkers is Adam Dreyer, with whom he has collaborated on various projects within the company. Their combined efforts contribute to the innovative environment at Nvidia Corporation.
Conclusion
Charles Tsen's contributions to the field of computing, particularly through his patent on correcting shift errors in fused multiply add operations, highlight his role as a significant inventor. His work continues to influence advancements in technology and computing efficiency.