Inventor · Rancho Cordova, CA, US

Kamal Sinha

122Patents
10h-index
101Co-inventors
75Inventor score

Filing activity: Apr 1, 2017 → Mar 11, 2024

Most-cited inventions

PatentTitleAreaCited byStatus
US10474458B2 Instructions and logic to perform floating-point and integer operations for machine learning Physics 40 Active
US10353706B2 Instructions and logic to perform floating-point and integer operations for machine learning Physics 40 Active
US11080046B2 Instructions and logic to perform floating point and integer operations for machine learning Physics 33 Active
US11360767B2 Instructions and logic to perform floating point and integer operations for machine learning Physics 33 Active
US11620256B2 Systems and methods for improving cache efficiency and utilization Physics 32 Active
US11169799B2 Instructions and logic to perform floating-point and integer operations for machine learning Physics 31 Active
US10304154B2 Coordination and increased utilization of graphics processors during inference Emerging Cross-Sectional Technologies 29 Active
US10186011B2 Programmable coarse grained and sparse matrix compute hardware with advanced scheduling Physics 29 Active
US10346944B2 Machine learning sparse computation mechanism Emerging Cross-Sectional Technologies 17 Active
US10706498B2 Machine learning sparse computation mechanism Emerging Cross-Sectional Technologies 11 Active
US10401954B2 Sensory enhanced augmented reality and virtual reality device Physics 10 Active
US11210760B2 Programmable coarse grained and sparse matrix compute hardware with advanced scheduling Physics 7 Active
US10102149B1 Replacement policies for a hybrid hierarchical cache Physics 7 Active
US10417731B2 Compute optimization mechanism for deep neural networks Emerging Cross-Sectional Technologies 6 Active
US10423415B2 Hierarchical general register file (GRF) for execution block Emerging Cross-Sectional Technologies 6 Active
US10444817B2 System, apparatus and method for increasing performance in a processor during a voltage ramp Physics 6 Active
US10410098B2 Compute optimizations for neural networks Physics 5 Active
US10261903B2 Extend GPU/CPU coherency to multi-GPU cores Emerging Cross-Sectional Technologies 5 Active
US10452397B2 Efficient multi-context thread distribution Physics 5 Active
US10983594B2 Sensory enhanced augmented reality and virtual reality device Physics 5 Active
US11360808B2 Efficient thread group scheduling Physics 4 Active
US11145105B2 Multi-tile graphics processor rendering Physics 4 Active
US10346166B2 Intelligent thread dispatch and vectorization of atomic operations Emerging Cross-Sectional Technologies 4 Active
US10909039B2 Data prefetching for graphics data processing Emerging Cross-Sectional Technologies 3 Active
US11074072B2 Compute optimizations for neural networks using bipolar binary weight Physics 3 Active

Source: USPTO / EPO open patent data. Inventor disambiguation is heuristic; counts are objective bibliographic measures.