Issued Patents 2021
Showing 1–9 of 9 patents
| Patent # | Title | Co-Inventors | Date |
|---|---|---|---|
| 11210584 | Memory efficient convolution operations in deep learning neural networks | Daniel Brand | 2021-12-28 |
| 11182457 | Matrix-factorization based gradient compression | Vinod Muthusamy | 2021-11-23 |
| 11093438 | Pipelining multi-directional reduction | Ulrich A. Finkler, David S. Kung | 2021-08-17 |
| 11093827 | Variable ISA vector-based compaction in distributed training of neural networks | Ulrich A. Finkler | 2021-08-17 |
| 11036829 | System, method, and recording medium for mirroring matrices for batched Cholesky decomposition on a graphic processing unit | David S. Kung, Ruchir Puri | 2021-06-15 |
| 10979531 | Pessimistic scheduling for topology optimized workload placement | Seetharami R. Seelam | 2021-04-13 |
| 10977552 | ISA-based compression in distributed training of neural networks | Ulrich A. Finkler | 2021-04-13 |
| 10956357 | Method for flexible, fast all-reduce on arbitrary tree topology | Jinho Lee, Soham Pranav Shah | 2021-03-23 |
| 10922606 | Multi-directional reduction in large scale deep-learning | Ulrich A. Finkler, David S. Kung, Li Zhang | 2021-02-16 |