Systems and methods for performing 16-bit floating-point matrix dot product instructions AF Heinecke, R Valentine, MJ Charney, R Sade, M Adelman, Z Sperber, ... US Patent 10,963,246, 2021 | 36 | 2021 |
Faster neural network training with approximate tensor operations M Adelman, K Levy, I Hakimi, M Silberstein Advances in Neural Information Processing Systems 34, 2021 | 33 | 2021 |
Systems, methods, and apparatuses for tile matrix multiplication and accumulation R Valentine, Z Sperber, MJ Charney, BL Toll, R Rappoport, ... US Patent 11,086,623, 2021 | 26 | 2021 |
Systems, methods, and apparatuses for dot product operations R Sade, S Rubanovich, A Gradstein, Z Sperber, A Heinecke, R Valentine, ... US Patent 11,669,326, 2023 | 23 | 2023 |
Systems and methods for performing 16-bit floating-point vector dot product instructions AF Heinecke, R Valentine, MJ Charney, R Sade, M Adelman, Z Sperber, ... US Patent 11,366,663, 2022 | 23 | 2022 |
Tensor processing primitives: a programming abstraction for efficiency and portability in deep learning workloads E Georganas, D Kalamkar, S Avancha, M Adelman, C Anderson, A Breuer, ... Proceedings of the International Conference for High Performance Computing …, 2021 | 22 | 2021 |
Systems and methods to zero a tile register pair R Sade, S Rubanovich, A Gradstein, Z Sperber, A Heinecke, R Valentine, ... US Patent 11,023,235, 2021 | 17 | 2021 |
Efficient and Generic 1D Dilated Convolution Layer for Deep Learning N Chaudhary, S Misra, D Kalamkar, A Heinecke, E Georganas, B Ziv, ... arXiv preprint arXiv:2104.08002, 2021 | 17 | 2021 |
Systems, methods, and apparatuses for matrix operations R Valentine, MJ Charney, E Ould-Ahmed-Vall, D Baum, Z Sperber, ... US Patent App. 16/487,421, 2020 | 16 | 2020 |
Apparatus and method for conjugate transpose and multiply M Adelman, R Valentine, D Towner, A Gradstein, MJ Charney US Patent App. 17/133,456, 2022 | 14 | 2022 |
Systems, methods, and apparatuses for dot production operations R Valentine, D Baum, Z Sperber, J Corbal, E Ould-Ahmed-Vall, BL Toll, ... US Patent 11,163,565, 2021 | 12 | 2021 |
Systems and methods to load a tile register pair R Sade, S Rubanovich, A Gradstein, Z Sperber, A Heinecke, R Valentine, ... US Patent 11,093,247, 2021 | 10 | 2021 |
Systems, methods, and apparatuses for tile transpose R Valentine, D Baum, Z Sperber, J Corbal, E Ould-Ahmed-Vall, BL Toll, ... US Patent App. 16/474,475, 2019 | 10 | 2019 |
Apparatuses, methods, and systems for instructions for moving data between tiles of a matrix operations accelerator and vector registers M Adelman, R Valentine, B Ziv, Y Pollak, G Stupp, A Gradstein, ... US Patent App. 16/914,347, 2021 | 9 | 2021 |
Systems, methods, and apparatuses for matrix operations R Sade, S Rubanovich, A Gradstein, Z Sperber, A Heinecke, R Valentine, ... US Patent 11,816,483, 2023 | 7 | 2023 |
Systems, methods, and apparatuses for tile load R Valentine, M Adelman, MB Girkar, Z Sperber, MJ Charney, BL Toll, ... US Patent 11,567,765, 2023 | 7 | 2023 |
Systems and methods to zero a tile register pair R Sade, S Rubanovich, A Gradstein, Z Sperber, A Heinecke, R Valentine, ... US Patent 11,645,077, 2023 | 6 | 2023 |
Matrix data scatter and gather by row CJ Hughes, AF Heinecke, R Valentine, M Adelman, E Georganas, ... US Patent App. 16/914,321, 2021 | 6 | 2021 |
Loading and storing matrix data with datatype conversion M Adelman, R Valentine, G Stupp, Y Pollak, A Gradstein, S Rubanovich, ... US Patent App. 16/914,317, 2021 | 6 | 2021 |
Systems, methods, and apparatus for tile configuration M Adelman, R Valentine, Z Sperber, MJ Charney, BL Toll, R Rappoport, ... US Patent 11,080,048, 2021 | 6 | 2021 |