公开(公告)号：US20230186149A1

公开(公告)日：2023-06-15

申请号：US17550882

申请日：2021-12-14

Applicant: Advanced Micro Devices, Inc.

Inventor： Saketh Venkata Rama ,  Ganesh Dasika ,  Laurent S. White

IPC: G06N20/00

CPC classification number: G06N20/00

Abstract: An approach is provided for using machine learning to provide compensation for roundoff error in algorithmic computations. The approach includes training a machine learning model based low precision data and corresponding high precision data. The low precision data includes pairs of low precision values of a specific datatype that correspond to pairs of high precision values from the high precision data. The high precision data includes pairs of high precision values of a specific datatype that correspond to the pairs of low precision values from the low precision data. When the machine learning model has been trained, the machine learning model is used as a basis for determining a compensation value is used to compensate for roundoff error in a particular algorithmic computation. Techniques discussed herein provide compensation for roundoff error during otherwise unstable computations, enabling high-performance computing and other scientific applications to use lower precision data types more readily.

公开(公告)号：US20230205837A1

公开(公告)日：2023-06-29

申请号：US17561227

申请日：2021-12-23

Applicant: Advanced Micro Devices, Inc.

Inventor： Laurent S. White ,  Ganesh Dasika ,  Saketh Venkata Rama

IPC: G06F17/13 ,  G06N20/00 ,  G01R33/24

CPC classification number: G06F17/13 ,  G06N20/00 ,  G01R33/24

Abstract: A physical system is simulated using a model including a plurality of elements in a mesh or grid. The elements are divided into partitions processed by different processing units. For some time steps, flux data is transmitted between partitions for updating the state of edge elements of the partitions. Periodically, transmission of flux data is suppressed and flux data is obtained by linear interpolation based on past flux data. Alternatively, flux data is obtained by processing state variables of an edge element and past flux data using a machine learning model, such as a DNN.