公开(公告)号：US20220334835A1

公开(公告)日：2022-10-20

申请号：US17644328

申请日：2021-12-14

Applicant: Intel Corporation

Inventor： Justin Gottschlich ,  Niranjan Hasabnis ,  Paul Petersen ,  Shengtian Zhou ,  Celine Lee

IPC: G06F8/75 ,  G06K9/62 ,  G06N3/08

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed that implement an automatically evolving code recommendation engine. In one example, the apparatus collects a user code snippet. The apparatus then determines a structured representation of the user code snippet. Next, the apparatus generates a recommended code snippet using the structured representation of the user code snippet. Then the apparatus obtains user-determined code snippet feedback comparing the user code snippet to the recommended code snippet, the user-determined code snippet feedback indicating one of a match, no match, or uncertain. Finally, the apparatus stores a code snippet training pair in a training database, the code snippet training pair including the user code snippet and the recommended code snippet.

公开(公告)号：US20220114137A1

公开(公告)日：2022-04-14

申请号：US17559556

申请日：2021-12-22

Applicant: Intel Corporation

Inventor： Celine Lee ,  Niranjan Hasabnis ,  Paul Petersen ,  Justin Gottschlich ,  Ramesh Peri

IPC: G06F15/80 ,  G06N20/00 ,  G06F13/40

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed to generate command lists to be offloaded to accelerator circuitry. An example apparatus includes kernel duration model circuitry to predict a duration of execution of a first kernel based on a first source location, a first name, a first property of a first argument, or an occupancy of the first kernel. The example apparatus includes subsequent kernel model circuitry to predict a tuple and a dependency of a second kernel based on a second source location, a second name, a second property of a second argument, or a time of submission of the previous kernel. The example apparatus includes reinforcement learning model circuitry to determine whether to bundle the first kernel into a command list based on the duration of execution of the first kernel, the tuple of the second kernel, or the dependency of the second kernel.

公开(公告)号：US12001382B2

公开(公告)日：2024-06-04

申请号：US17559556

申请日：2021-12-22

Applicant: Intel Corporation

Inventor： Celine Lee ,  Niranjan Hasabnis ,  Paul Petersen ,  Justin Gottschlich ,  Ramesh Peri

IPC: G06F15/80 ,  G06F3/06 ,  G06F9/455 ,  G06F9/48 ,  G06F9/50 ,  G06F13/40 ,  G06N3/006 ,  G06N3/045 ,  G06N3/08 ,  G06N3/084 ,  G06N5/01 ,  G06N20/00

CPC classification number: G06F15/80 ,  G06F13/4068 ,  G06N20/00

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed to generate command lists to be offloaded to accelerator circuitry. An example apparatus includes kernel duration model circuitry to predict a duration of execution of a first kernel based on a first source location, a first name, a first property of a first argument, or an occupancy of the first kernel. The example apparatus includes subsequent kernel model circuitry to predict a tuple and a dependency of a second kernel based on a second source location, a second name, a second property of a second argument, or a time of submission of the previous kernel. The example apparatus includes reinforcement learning model circuitry to determine whether to bundle the first kernel into a command list based on the duration of execution of the first kernel, the tuple of the second kernel, or the dependency of the second kernel.

公开(公告)号：US11977605B2

公开(公告)日：2024-05-07

申请号：US17644328

申请日：2021-12-14

Applicant: Intel Corporation

Inventor： Justin Gottschlich ,  Niranjan Hasabnis ,  Paul Petersen ,  Shengtian Zhou ,  Celine Lee

IPC: G06F11/36 ,  G06F8/71 ,  G06F8/75 ,  G06F9/451 ,  G06F18/214 ,  G06F18/22 ,  G06F18/2413 ,  G06N3/08 ,  G06F16/9535 ,  G06Q30/0282

CPC classification number: G06F18/22 ,  G06F8/71 ,  G06F8/75 ,  G06F9/453 ,  G06F18/214 ,  G06F18/2155 ,  G06F18/24147 ,  G06N3/08 ,  G06F16/9535 ,  G06Q30/0282

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed that implement an automatically evolving code recommendation engine. In one example, the apparatus collects a user code snippet. The apparatus then determines a structured representation of the user code snippet. Next, the apparatus generates a recommended code snippet using the structured representation of the user code snippet. Then the apparatus obtains user-determined code snippet feedback comparing the user code snippet to the recommended code snippet, the user-determined code snippet feedback indicating one of a match, no match, or uncertain. Finally, the apparatus stores a code snippet training pair in a training database, the code snippet training pair including the user code snippet and the recommended code snippet.