公开(公告)号：US20220156639A1

公开(公告)日：2022-05-19

申请号：US17414416

申请日：2019-08-07

Applicant: Hewlett-Packard Development Company, L.P.

Inventor： Vitor Henrique da Silva ,  Thomas da Silva Paula ,  Pedro Henrique Garcez Monteiro

IPC: G06N20/00 ,  G06F9/50

Abstract: Examples of methods for predicting processing workloads are described herein. In some examples, a method may include predicting a processing workload for a set of machine learning models. In some examples, the method may include loading a machine learning model of the set of machine learning models from non-volatile memory based on the predicted processing workload.

公开(公告)号：US20220147430A1

公开(公告)日：2022-05-12

申请号：US17415766

申请日：2019-07-25

Applicant: Hewlett-Packard Development Company, L.P.

Inventor： Carlos Haas Costa ,  Christian Makaya ,  Madhu Sudan Athreya ,  Raphael Gay ,  Pedro Henrique Garcez Monteiro

IPC: G06F11/34 ,  G06N20/00

Abstract: For each of a number of workloads, time intervals within execution performance information that was collected during execution of the workload on a first hardware platform are correlated with corresponding time intervals within execution performance information that was collected during execution of the workload on a second hardware platform. For a workload, the time intervals within the execution performance information on the second hardware platform are correlated to the time intervals within the execution performance information the first hardware platform during which the same parts of the workload were executed. A machine learning model that outputs predicted performance on the second hardware platform relative to known performance on the first hardware platform is trained. The model is trained from the correlated time intervals within the execution performance information for each workload on the hardware platforms.

公开(公告)号：US11294788B2

公开(公告)日：2022-04-05

申请号：US16482358

申请日：2017-07-20

Applicant: HEWLETT-PACKARD DEVELOPMENT COMPANY, L.P.

Inventor： Taciano Perez ,  Tadeu Marchese ,  Pedro Henrique Garcez Monteiro ,  Raphael Gay

IPC: G06F11/34 ,  G06F16/901 ,  G06F11/30 ,  G06F11/32

Abstract: A program is executed on a first computer system and the execution of the program is monitored. A plurality of operation records are created based upon the monitoring, where each operation record is associated with an operation carried out during execution of the program. A first value of a cumulative performance indicator associated with the execution of the program on the first computer system is determined. For each operation record, a value of a performance indicator associated with carrying out the operation on a second computer system is predicted. For an operation record, the value of the performance indicator is predicted based on a performance model associated with carrying out operations on the second computer system. A second value of the cumulative performance indicator is determined, which is associated with execution of the program on the second computer system and is based on the predicted values of the performance indicator.