公开(公告)号：US20240345936A1

公开(公告)日：2024-10-17

申请号：US18632236

申请日：2024-04-10

申请人： SambaNova Systems, Inc.

发明人： Muthiah ANNAMALAI ,  Anders RAVNBORG

IPC分类号： G06F11/34 ,  G06F11/36

CPC分类号： G06F11/3452 ,  G06F11/3428 ,  G06F11/3616

摘要： A system comprising a tool for providing actionable insight for bring up and performance debug of performant dataflow graphs on CGRA. A system comprising a tool for providing hierarchical traceable graph transformation of dataflow graph and annotated with runtime information after the compilation and execution back onto higher levels of stack from hardware metrics. A system comprising a tool for system performance monitoring and tuning by composition of compile time and runtime information of a workload dataflow graph on CGRA.

公开(公告)号：US12026078B1

公开(公告)日：2024-07-02

申请号：US18346082

申请日：2023-06-30

申请人： NOTA, INC.

发明人： Sanggeon Park ,  Jimin Lee

IPC分类号： G06F11/34

CPC分类号： G06F11/3428 ,  G06F11/3423

摘要： Disclosed is a method for providing a benchmark result, which is performed by a computing device. The method may include obtaining a first input data including an inference task and a dataset. The method may include determining a target model which is a subject of a benchmark for the inference task and at least one target node at which the inference task of the target model is to be executed. The determined target model corresponds to an artificial intelligence-based model, and the benchmark for the inference task of the determined target model is performed at the at least one target node based on the dataset. The method may include providing the benchmark result obtained by executing the target model at the at least one target node.

公开(公告)号：US20240211370A1

公开(公告)日：2024-06-27

申请号：US18145576

申请日：2022-12-22

申请人： Optum Services (Ireland) Limited

发明人： Riccardo MATTIVI

IPC分类号： G06F11/34 ,  G06N20/00

CPC分类号： G06F11/3428 ,  G06N20/00

摘要： Various embodiments of the present disclosure disclose model training processes facilitated by specialized interactive user interfaces. A user interface may provide a natural language rule described by natural language text to a user for review. The user may verify or modify the natural language rule to establish a performance condition for a machine learning model. The natural language rule may be converted to a computer interpretable rule that may be used as a labeling function for the performance condition. Using weak supervision techniques, the labeling function may be applied to a training dataset to generate a labeled training dataset. A machine learning model may be trained using the labeled training dataset. Once trained, the model may be evaluated and evaluation data may be provided to the user. The user may perform multiple iterations of the training process to refine, add, or delete natural language rules for training the model.

公开(公告)号：US12020077B1

公开(公告)日：2024-06-25

申请号：US17246459

申请日：2021-04-30

申请人： SPLUNK INC.

发明人： Sumeet Rohatgi ,  Gaurav Chawla ,  Mehul Goyal ,  Tyler Bakke ,  Aditya Kamble ,  Hao Zhan ,  Rahul Merulinath ,  Ryan Flynn ,  Matteo Merli ,  Boyang Peng ,  Sanjeev Kulkarni ,  Ivan Kelly ,  Karthikeyan Ramasamy

IPC分类号： G06F9/46 ,  G06F9/50 ,  G06F11/14 ,  G06F11/20 ,  G06F11/34

CPC分类号： G06F9/505 ,  G06F9/5022 ,  G06F9/5038 ,  G06F11/1492 ,  G06F11/203 ,  G06F11/3428

摘要： A process for facilitating autoscaling in a stateful system is described herein. In embodiments, a set of metrics associated with a set of components of a stateful service is obtained. The set of metrics may generally indicate a utilization or a load of each of the components of the set of components (e.g., message managers and/or data stores). Thereafter, it is determined to initiate a scaling event at the stateful service in association with the set of components of the stateful service based on at least a portion of the set of metrics attaining a metric threshold indicating a threshold value for determining whether to scale stateful service components. A scaling request can then be provided to the stateful service to initiate the scaling event at the stateful service in association with the set of components of the stateful service.

公开(公告)号：US12013824B1

公开(公告)日：2024-06-18

申请号：US17161334

申请日：2021-01-28

申请人： Liquibase Inc.

发明人： Robert E. Reeves ,  Peter James Pickerill

IPC分类号： G06F16/21 ,  G06F11/34 ,  G06F16/23 ,  G06F16/245 ,  G06F16/2453 ,  G06N5/025 ,  G06N20/00

CPC分类号： G06F16/213 ,  G06F11/3428 ,  G06F16/2358 ,  G06F16/24542 ,  G06N5/025 ,  G06N20/00

摘要： Embodiments provide systems, methods, and computer program products that utilize artificial intelligence/machine learning to process database change data and correlated performance data to predict the impact of database changes and generate rules with respect to database changes to prevent undesired behavior or promote increased performance.

公开(公告)号：US12004013B2

公开(公告)日：2024-06-04

申请号：US17323680

申请日：2021-05-18

申请人： Microsoft Technology Licensing, LLC

发明人： Bozidar Radunovic ,  Sanjeev Mehrotra ,  Yongguang Zhang ,  Paramvir Bahl ,  Xenofon Foukas

IPC分类号： H04W28/16 ,  G06F11/32 ,  G06F11/34 ,  G06N20/00 ,  H04W24/10 ,  H04W28/02 ,  H04W72/21 ,  H04W72/23

CPC分类号： H04W28/16 ,  G06F11/324 ,  G06F11/3428 ,  G06N20/00 ,  H04W24/10 ,  H04W28/0268 ,  H04W72/21 ,  H04W72/23

摘要： Described are examples for monitoring performance metrics of one or more workloads in a cloud-computing environment and reallocating compute resources based on the monitoring. Reallocating compute resources can include migrating workloads among nodes or other resources in the cloud-computing environment, reallocating hardware accelerator resources, adjusting transmit power for virtual radio access network (vRAN) workloads, and/or the like.

公开(公告)号：US20240134769A1

公开(公告)日：2024-04-25

申请号：US18485811

申请日：2023-10-12

申请人： Arizona Board of Regents on Behalf of Arizona State University ,  University of California

发明人： Shail Dave ,  Aviral Shrivastava ,  Tony Nowatzki

IPC分类号： G06F11/34 ,  G06F8/41

CPC分类号： G06F11/3428 ,  G06F8/4441

摘要： A system applies bottleneck analysis for design and optimization of computing systems. In particular, the system constructs a bottleneck model, including a bottleneck cost graph for a workload or a function, through which factors corresponding to the execution costs of an arbitrary processor can be modeled. By using the bottleneck analysis, the system can determine bottleneck factors for an obtained cost value (e.g., time taken by an application's execution on a processor) and can reason about obtained high cost. The system determines and uses information about parameters impacting bottlenecks for execution costs and their approximate relationship with the bottlenecks to produce an optimized hardware-software configuration for execution of one or more workloads. Systematic, bottleneck-guided analysis and optimization can introduce explainability in the design and optimization process and can achieve more efficient design configurations much faster.

公开(公告)号：US20240126672A1

公开(公告)日：2024-04-18

申请号：US17980394

申请日：2022-11-03

申请人： Dell Products L.P.

发明人： Hongwei YUE ,  Shunhua XIE

IPC分类号： G06F11/34 ,  G06F11/30 ,  G06K9/62

CPC分类号： G06F11/3428 ,  G06F11/3006 ,  G06K9/6256

摘要： An information handling system may include at least one processor and a memory. The information handling system may be configured to: receive telemetry information regarding a target workload; receive configuration data regarding a computing cluster that is to execute a simulation of the target workload; train a workload artificial intelligence (AI) model based on the telemetry information and the configuration data to create the simulation of the target workload; generate a benchmarking configuration file based on the workload AI model; and deploy the benchmarking configuration file to the computing cluster for execution.

公开(公告)号：US20240069998A1

公开(公告)日：2024-02-29

申请号：US18224513

申请日：2023-07-20

申请人： NVIDIA Corporation

发明人： Sanjay Chatterjee ,  Dileep Ranganathan ,  Sukesh Roy ,  Nivedita Viswanath ,  Neeraj Kapoor

IPC分类号： G06F11/07 ,  G06F9/50 ,  G06F11/34

CPC分类号： G06F11/0709 ,  G06F9/5038 ,  G06F11/3428 ,  G06F2209/5022

摘要： Apparatuses, systems, and techniques for selecting computing resources based on software programs scoring past performance of computing resources. In at least one embodiment, a processor comprising circuitry may cause software programs to be performed using computing resources based on software programs to score past performance of the one or more computing resources. In at least one embodiment, a processor selects a computing system to perform a software workload based on attributes of computer systems that are recorded over periods of time. In at least one embodiment, a processor comprises circuits to use attribute values of computer systems that are computed over periods of time to select computer systems to perform software workloads.

公开(公告)号：US11874810B2

公开(公告)日：2024-01-16

申请号：US17189944

申请日：2021-03-02

申请人： Saudi Arabian Oil Company

发明人： Mohammed Jebreel Hakami ,  Ning Li ,  Paul Lecaillon

IPC分类号： G06F16/20 ,  G06F16/215 ,  G06F16/28 ,  G06F11/34 ,  G06F11/30 ,  G06F16/2455

CPC分类号： G06F16/215 ,  G06F11/3072 ,  G06F11/3428 ,  G06F16/24564 ,  G06F16/287

摘要： Some implementations of the present disclosure provide a method that include: accessing a plurality of tables that describe a relational data model for a hierarch of exploration data assets, wherein the relational data model defines the hierarchy of exploration data assets and attributes at each level of the hierarchy, and wherein multiple databases encode the hierarchy of exploration data assets; querying the hierarch of exploration data assets according to one or more data quality rules capable of identifying defects in the hierarchy of exploration data assets; identifying instances of data records that fail to meet the one or more data quality rules; based on the identified instances, calculating one or more data quality metrics for the hierarchy of exploration data assets; and generating an alert that includes: the identified instances of data records, or the calculated one or more data quality metrics.