中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

123 records (took 0.038 seconds)

    Tuning of loop orders in blocked dense basic linear algebra subroutines
    21.
    发明授权
    Tuning of loop orders in blocked dense basic linear algebra subroutines 有权

    公开(公告)号：US11354564B2

    公开(公告)日：2022-06-07

    申请号：US16454318

    申请日：2019-06-27

    Applicant: Intel Corporation

    Inventor： Alexander Heinecke ,  Evangelos Georganas ,  Justin Gottschlich

    IPC: G06N3/04 ,  G06F9/30 ,  G06N3/08 ,  G06F8/41 ,  G06F9/455

    Abstract: An example includes a sequence generator to generate a plurality of sequence pairs, a first one of the sequence pairs including: (i) a first input sequence representing first accesses to first tensors in a first loop nest of a first computer program, and (ii) a first output sequence representing a first tuned loop nest corresponding to the first accesses to the first tensors in the first loop nest; a model trainer to train a recurrent neural network based on the sequence pairs as training data, the recurrent neural network to be trained to tune loop ordering of a second computer program based on a second input sequence representing second accesses to a second tensor in a second loop nest of the second computer program; and a memory interface to store, in memory, a trained model corresponding to the recurrent neural network.

    Methods and apparatus to recommend instruction adaptations to improve compute performance
    23.
    发明授权
    Methods and apparatus to recommend instruction adaptations to improve compute performance 有权

    公开(公告)号：US11340874B2

    公开(公告)日：2022-05-24

    申请号：US16455259

    申请日：2019-06-27

    Applicant: Intel Corporation

    Inventor： Thijs Metsch ,  Mohammad Mejbah Ul Alam ,  Justin Gottschlich

    IPC: G06F8/35 ,  G06N20/00

    Abstract: Methods, apparatus, systems and articles of manufacture to recommend instruction adaptations to improve compute performance are disclosed. An example apparatus includes a pattern detector to detect an execution pattern from an execution profile provided by a server, the execution profile associated with an instruction stored in an instruction repository. An adaptation identifier is to identify a possible instruction adaptation that may be applied to the instruction associated with the execution pattern. A model processor is to predict, using a machine learning model, an expected performance improvement of the adaptation. A result comparator is to determine whether the expected performance improvement meets an threshold. An instruction editor is to, in response to the result comparator determining that the expected performance improvement meets the threshold, apply the possible instruction adaptation to the instruction in the instruction repository.

    METHODS AND APPARATUS TO DETERMINE EXECUTION COST
    25.
    发明申请
    METHODS AND APPARATUS TO DETERMINE EXECUTION COST 有权

    公开(公告)号：US20220091895A1

    公开(公告)日：2022-03-24

    申请号：US17541016

    申请日：2021-12-02

    Applicant: Intel Corporation

    Inventor： Niranjan Hasabnis ,  Justin Gottschlich ,  Jesmin Jahan Tithi ,  Anand Venkat

    IPC: G06F9/50 ,  G06F9/30 ,  G06F9/455 ,  G06F9/38

    Abstract: Methods, apparatus, systems, and articles of manufacture to determine execution cost are disclosed. An example apparatus includes memory; instructions included in the apparatus; and processor circuitry to execute the instruction to: cause a plurality of instructions corresponding to a mnemonic to be executed; determine an average execution cost of the plurality of instructions; determine a standard deviation of execution costs of the plurality of instructions; and generate a mapping table including an entry, the entry including the mnemonic in association with the average and the standard deviation.

    Methods and apparatus for intentional programming for heterogeneous systems
    26.
    发明授权
    Methods and apparatus for intentional programming for heterogeneous systems 有权

    公开(公告)号：US11269639B2

    公开(公告)日：2022-03-08

    申请号：US16455388

    申请日：2019-06-27

    Applicant: Intel Corporation

    Inventor： Adam Herr ,  Derek Gerstmann ,  Justin Gottschlich ,  Mikael Bourges-Sevenier ,  Sridhar Sharma

    IPC: G06F9/30 ,  G06F8/30 ,  G06F8/52 ,  G06F8/76 ,  G06F8/75 ,  G06N3/04 ,  G06F9/38

    Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for intentional programming for heterogeneous systems. An example apparatus includes a code lifter to identify annotated code corresponding to an algorithm to be executed on the heterogeneous system based on an identifier being associated with the annotated code, and convert the annotated code in the first representation to intermediate code in a second representation by identifying the intermediate code as having a first algorithmic intent that corresponds to a second algorithmic intent of the annotated code, a domain specific language (DSL) generator to translate the intermediate code in the second representation to DSL code in a third representation when the first algorithmic intent matches the second algorithmic intent, the third representation corresponding to a DSL representation, and a code replacer to invoke a compiler to generate an executable including variant binaries based on the DSL code.

    Apparatus and methods for object manipulation via action sequence optimization
    28.
    发明授权
    Apparatus and methods for object manipulation via action sequence optimization 有权

    公开(公告)号：US11213947B2

    公开(公告)日：2022-01-04

    申请号：US16455263

    申请日：2019-06-27

    Applicant: Intel Corporation

    Inventor： Javier Felip Leon ,  David Israel Gonzalez Aguirre ,  Javier Sebastián Turek ,  Ignacio Javier Alvarez ,  Luis Carlos Maria Remis ,  Justin Gottschlich

    IPC: B25J9/16

    Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for object manipulation via action sequence optimization. An example method disclosed herein includes determining an initial state of a scene, generating a first action phase sequence to transform the initial state of the scene to a solution state of the scene by selecting a plurality of action phases based on action phase probabilities, determining whether a first simulated outcome of executing the first action phase sequence satisfies an acceptability criterion and, when the first simulated outcome does not satisfy the acceptability criterion, calculating a first cost function output based on a difference between the first simulated outcome and the solution state of the scene, the first cost function output utilized to generate updated action phase probabilities.

Patent Agency Ranking