-
公开(公告)号：US11599367B2
公开(公告)日：2023-03-07
申请号：US16752239
申请日：2020-01-24
申请人： CORNAMI, INC.
发明人： Tianfang Liu
摘要： A system and method to compress application control data, such as weights for a layer of a convolutional neural network, is disclosed. A multi-core system for executing at least one layer of the convolutional neural network includes a storage device storing a compressed weight matrix of a set of weights of the at least one layer of the convolutional network and a decompression matrix. The compressed weight matrix is formed by matrix factorization and quantization of a floating point value of each weight to a floating point format. A decompression module is operable to obtain an approximation of the weight values by decompressing the compressed weight matrix through the decompression matrix. A plurality of cores executes the at least one layer of the convolutional neural network with the approximation of weight values to produce an inference output.
-
公开(公告)号：US11693662B2
公开(公告)日：2023-07-04
申请号：US16743257
申请日：2020-01-15
申请人： Cornami Inc.
CPC分类号： G06F9/3812 , G06F7/4806 , G06F8/4436 , G06F17/142
摘要： Systems and methods for configuring a reduced instruction set computer processor architecture to execute fully homomorphic encryption (FHE) logic gates as a streaming topology. The method includes parsing sequential FHE logic gate code, transforming the FHE logic gate code into a set of code modules that each have in input and an output that is a function of the input and which do not pass control to other functions, creating a node wrapper around each code module, configuring at least one of the primary processing cores to implement the logic element equivalents of each element in a manner which operates in a streaming mode wherein data streams out of corresponding arithmetic logic units into the main memory and other ones of the plurality arithmetic logic units.
-
公开(公告)号：US20200213079A1
公开(公告)日：2020-07-02
申请号：US16743257
申请日：2020-01-15
申请人： Cornami Inc.
发明人： Mache Kreeger , Tianfang Liu , Frederick Furtek , Paul L. Master
摘要： Systems and methods for configuring a reduced instruction set computer processor architecture to execute fully homomorphic encryption (FHE) logic gates as a streaming topology. The method includes parsing sequential FHE logic gate code, transforming the FHE logic gate code into a set of code modules that each have in input and an output that is a function of the input and which do not pass control to other functions, creating a node wrapper around each code module, configuring at least one of the primary processing cores to implement the logic element equivalents of each element in a manner which operates in a streaming mode wherein data streams out of corresponding arithmetic logic units into the main memory and other ones of the plurality arithmetic logic units.
-
公开(公告)号：US12126367B2
公开(公告)日：2024-10-22
申请号：US18178237
申请日：2023-03-03
申请人： Cornami, Inc.
发明人： Tianfang Liu
IPC分类号： H03M7/30 , G06F7/483 , G06F9/445 , G06F18/2135 , G06F18/24 , G06N3/02 , G06N3/0464 , G06V10/764 , G06V10/77 , G06V10/82 , H03M7/02 , H03M7/40
CPC分类号： H03M7/3059 , G06F7/483 , G06F9/445 , G06F18/2135 , G06F18/24 , G06N3/02 , G06N3/0464 , G06V10/764 , G06V10/7715 , G06V10/82 , H03M7/02 , H03M7/40 , H03M7/6005 , H03M7/6011 , H03M7/70
摘要： A system and method to compress application control data, such as weights for a layer of a convolutional neural network, is disclosed. A multi-core system for executing at least one layer of the convolutional neural network includes a storage device storing a compressed weight matrix of a set of weights of the at least one layer of the convolutional network and a decompression matrix. The compressed weight matrix is formed by matrix factorization and quantization of a floating point value of each weight to a floating point format. A decompression module is operable to obtain an approximation of the weight values by decompressing the compressed weight matrix through the decompression matrix. A plurality of cores executes the at least one layer of the convolutional neural network with the approximation of weight values to produce an inference output.
-
公开(公告)号：US20230350684A1
公开(公告)日：2023-11-02
申请号：US18205929
申请日：2023-06-05
申请人： Cornami Inc.
CPC分类号： G06F9/3812 , G06F7/4806 , G06F17/142 , G06F8/4436
摘要： Systems and methods for configuring a reduced instruction set computer processor architecture to execute fully homomorphic encryption (FHE) logic gates as a streaming topology. The method includes parsing sequential FHE logic gate code, transforming the FHE logic gate code into a set of code modules that each have in input and an output that is a function of the input and which do not pass control to other functions, creating a node wrapper around each code module, configuring at least one of the primary processing cores to implement the logic element equivalents of each element in a manner which operates in a streaming mode wherein data streams out of corresponding arithmetic logic units into the main memory and other ones of the plurality arithmetic logic units.
-
公开(公告)号：US20220360428A1
公开(公告)日：2022-11-10
申请号：US17860475
申请日：2022-07-08
申请人： Cornami, Inc.
摘要： Systems and methods for configuring a reduced instruction set computer processor architecture to execute fully homomorphic encryption (FHE) logic gates as a streaming topology. The method includes parsing sequential FHE logic gate code, transforming the FHE logic gate code into a set of code modules that each have in input and an output that is a function of the input and which do not pass control to other functions, creating a node wrapper around each code module, configuring at least one of the primary processing cores to implement the logic element equivalents of each element in a manner which operates in a streaming mode wherein data streams out of corresponding arithmetic logic units into the main memory and other ones of the plurality arithmetic logic units.
-
-
-
-
-
搜索结果
6 条记录 (耗时 0.014 秒)
