公开(公告)号：US20230229591A1

公开(公告)日：2023-07-20

申请号：US18126228

申请日：2023-03-24

Applicant: Meta Platforms Technologies, LLC

Inventor： Ganesh Venkatesh ,  Liangzhen Lai ,  Pierce I-Jen Chuang ,  Meng Li

IPC: G06F12/04

CPC classification number: G06F12/04 ,  G06F2212/1028 ,  G06N3/08

Abstract: Disclosed herein includes a system, a method, and a device for reading and writing sparse data in a neural network accelerator. A mask identifying byte positions within a data word having non-zero values in memory can be accessed. Each bit of the mask can have a first value or a second value, the first value indicating that a byte of the data word corresponds to a non-zero byte value, the second value indicating that the byte of the data word corresponds to a zero byte value. The data word can be modified to have non-zero byte values stored at an end of a first side of the data word in the memory, and any zero byte values stored in a remainder of the data word. The modified data word can be written to the memory via at least a first slice of a plurality of slices that is configured to access the first side of the data word in the memory.

公开(公告)号：US11630770B2

公开(公告)日：2023-04-18

申请号：US16509138

申请日：2019-07-11

Applicant: Meta Platforms Technologies, LLC

Inventor： Ganesh Venkatesh ,  Liangzhen Lai ,  Pierce I-Jen Chuang ,  Meng Li

IPC: G06F12/04 ,  G06N3/08

Abstract: Disclosed herein includes a system, a method, and a device for reading and writing sparse data in a neural network accelerator. A plurality of slices can be established to access a memory having an access size of a data word. A first slice can be configured to access a first side of the data word in memory. Circuitry can access a mask identifying byte positions within the data word having non-zero values. The circuitry can modify the data word to have non-zero byte values stored starting at an end of the first side, and any zero byte values stored in a remainder of the data word. A determination can be made whether a number of non-zero byte values is less than or equal to a first access size of the first slice. The circuitry can write the modified data word to the memory via at least the first slice.

公开(公告)号：US11507814B1

公开(公告)日：2022-11-22

申请号：US16899806

申请日：2020-06-12

Applicant: Meta Platforms Technologies, LLC

Inventor： Meng Li ,  Yilei Li

IPC: G06N3/063 ,  G06F17/16 ,  G06F7/544

Abstract: Disclosed herein includes a system, a method, and a device for improving power efficiency of a neural network implemented in an AI chip. In a neural network, large amounts of computations for multiply and accumulate can result in frequent toggles or transitions in states of logic circuits in the AI chip. Such frequent toggles or transitions of states of logic circuits can cause a large overall power consumption. In one aspect, to minimize the number of toggles, a sequence or order of computations can be rearranged. In one approach, total hamming distances for weights or input strings in different arrangements or sequences can be identified, and an arrangement or a sequence of weights or input strings with a reduced or minimum total hamming distance can be identified. An arrangement or a sequence of weights that render a reduced total hamming distance can be identified.

公开(公告)号：US11675998B2

公开(公告)日：2023-06-13

申请号：US16511544

申请日：2019-07-15

Applicant: Meta Platforms Technologies, LLC

Inventor： Ganesh Venkatesh ,  Liangzhen Lai ,  Pierce I-Jen Chuang ,  Meng Li

IPC: G06N3/04 ,  G06N3/063

CPC classification number: G06N3/04 ,  G06N3/063

Abstract: Disclosed herein includes a system, a method, and a device for receiving input data to generate a plurality of outputs for a layer of a neural network. The plurality of outputs are arranged in a first array. Dimensions of the first array may be compared with dimensions of a processing unit (PE) array including a plurality of PEs. According to a result of the comparing, the first array is partitioned into subarrays by the processor. Each of the subarrays has dimensions less than or equal to the dimensions of the PE array. A first group of PEs in the PE array is assigned to a first one of the subarrays. A corresponding output of the plurality of outputs is generated using a portion of the input data by each PE of the first group of PEs assigned to the first one of the subarrays.

公开(公告)号：US11954025B2

公开(公告)日：2024-04-09

申请号：US18126228

申请日：2023-03-24

Applicant: Meta Platforms Technologies, LLC

Inventor： Ganesh Venkatesh ,  Liangzhen Lai ,  Pierce I-Jen Chuang ,  Meng Li

IPC: G06F12/04 ,  G06N3/08

CPC classification number: G06F12/04 ,  G06F2212/1028 ,  G06N3/08

Abstract: Disclosed herein includes a system, a method, and a device for reading and writing sparse data in a neural network accelerator. A mask identifying byte positions within a data word having non-zero values in memory can be accessed. Each bit of the mask can have a first value or a second value, the first value indicating that a byte of the data word corresponds to a non-zero byte value, the second value indicating that the byte of the data word corresponds to a zero byte value. The data word can be modified to have non-zero byte values stored at an end of a first side of the data word in the memory, and any zero byte values stored in a remainder of the data word. The modified data word can be written to the memory via at least a first slice of a plurality of slices that is configured to access the first side of the data word in the memory.