-
公开(公告)号:US20220199185A1
公开(公告)日:2022-06-23
申请号:US17392781
申请日:2021-08-03
Applicant: Samsung Electronics Co., Ltd.
Inventor: Jongyoon YOON , Hyeonjong SONG , Seonghyeog CHOI , Hongrak SON
Abstract: In a method of predicting a remaining lifetime of the nonvolatile memory device, a read sequence is performed. The read sequence includes a plurality of read operations, and at least one of the plurality of read operations is sequentially performed until read data stored in the nonvolatile memory device is successfully retrieved. Sequence class and error correction code (ECC) decoding information are generated. A life stage of the nonvolatile memory device is determined based on at least one of the sequence class and the ECC decoding information. When it is determined that the nonvolatile memory device corresponds to a first life stage, a coarse prediction on the remaining lifetime of the nonvolatile memory device is performed. When it is determined that the nonvolatile memory device corresponds to a second life stage after the first life stage, a fine prediction on the remaining lifetime of the nonvolatile memory device is performed.
-
公开(公告)号:US20250077179A1
公开(公告)日:2025-03-06
申请号:US18586002
申请日:2024-02-23
Applicant: SAMSUNG ELECTRONICS CO., LTD.
Inventor: Hoyoun KIM , Jongyoon YOON
Abstract: A memory device for supporting machine learning includes a first cell array configured to store weight data at a first precision or a second precision, a second cell array configured to store loss data at the first precision, a third cell array configured to store gradient data at the first precision, and a computation circuit configured to perform at least one of a multiplying operation, a dividing operation, or a rounding operation corresponding to a scaling factor during mixed-precision training that uses the first precision and the second precision.
-