Invention Grant
- Patent Title: Deep learning accelerator architecture with chunking GEMM
-
Application No.: US15957711Application Date: 2018-04-19
-
Publication No.: US10657442B2Publication Date: 2020-05-19
- Inventor: Naigang Wang , Jungwook Choi , Kailash Gopalakrishnan , Daniel Brand
- Applicant: International Business Machines Corporation
- Applicant Address: US NY Armonk
- Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
- Current Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
- Current Assignee Address: US NY Armonk
- Agency: Garg Law Firm, PLLC
- Agent Rakesh Garg; Joseph Petrokaitis
- Main IPC: G06N3/08
- IPC: G06N3/08 ; G06F17/16 ; G06F7/483

Abstract:
A compute matrix is configured to include a set of compute units, each compute unit including a multiplier and an accumulator, each of the multiplier and the accumulator formed using at least one floating point unit (FPU). An accumulator array is configured to include a set of external accumulators. The compute matrix is operated to produce a chunk dot-product using a first chunk of a first input vector and a first chunk of a second input vector. The accumulator array is operated to output a dot-product of the first input vector and the second input vector using the chunk dot-product.
Public/Granted literature
- US20190325301A1 DEEP LEARNING ACCELERATOR ARCHITECTURE WITH CHUNKING GEMM Public/Granted day:2019-10-24
Information query