DC Field | Value | Language |
---|---|---|
dc.contributor.author | Um, Soyeon | ko |
dc.contributor.author | Kim, Sangyeob | ko |
dc.contributor.author | Kim, Sangjin | ko |
dc.contributor.author | Yoo, Hoi-Jun | ko |
dc.date.accessioned | 2021-06-01T02:10:24Z | - |
dc.date.available | 2021-06-01T02:10:24Z | - |
dc.date.created | 2021-06-01 | - |
dc.date.created | 2021-06-01 | - |
dc.date.issued | 2021-05 | - |
dc.identifier.citation | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS, v.68, no.5, pp.1605 - 1609 | - |
dc.identifier.issn | 1549-7747 | - |
dc.identifier.uri | http://hdl.handle.net/10203/285390 | - |
dc.description.abstract | Recently, Computing-In-Memory (CIM) processors have been proposed to achieve high energy-efficiency by reducing data movement and solving memory bottlenecks. Furthermore, a network with high accurate image classification has been introduced through the Absolute-Difference-Accumulation (ADA) operation instead of the multiplication-and-accumulation operation, which is widely used in DNN. ADA operation provides not only opportunities for high energy-efficient DNN accelerating by reducing multiplication but also a chance to reuse computation results. However, the previous CIM processor cannot reuse previous computation results for other computations. In this brief, we propose a high accurate and high energy-efficient ADA-CIM processor that with two key features: 1) computation reuse for low-power, resulting in a 49.5% CIM operation power reduction, and 2) low-cost sign prediction core with 3-bit activation and weight quantization for high utilization. From the two key features, the proposed ADA-CIM processor is simulated in 28 nm CMOS technology and occupies 3.78 mm(2). It consumes 2.77mW and achieves 43.1 TOPS/W energy-efficiency with a high-accuracy of 91.62% at CIFAR-10 (ResNet-20). | - |
dc.language | English | - |
dc.publisher | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC | - |
dc.title | A 43.1TOPS/W Energy-Efficient Absolute-Difference-Accumulation Operation Computing-In-Memory With Computation Reuse | - |
dc.type | Article | - |
dc.identifier.wosid | 000645863300010 | - |
dc.identifier.scopusid | 2-s2.0-85103250029 | - |
dc.type.rims | ART | - |
dc.citation.volume | 68 | - |
dc.citation.issue | 5 | - |
dc.citation.beginningpage | 1605 | - |
dc.citation.endingpage | 1609 | - |
dc.citation.publicationname | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS | - |
dc.identifier.doi | 10.1109/TCSII.2021.3067327 | - |
dc.contributor.localauthor | Yoo, Hoi-Jun | - |
dc.contributor.nonIdAuthor | Um, Soyeon | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Random access memory | - |
dc.subject.keywordAuthor | Program processors | - |
dc.subject.keywordAuthor | Energy efficiency | - |
dc.subject.keywordAuthor | Common Information Model (computing) | - |
dc.subject.keywordAuthor | Quantization (signal) | - |
dc.subject.keywordAuthor | Computer architecture | - |
dc.subject.keywordAuthor | Circuits and systems | - |
dc.subject.keywordAuthor | Absolute-difference-accumulation (ADA) operation | - |
dc.subject.keywordAuthor | computation reuse | - |
dc.subject.keywordAuthor | computing-in-memory (CIM) | - |
dc.subject.keywordAuthor | energy-efficient | - |
dc.subject.keywordAuthor | low-cost absolute difference sign prediction | - |
dc.subject.keywordAuthor | SRAM | - |
dc.subject.keywordPlus | SRAM | - |
dc.subject.keywordPlus | MACRO | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.