DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 유회준 | - |
dc.contributor.author | Hong, Seongyon | - |
dc.contributor.author | 홍성연 | - |
dc.date.accessioned | 2024-07-25T19:31:12Z | - |
dc.date.available | 2024-07-25T19:31:12Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1045892&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/320663 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2023.8,[iii, 18 p. :] | - |
dc.description.abstract | Recent computing-in-memory (CIM) achieves high energy efficiency with charge-domain computation and multi-bit input driving. However, the previous works still require high power consumption and trade computation signal-to-noise ratio (SNR) for energy efficiency. This work proposes an energy-efficient and accurate multi-bit input/weight-parallel CIM processor with four key features: 1) a 10T2C sign-magnitude cell with voltage-capacitance-ratio (VCR) decoding for 5-bit analog inputs with only 2-level supply voltages, 2) a computation word line (CWL) charge reuse method for input driver power reduction, 3) a signal-amplifying noise canceling voltage-to-time converter (SANC-VTC) for SNR improvement, and 4) a distribution-aware time-to-digital converter (DA-TDC) for ADC power reduction. The proposed CIM processor is simulated in 28 nm CMOS technology with 1.25 mm$^2$ area. As a result, it achieves 4.44 mW power consumption and 332 TOPS/W energy efficiency with 72.43% benchmark accuracy (@ ImageNet, ResNet50, 5-bit input/5-bit weight). | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 컴퓨팅 인 메모리▼a심층 신경망▼a에너지 효율▼aSRAM▼a시간 기반 ADC | - |
dc.subject | Computing-in-memory▼aDeep neural network (DNN)▼aEnergy efficiency▼aSRAM▼aTime-based ADC | - |
dc.title | (An) in-memory parallel computing processor for energy-efficient and high accuracy deep neural network computation | - |
dc.title.alternative | 에너지 효율적인 고정확도 심층 신경망 연산을 위한 메모리 내 병렬 연산 프로세서 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
dc.contributor.alternativeauthor | Yoo, Hoi-jun | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.