Matches in SemOpenAlex for { <https://semopenalex.org/work/W3135380350> ?p ?o ?g. }
- W3135380350 abstract "With the growing number of data-intensive workloads, GPU, which is the state-of-the-art single-instruction-multiple-thread (SIMT) processor, is hindered by the memory bandwidth wall. To alleviate this bottleneck, previously proposed 3D-stacking near-bank computing accelerators benefit from abundant bank-internal bandwidth by bringing computations closer to the DRAM banks. However, these accelerators are specialized for certain application domains with simple architecture data paths and customized software mapping schemes. For general purpose scenarios, lightweight hardware designs for diverse data paths, architectural supports for the SIMT programming model, and end-to-end software optimizations remain challenging. To address these issues, we propose MPU (Memory-centric Processing Unit), the first SIMT processor based on 3D-stacking near-bank computing architecture. First, to realize diverse data paths with small overheads while leveraging bank-level bandwidth, MPU adopts a hybrid pipeline with the capability of offloading instructions to near-bank compute-logic. Second, we explore two architectural supports for the SIMT programming model, including a near-bank shared memory design and a multiple activated row-buffers enhancement. Third, we present an end-to-end compilation flow for MPU to support CUDA programs. To fully utilize MPU's hybrid pipeline, we develop a backend optimization for the instruction offloading decision. The evaluation results of MPU demonstrate 3.46x speedup and 2.57x energy reduction compared with an NVIDIA Tesla V100 GPU on a set of representative data-intensive workloads." @default.
- W3135380350 created "2021-03-15" @default.
- W3135380350 creator A5046896624 @default.
- W3135380350 creator A5047767267 @default.
- W3135380350 creator A5048052285 @default.
- W3135380350 creator A5068606980 @default.
- W3135380350 creator A5078298822 @default.
- W3135380350 creator A5082076121 @default.
- W3135380350 date "2021-03-11" @default.
- W3135380350 modified "2023-09-23" @default.
- W3135380350 title "MPU: Towards Bandwidth-abundant SIMT Processor via Near-bank Computing" @default.
- W3135380350 cites W1975237352 @default.
- W3135380350 cites W1979527452 @default.
- W3135380350 cites W1982825626 @default.
- W3135380350 cites W1999085092 @default.
- W3135380350 cites W2005395913 @default.
- W3135380350 cites W2034861439 @default.
- W3135380350 cites W2048466306 @default.
- W3135380350 cites W2055312318 @default.
- W3135380350 cites W2079787774 @default.
- W3135380350 cites W2080592089 @default.
- W3135380350 cites W2086112773 @default.
- W3135380350 cites W2092324191 @default.
- W3135380350 cites W2094332102 @default.
- W3135380350 cites W2112181056 @default.
- W3135380350 cites W2112980698 @default.
- W3135380350 cites W2114440330 @default.
- W3135380350 cites W2116784058 @default.
- W3135380350 cites W2118231264 @default.
- W3135380350 cites W2118703320 @default.
- W3135380350 cites W2129991978 @default.
- W3135380350 cites W2141546789 @default.
- W3135380350 cites W2150909864 @default.
- W3135380350 cites W2169880332 @default.
- W3135380350 cites W2291192530 @default.
- W3135380350 cites W2291750097 @default.
- W3135380350 cites W2408724663 @default.
- W3135380350 cites W2414912620 @default.
- W3135380350 cites W2488627141 @default.
- W3135380350 cites W2497599918 @default.
- W3135380350 cites W2508602506 @default.
- W3135380350 cites W2513721464 @default.
- W3135380350 cites W2517869808 @default.
- W3135380350 cites W2518281301 @default.
- W3135380350 cites W2536390129 @default.
- W3135380350 cites W2543989436 @default.
- W3135380350 cites W2560443438 @default.
- W3135380350 cites W2562213348 @default.
- W3135380350 cites W2575503711 @default.
- W3135380350 cites W2605347906 @default.
- W3135380350 cites W2612654866 @default.
- W3135380350 cites W2761132374 @default.
- W3135380350 cites W2764172931 @default.
- W3135380350 cites W2767588966 @default.
- W3135380350 cites W2887018074 @default.
- W3135380350 cites W2896090304 @default.
- W3135380350 cites W2897709121 @default.
- W3135380350 cites W2904929935 @default.
- W3135380350 cites W2907701003 @default.
- W3135380350 cites W2912888224 @default.
- W3135380350 cites W2913789423 @default.
- W3135380350 cites W2949989598 @default.
- W3135380350 cites W2967163762 @default.
- W3135380350 cites W2979823675 @default.
- W3135380350 cites W2980235049 @default.
- W3135380350 cites W2985229340 @default.
- W3135380350 cites W2991330024 @default.
- W3135380350 cites W3011446983 @default.
- W3135380350 cites W3039617420 @default.
- W3135380350 cites W3042598257 @default.
- W3135380350 cites W3098742859 @default.
- W3135380350 cites W3099365437 @default.
- W3135380350 doi "https://doi.org/10.48550/arxiv.2103.06653" @default.
- W3135380350 hasPublicationYear "2021" @default.
- W3135380350 type Work @default.
- W3135380350 sameAs 3135380350 @default.
- W3135380350 citedByCount "0" @default.
- W3135380350 crossrefType "posted-content" @default.
- W3135380350 hasAuthorship W3135380350A5046896624 @default.
- W3135380350 hasAuthorship W3135380350A5047767267 @default.
- W3135380350 hasAuthorship W3135380350A5048052285 @default.
- W3135380350 hasAuthorship W3135380350A5068606980 @default.
- W3135380350 hasAuthorship W3135380350A5078298822 @default.
- W3135380350 hasAuthorship W3135380350A5082076121 @default.
- W3135380350 hasBestOaLocation W31353803501 @default.
- W3135380350 hasConcept C111919701 @default.
- W3135380350 hasConcept C118524514 @default.
- W3135380350 hasConcept C149635348 @default.
- W3135380350 hasConcept C173608175 @default.
- W3135380350 hasConcept C188045654 @default.
- W3135380350 hasConcept C199360897 @default.
- W3135380350 hasConcept C2776257435 @default.
- W3135380350 hasConcept C2777904410 @default.
- W3135380350 hasConcept C2778119891 @default.
- W3135380350 hasConcept C2780513914 @default.
- W3135380350 hasConcept C31258907 @default.
- W3135380350 hasConcept C34165917 @default.
- W3135380350 hasConcept C41008148 @default.
- W3135380350 hasConcept C43521106 @default.
- W3135380350 hasConcept C68339613 @default.