Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387667075> ?p ?o ?g. }
- W4387667075 endingPage "30" @default.
- W4387667075 startingPage "1" @default.
- W4387667075 abstract "Binary spatter code (BSC)-based hyperdimensional computing (HDC) is a highly error-resilient approximate computational paradigm suited for error-prone, emerging hardware platforms. In BSC HDC, the basic datatype is a hypervector , a typically large binary vector, where the size of the hypervector has a significant impact on the fidelity and resource usage of the computation. Typically, the hypervector size is dynamically tuned to deliver the desired accuracy; this process is time-consuming and often produces hypervector sizes that lack accuracy guarantees and produce poor results when reused for very similar workloads. We present Heim, a hardware-aware static analysis and optimization framework for BSC HD computations. Heim analytically derives the minimum hypervector size that minimizes resource usage and meets the target accuracy requirement. Heim guarantees the optimized computation converges to the user-provided accuracy target on expectation, even in the presence of hardware error. Heim deploys a novel static analysis procedure that unifies theoretical results from the neuroscience community to systematically optimize HD computations. We evaluate Heim against dynamic tuning-based optimization on 25 benchmark data structures. Given a 99% accuracy requirement, Heim-optimized computations achieve a 99.2%-100.0% median accuracy, up to 49.5% higher than dynamic tuning-based optimization, while achieving 1.15x-7.14x reductions in hypervector size compared to HD computations that achieve comparable query accuracy and finding parametrizations 30.0x-100167.4x faster than dynamic tuning-based approaches. We also use Heim to systematically evaluate the performance benefits of using analog CAMs and multiple-bit-per-cell ReRAM over conventional hardware, while maintaining iso-accuracy – for both emerging technologies, we find usages where the emerging hardware imparts significant benefits." @default.
- W4387667075 created "2023-10-17" @default.
- W4387667075 creator A5030371417 @default.
- W4387667075 creator A5060778165 @default.
- W4387667075 date "2023-10-16" @default.
- W4387667075 modified "2023-10-17" @default.
- W4387667075 title "Hardware-Aware Static Optimization of Hyperdimensional Computations" @default.
- W4387667075 cites W1967540748 @default.
- W4387667075 cites W1986707196 @default.
- W4387667075 cites W2000150040 @default.
- W4387667075 cites W2001972101 @default.
- W4387667075 cites W2042588857 @default.
- W4387667075 cites W2117655830 @default.
- W4387667075 cites W2147657366 @default.
- W4387667075 cites W2165121732 @default.
- W4387667075 cites W2554538030 @default.
- W4387667075 cites W2583857261 @default.
- W4387667075 cites W2613567208 @default.
- W4387667075 cites W2624514417 @default.
- W4387667075 cites W2770867854 @default.
- W4387667075 cites W2771100829 @default.
- W4387667075 cites W2809429070 @default.
- W4387667075 cites W2895774650 @default.
- W4387667075 cites W2895910969 @default.
- W4387667075 cites W2906925223 @default.
- W4387667075 cites W2908729710 @default.
- W4387667075 cites W2914250845 @default.
- W4387667075 cites W2941283033 @default.
- W4387667075 cites W2950268605 @default.
- W4387667075 cites W2951498229 @default.
- W4387667075 cites W2962967667 @default.
- W4387667075 cites W2963650979 @default.
- W4387667075 cites W2971739303 @default.
- W4387667075 cites W2981174422 @default.
- W4387667075 cites W2990838824 @default.
- W4387667075 cites W3005885031 @default.
- W4387667075 cites W3036478084 @default.
- W4387667075 cites W3101009183 @default.
- W4387667075 cites W3133016223 @default.
- W4387667075 cites W3186511633 @default.
- W4387667075 cites W3190134129 @default.
- W4387667075 cites W3201072697 @default.
- W4387667075 cites W3204836041 @default.
- W4387667075 cites W3211097789 @default.
- W4387667075 cites W3212827699 @default.
- W4387667075 cites W3214202924 @default.
- W4387667075 cites W4230025805 @default.
- W4387667075 cites W4243061752 @default.
- W4387667075 cites W4293023530 @default.
- W4387667075 cites W4293216350 @default.
- W4387667075 cites W4308481194 @default.
- W4387667075 cites W4312242794 @default.
- W4387667075 cites W4312787185 @default.
- W4387667075 cites W4319302948 @default.
- W4387667075 cites W4362459302 @default.
- W4387667075 cites W4376604041 @default.
- W4387667075 cites W4386556380 @default.
- W4387667075 doi "https://doi.org/10.1145/3622797" @default.
- W4387667075 hasPublicationYear "2023" @default.
- W4387667075 type Work @default.
- W4387667075 citedByCount "0" @default.
- W4387667075 crossrefType "journal-article" @default.
- W4387667075 hasAuthorship W4387667075A5030371417 @default.
- W4387667075 hasAuthorship W4387667075A5060778165 @default.
- W4387667075 hasBestOaLocation W43876670751 @default.
- W4387667075 hasConcept C111919701 @default.
- W4387667075 hasConcept C113775141 @default.
- W4387667075 hasConcept C11413529 @default.
- W4387667075 hasConcept C13280743 @default.
- W4387667075 hasConcept C139571649 @default.
- W4387667075 hasConcept C169590947 @default.
- W4387667075 hasConcept C173608175 @default.
- W4387667075 hasConcept C177264268 @default.
- W4387667075 hasConcept C185798385 @default.
- W4387667075 hasConcept C199360897 @default.
- W4387667075 hasConcept C205649164 @default.
- W4387667075 hasConcept C2776760102 @default.
- W4387667075 hasConcept C33923547 @default.
- W4387667075 hasConcept C41008148 @default.
- W4387667075 hasConcept C45374587 @default.
- W4387667075 hasConcept C48372109 @default.
- W4387667075 hasConcept C9390403 @default.
- W4387667075 hasConcept C94375191 @default.
- W4387667075 hasConcept C98045186 @default.
- W4387667075 hasConceptScore W4387667075C111919701 @default.
- W4387667075 hasConceptScore W4387667075C113775141 @default.
- W4387667075 hasConceptScore W4387667075C11413529 @default.
- W4387667075 hasConceptScore W4387667075C13280743 @default.
- W4387667075 hasConceptScore W4387667075C139571649 @default.
- W4387667075 hasConceptScore W4387667075C169590947 @default.
- W4387667075 hasConceptScore W4387667075C173608175 @default.
- W4387667075 hasConceptScore W4387667075C177264268 @default.
- W4387667075 hasConceptScore W4387667075C185798385 @default.
- W4387667075 hasConceptScore W4387667075C199360897 @default.
- W4387667075 hasConceptScore W4387667075C205649164 @default.
- W4387667075 hasConceptScore W4387667075C2776760102 @default.
- W4387667075 hasConceptScore W4387667075C33923547 @default.
- W4387667075 hasConceptScore W4387667075C41008148 @default.