Matches in SemOpenAlex for { <https://semopenalex.org/work/W2896819009> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W2896819009 abstract "Domain specific accelerators present new challenges and opportunities for code generation onto novel instruction sets, communication fabrics, and memory architectures. In this paper we introduce an intermediate representation (IR) which enables both deep learning computational kernels and hardware capabilities to be described in the same IR. We then formulate and apply instruction mapping to determine the possible ways a computation can be performed on a hardware system. Next, our scheduler chooses a specific mapping and determines the data movement and computation order. In order to manage the large search space of mappings and schedules, we developed a flexible framework that allows heuristics, cost models, and potentially machine learning to facilitate this search problem. With this system, we demonstrate the automated extraction of matrix multiplication kernels out of recent deep learning kernels such as depthwise-separable convolution. In addition, we demonstrate two to five times better performance on DeepBench sized GEMMs and GRU RNN execution when compared to state-of-the-art (SOTA) implementations on new hardware and up to 85% of the performance for SOTA implementations on existing hardware." @default.
- W2896819009 created "2018-10-26" @default.
- W2896819009 creator A5016696594 @default.
- W2896819009 creator A5022752991 @default.
- W2896819009 creator A5024066643 @default.
- W2896819009 creator A5057664789 @default.
- W2896819009 creator A5078599914 @default.
- W2896819009 creator A5079165150 @default.
- W2896819009 date "2018-10-12" @default.
- W2896819009 modified "2023-09-23" @default.
- W2896819009 title "ISA Mapper: A Compute and Hardware Agnostic Deep Learning Compiler" @default.
- W2896819009 cites W2055312318 @default.
- W2896819009 cites W2057455870 @default.
- W2896819009 cites W2147405597 @default.
- W2896819009 cites W2784561332 @default.
- W2896819009 cites W2786320458 @default.
- W2896819009 cites W2787513823 @default.
- W2896819009 cites W2788183263 @default.
- W2896819009 cites W2886857292 @default.
- W2896819009 cites W2949251082 @default.
- W2896819009 cites W2963960923 @default.
- W2896819009 hasPublicationYear "2018" @default.
- W2896819009 type Work @default.
- W2896819009 sameAs 2896819009 @default.
- W2896819009 citedByCount "0" @default.
- W2896819009 crossrefType "posted-content" @default.
- W2896819009 hasAuthorship W2896819009A5016696594 @default.
- W2896819009 hasAuthorship W2896819009A5022752991 @default.
- W2896819009 hasAuthorship W2896819009A5024066643 @default.
- W2896819009 hasAuthorship W2896819009A5057664789 @default.
- W2896819009 hasAuthorship W2896819009A5078599914 @default.
- W2896819009 hasAuthorship W2896819009A5079165150 @default.
- W2896819009 hasConcept C108583219 @default.
- W2896819009 hasConcept C111919701 @default.
- W2896819009 hasConcept C113775141 @default.
- W2896819009 hasConcept C118524514 @default.
- W2896819009 hasConcept C127705205 @default.
- W2896819009 hasConcept C154945302 @default.
- W2896819009 hasConcept C169590947 @default.
- W2896819009 hasConcept C173608175 @default.
- W2896819009 hasConcept C199360897 @default.
- W2896819009 hasConcept C26713055 @default.
- W2896819009 hasConcept C41008148 @default.
- W2896819009 hasConcept C45374587 @default.
- W2896819009 hasConceptScore W2896819009C108583219 @default.
- W2896819009 hasConceptScore W2896819009C111919701 @default.
- W2896819009 hasConceptScore W2896819009C113775141 @default.
- W2896819009 hasConceptScore W2896819009C118524514 @default.
- W2896819009 hasConceptScore W2896819009C127705205 @default.
- W2896819009 hasConceptScore W2896819009C154945302 @default.
- W2896819009 hasConceptScore W2896819009C169590947 @default.
- W2896819009 hasConceptScore W2896819009C173608175 @default.
- W2896819009 hasConceptScore W2896819009C199360897 @default.
- W2896819009 hasConceptScore W2896819009C26713055 @default.
- W2896819009 hasConceptScore W2896819009C41008148 @default.
- W2896819009 hasConceptScore W2896819009C45374587 @default.
- W2896819009 hasOpenAccess W2896819009 @default.
- W2896819009 hasRelatedWork W1982970530 @default.
- W2896819009 hasRelatedWork W2510196892 @default.
- W2896819009 hasRelatedWork W2733902982 @default.
- W2896819009 hasRelatedWork W2912704281 @default.
- W2896819009 hasRelatedWork W2941583287 @default.
- W2896819009 hasRelatedWork W2962929233 @default.
- W2896819009 hasRelatedWork W2978910060 @default.
- W2896819009 hasRelatedWork W2984989518 @default.
- W2896819009 hasRelatedWork W3000852400 @default.
- W2896819009 hasRelatedWork W3017000457 @default.
- W2896819009 hasRelatedWork W3030491772 @default.
- W2896819009 hasRelatedWork W3096977550 @default.
- W2896819009 hasRelatedWork W3098369763 @default.
- W2896819009 hasRelatedWork W3101746700 @default.
- W2896819009 hasRelatedWork W3108220046 @default.
- W2896819009 hasRelatedWork W3129512572 @default.
- W2896819009 hasRelatedWork W3159018959 @default.
- W2896819009 hasRelatedWork W3169689544 @default.
- W2896819009 hasRelatedWork W3186087474 @default.
- W2896819009 hasRelatedWork W3189524238 @default.
- W2896819009 isParatext "false" @default.
- W2896819009 isRetracted "false" @default.
- W2896819009 magId "2896819009" @default.
- W2896819009 workType "article" @default.