About: MemMAP: Compact and Generalizable Meta-LSTM Models for Memory Access Prediction

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: MemMAP: Compact and Generalizable Meta-LSTM Models for Memory Access Prediction Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : covidontheweb.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	MemMAP: Compact and Generalizable Meta-LSTM Models for Memory Access Prediction
Creator	Kannan, Rajgopal Prasanna, Viktor Srivastava, Ajitesh Augusto, Cesar De Rose, F Wang, Ta-Yang Zhang, Pengmiao
Source	PMC
abstract	With the rise of Big Data, there has been a significant effort in increasing compute power through GPUs, TPUs, and heterogeneous architectures. As a result, many applications are memory bound, i.e., they are bottlenecked by the movement of data from main memory to compute units. One way to address this issue is through data prefetching, which relies on accurate prediction of memory accesses. While recent deep learning models have performed well on sequence prediction problems, they are far too heavy in terms of model size and inference latency to be practical for data prefetching. Here, we propose extremely compact LSTM models that can predict the next memory access with high accuracy. Prior LSTM based work on access prediction has used orders of magnitude more parameters and developed one model for each application (trace). While one (specialized) model per application can result in more accuracy, it is not a scalable approach. In contrast, our models can predict for a class of applications by trading off specialization at the cost of few retraining steps at runtime, for a more generalizable compact meta-model. Our experiments on 13 benchmark applications demonstrate that three compact meta-models can obtain accuracy close to specialized models using few batches of retraining for majority of the applications.
has issue date	2020-04-17(xsd:dateTime)
bibo:doi	10.1007/978-3-030-47436-2_5
has license	no-cc
sha1sum (hex)	66178117b0f45a4700170b95b866658cf8c89323
schema:url	https://doi.org/10.1007/978-3-030-47436-2_5
resource representing a document's title	MemMAP: Compact and Generalizable Meta-LSTM Models for Memory Access Prediction
has PubMed Central identifier	PMC7206302
schema:publication	Advances in Knowledge Discovery and Data Mining
resource representing a document's body	covid:66178117b0f45a4700170b95b866658cf8c89323#body_text
is schema:about of	named entity 'MODELS' named entity 'UNITS' named entity 'ADDRESS' named entity 'ACCURACY' named entity 'ACCESSES' named entity 'EXTREMELY' named entity 'HERE' named entity 'MEMORY' named entity 'MAIN' named entity 'SEQUENCE' covid:arg/66178117b0f45a4700170b95b866658cf8c89323 named entity 'prefetching' named entity 'prefetching' named entity 'Big Data' named entity 'LSTM' named entity 'compute units' named entity 'memory bound' named entity 'orders of magnitude' named entity 'k-means clustering' named entity 'concatenating' named entity 'LSTM' named entity 'deep learning' named entity 'metamodel' named entity 'LSTM' named entity 'intuition' named entity 'meta-model' named entity 'meta-model' named entity 'activation function' named entity 'neural network' named entity 'meta-model' named entity 'binary representation' named entity 'LSTM' named entity 'LSTM' named entity 'memory locations' named entity 'inference' named entity 'data center' named entity 'decimal' named entity 'MAML' named entity 'meta-model' named entity 'memory footprint' named entity 'machine learning' named entity 'cross-entropy' named entity 'MAML' named entity 'Agnostic' named entity 'LSTM' named entity 'Swaption' named entity 'LSTM' named entity 'logistic regression' named entity 'low latency' named entity 'LSTM' named entity 'decision trees' named entity 'meta-model' named entity 'inference' named entity 'PARSEC' named entity 'LSTM' named entity 'binary representation' named entity 'inference' named entity 'standard deviation' named entity 'Concatenated' named entity 'LSTM' named entity 'Raytrace' named entity 'Raytrace' named entity 'k-means' named entity 'Neural networks' named entity 'prefetchers' named entity 'LSTM' named entity 'meta-model' named entity 'PARSEC' named entity 'deep learning' named entity 'Raytrace' named entity 'memory locations' named entity 'training set' named entity 'time-series data' named entity 'Raytrace' named entity 'cache misses' named entity 'LSTM' named entity 'memory addresses'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software