NAPEL: Near-memory computing application performance prediction via ensemble learning

Gagandeep Singh, Juan Gómez-Luna, Giovanni Mariani, Geraldo F. Oliveira, Stefano Corda, Sander Stuijk, Onur Mutlu, Henk Corporaal

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

3 Citaties (Scopus)

Uittreksel

The cost of moving data between the memory/storage units and the compute units is a major contributor to the execution time and energy consumption of modern workloads in computing systems. A promising paradigm to alleviate this data movement bottleneck is near-memory computing (NMC), which consists of placing compute units close to the memory/storage units. There is substantial research effort that proposesNMCarchitectures and identifiesworkloads that can benefit from NMC. System architects typically use simulation techniques to evaluate the performance and energy consumption of their designs. However, simulation is extremely slow, imposing long times for design space exploration. In order to enable fast early-stage design space exploration of NMC architectures, we need high-level performance and energy models. We present NAPEL, a high-level performance and energy estimation framework for NMC architectures. NAPEL leverages ensemble learning to develop a model that is based on microarchitectural parameters and application characteristics. NAPEL training uses a statistical technique, called design of experiments, to collect representative training data efficiently. NAPEL provides early design space exploration 220× faster than a state-of-the-artNMCsimulator, on average, with error rates of to 8.5% and 11.6% for performance and energy estimations, respectively, compared to the NMC simulator. NAPEL is also capable of making accurate predictions for previously-unseen applications.

TaalEngels
TitelProceedings of the 56th Annual Design Automation Conference 2019, DAC 2019
UitgeverijInstitute of Electrical and Electronics Engineers
Aantal pagina's6
ISBN van elektronische versie9781450367257
DOI's
StatusGepubliceerd - 2 jun 2019
Evenement56th Annual Design Automation Conference, (DAC2019) - Las Vegas, Verenigde Staten van Amerika
Duur: 2 jun 20196 jun 2019
https://dac.com/

Congres

Congres56th Annual Design Automation Conference, (DAC2019)
Verkorte titelDAC2019
LandVerenigde Staten van Amerika
StadLas Vegas
Periode2/06/196/06/19
Internet adres

Vingerafdruk

Ensemble Learning
Performance Prediction
Data storage equipment
Computing
Design Space Exploration
Unit
Energy Consumption
Energy utilization
Energy Model
Design of Experiments
Performance Model
Energy
Leverage
Design of experiments
Execution Time
Workload
Error Rate
Simulation
Simulator
Simulators

Citeer dit

Singh, G., Gómez-Luna, J., Mariani, G., Oliveira, G. F., Corda, S., Stuijk, S., ... Corporaal, H. (2019). NAPEL: Near-memory computing application performance prediction via ensemble learning. In Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019 [a27] Institute of Electrical and Electronics Engineers. DOI: 10.1145/3316781.3317867
Singh, Gagandeep ; Gómez-Luna, Juan ; Mariani, Giovanni ; Oliveira, Geraldo F. ; Corda, Stefano ; Stuijk, Sander ; Mutlu, Onur ; Corporaal, Henk. / NAPEL : Near-memory computing application performance prediction via ensemble learning. Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019. Institute of Electrical and Electronics Engineers, 2019.
@inproceedings{1a4b705b437444cf95523e159ce3ae50,
title = "NAPEL: Near-memory computing application performance prediction via ensemble learning",
abstract = "The cost of moving data between the memory/storage units and the compute units is a major contributor to the execution time and energy consumption of modern workloads in computing systems. A promising paradigm to alleviate this data movement bottleneck is near-memory computing (NMC), which consists of placing compute units close to the memory/storage units. There is substantial research effort that proposesNMCarchitectures and identifiesworkloads that can benefit from NMC. System architects typically use simulation techniques to evaluate the performance and energy consumption of their designs. However, simulation is extremely slow, imposing long times for design space exploration. In order to enable fast early-stage design space exploration of NMC architectures, we need high-level performance and energy models. We present NAPEL, a high-level performance and energy estimation framework for NMC architectures. NAPEL leverages ensemble learning to develop a model that is based on microarchitectural parameters and application characteristics. NAPEL training uses a statistical technique, called design of experiments, to collect representative training data efficiently. NAPEL provides early design space exploration 220× faster than a state-of-the-artNMCsimulator, on average, with error rates of to 8.5{\%} and 11.6{\%} for performance and energy estimations, respectively, compared to the NMC simulator. NAPEL is also capable of making accurate predictions for previously-unseen applications.",
author = "Gagandeep Singh and Juan G{\'o}mez-Luna and Giovanni Mariani and Oliveira, {Geraldo F.} and Stefano Corda and Sander Stuijk and Onur Mutlu and Henk Corporaal",
year = "2019",
month = "6",
day = "2",
doi = "10.1145/3316781.3317867",
language = "English",
booktitle = "Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Singh, G, Gómez-Luna, J, Mariani, G, Oliveira, GF, Corda, S, Stuijk, S, Mutlu, O & Corporaal, H 2019, NAPEL: Near-memory computing application performance prediction via ensemble learning. in Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019., a27, Institute of Electrical and Electronics Engineers, Las Vegas, Verenigde Staten van Amerika, 2/06/19. DOI: 10.1145/3316781.3317867

NAPEL : Near-memory computing application performance prediction via ensemble learning. / Singh, Gagandeep; Gómez-Luna, Juan; Mariani, Giovanni; Oliveira, Geraldo F.; Corda, Stefano; Stuijk, Sander; Mutlu, Onur; Corporaal, Henk.

Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019. Institute of Electrical and Electronics Engineers, 2019. a27.

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

TY - GEN

T1 - NAPEL

T2 - Near-memory computing application performance prediction via ensemble learning

AU - Singh,Gagandeep

AU - Gómez-Luna,Juan

AU - Mariani,Giovanni

AU - Oliveira,Geraldo F.

AU - Corda,Stefano

AU - Stuijk,Sander

AU - Mutlu,Onur

AU - Corporaal,Henk

PY - 2019/6/2

Y1 - 2019/6/2

N2 - The cost of moving data between the memory/storage units and the compute units is a major contributor to the execution time and energy consumption of modern workloads in computing systems. A promising paradigm to alleviate this data movement bottleneck is near-memory computing (NMC), which consists of placing compute units close to the memory/storage units. There is substantial research effort that proposesNMCarchitectures and identifiesworkloads that can benefit from NMC. System architects typically use simulation techniques to evaluate the performance and energy consumption of their designs. However, simulation is extremely slow, imposing long times for design space exploration. In order to enable fast early-stage design space exploration of NMC architectures, we need high-level performance and energy models. We present NAPEL, a high-level performance and energy estimation framework for NMC architectures. NAPEL leverages ensemble learning to develop a model that is based on microarchitectural parameters and application characteristics. NAPEL training uses a statistical technique, called design of experiments, to collect representative training data efficiently. NAPEL provides early design space exploration 220× faster than a state-of-the-artNMCsimulator, on average, with error rates of to 8.5% and 11.6% for performance and energy estimations, respectively, compared to the NMC simulator. NAPEL is also capable of making accurate predictions for previously-unseen applications.

AB - The cost of moving data between the memory/storage units and the compute units is a major contributor to the execution time and energy consumption of modern workloads in computing systems. A promising paradigm to alleviate this data movement bottleneck is near-memory computing (NMC), which consists of placing compute units close to the memory/storage units. There is substantial research effort that proposesNMCarchitectures and identifiesworkloads that can benefit from NMC. System architects typically use simulation techniques to evaluate the performance and energy consumption of their designs. However, simulation is extremely slow, imposing long times for design space exploration. In order to enable fast early-stage design space exploration of NMC architectures, we need high-level performance and energy models. We present NAPEL, a high-level performance and energy estimation framework for NMC architectures. NAPEL leverages ensemble learning to develop a model that is based on microarchitectural parameters and application characteristics. NAPEL training uses a statistical technique, called design of experiments, to collect representative training data efficiently. NAPEL provides early design space exploration 220× faster than a state-of-the-artNMCsimulator, on average, with error rates of to 8.5% and 11.6% for performance and energy estimations, respectively, compared to the NMC simulator. NAPEL is also capable of making accurate predictions for previously-unseen applications.

UR - http://www.scopus.com/inward/record.url?scp=85067833914&partnerID=8YFLogxK

U2 - 10.1145/3316781.3317867

DO - 10.1145/3316781.3317867

M3 - Conference contribution

BT - Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019

PB - Institute of Electrical and Electronics Engineers

ER -

Singh G, Gómez-Luna J, Mariani G, Oliveira GF, Corda S, Stuijk S et al. NAPEL: Near-memory computing application performance prediction via ensemble learning. In Proceedings of the 56th Annual Design Automation Conference 2019, DAC 2019. Institute of Electrical and Electronics Engineers. 2019. a27. Beschikbaar vanaf, DOI: 10.1145/3316781.3317867