SlideShare a Scribd company logo
1 of 13
Download to read offline
21st International Middleware Conference
December 7 - 11, 2020, Delft, Netherlands
PipeTune: Pipeline Parallelism of Hyper and System
Parameters Tuning for Deep Learning Clusters
Isabelly Rocha1, Nathaniel Morris2, Lydia Y. Chen3, Pascal Felber1, Robert Birke4, Valerio Schiavoni1
1University of Neuchâtel, 2The Ohio State University, 3TU Delft, 4ABB Research
Deep Learning
How many
neurons
should each
layer have?
How many
epochs
should it
run?
Which
learning rate
to define?
How many
layers to
use?
Hyperparameters Tuning
run
optimize()
Hyperparameters Model Parameters Score
n_layers = 3
n_neuros = 1024
learning_rate = 0.1
Weights
Optimization
60%
n_layers = 5
n_neuros = 512
learning_rate = 0.1
Weights
Optimization
75%
Hyperparameters Tuning
Hey, you going
to sleep?
Yes, now
shut up
What if you try 0.01 as
a learning rate?
Hyperparameters Autotuning
model dataset
hyper-
parameters
ranges
metric optimization
function
Hyperparameter Tuner
trained model
optimal hyper-
parameters
Google
Vizier
user
Auto-tuning: What is the problem?
Estimated Cost of
Tuning 6 Parameters
Cost[$]
0
22,5
45
67,5
90
EC2 Instances
m4.4xlarge m4.8xlarge m5.12xlarge m5.16xlarge m5.24xlarge
Tuning Time by
Number of Parameters
TuningTime
[hours]
0
1
2
3
4
Number of Parameters
1 2 3 4 5 6
The user can define only one objective function in the existing auto-tuning tools.
The chosen function is typically accuracy and the tuning performance is ignored.
Tuning duration grows exponentially with the number of parameters to be tuned.
Using more resources to improve the tuning performance is an expensive solution.
Auto-tuning: How to improve it?
1. Hyperparameters not only impact accuracy but also tuning duration and energy.
2. The optimal system parameters depend on the chosen hyperparameters.
Batch Size Impact
Difference[%]
-70
-60
-50
-40
-30
-20
-10
0
Batch Size
64 256 1024
Accuracy Duration Energy Cores Impact on Duration
DurationDifference[%]
-45
-30
-15
0
15
30
45
60
Number of Cores
2 4 8
Batch 64 Batch 256 Batch 1024
Baseline: batch size = 32. Baseline: number of cores = 1.
Hyperparameter Tuner
Hyper-parameter Tuner: Optimized
Evaluation: Setup
Baseline
Tune: Hyperparameter tuning only (i.e.,
no system parameter considered)
Workloads
Scenarios
Environment
I. Single Node (Intel E5-2620 with 8 cores)
Implemented on top of Keras and TensorFlow
II. Distributed Cluster (4x Intel E3-1275 with 8 cores)
Implemented on top of Spark using BigDL
I. Single-Tenancy
“Offline mode” showing results of running an
independent unseen HPT Job.
II. Multi-Tenancy
“Online mode” showing the averaged response
time of a synthetic trace with 90% load.
Evaluation Scenario I (Single-Node)
Model AccuracyAccuracy[%]
0
20
40
60
80
jacobi spkmeans bfs
Tune PipeTune
Tuning Duration
Time[s]
0
750
1500
2250
3000
jacobi spkmeans bfs
Tune PipeTune
Training Duration
Time[s]
0
10
20
30
40
jacobi spkmeans bfs
Tune PipeTune
Tuning Energy
Energy[kJ]
0
0,225
0,45
0,675
0,9
jacobi spkmeans bfs
Tune PipeTune
Evaluation Scenario II
Averaged Response Time
Time[s]
0
3000
6000
9000
12000
jacobi spkmeans bfs
Tune PipeTune
Single Node Distributed Cluster
Averaged Response Time
Time[s]
0
2375
4750
7125
9500
mnist new20 all
Tune PipeTune
Summary
• PipeTune is a novel approach for DNN tuning jobs;
• Leverages the combination of hyper with system parameter tuning to achieve high
model accuracy under low runtime and energy consumption;
• Experimental evaluation performed under various scenarios and using different state-
of-the-art workloads indicates promising results;
• Reduces the tuning time up to 23%;
• Speeds up the training time by up to 1.7x;
• Lowers energy consumption up to 20%;
• Refer to the paper for: more detailed evaluation and intermediate solution;
• Source code available in: https://github.com/isabellyrocha/pipetune.

More Related Content

What's hot

Cloud Computing
Cloud ComputingCloud Computing
Cloud Computingbutest
 
IT talk: Как я перестал бояться и полюбил TestNG
IT talk: Как я перестал бояться и полюбил TestNGIT talk: Как я перестал бояться и полюбил TestNG
IT talk: Как я перестал бояться и полюбил TestNGDataArt
 
"Эффективность и оптимизация кода в Java 8" Сергей Моренец
"Эффективность и оптимизация кода в Java 8" Сергей Моренец"Эффективность и оптимизация кода в Java 8" Сергей Моренец
"Эффективность и оптимизация кода в Java 8" Сергей МоренецFwdays
 
Introduction to Chainer
Introduction to ChainerIntroduction to Chainer
Introduction to ChainerShunta Saito
 
Multithreading to Construct Neural Networks
Multithreading to Construct Neural NetworksMultithreading to Construct Neural Networks
Multithreading to Construct Neural NetworksAltoros
 
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16MLconf
 
TensorFlow in Context
TensorFlow in ContextTensorFlow in Context
TensorFlow in ContextAltoros
 
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Lalit Jain
 
Applying your Convolutional Neural Networks
Applying your Convolutional Neural NetworksApplying your Convolutional Neural Networks
Applying your Convolutional Neural NetworksDatabricks
 
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016MLconf
 
Python library
Python libraryPython library
Python libraryToniyaP1
 
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...MLconf
 
Multithreading
MultithreadingMultithreading
MultithreadingF K
 
Anomaly Detection at Scale
Anomaly Detection at ScaleAnomaly Detection at Scale
Anomaly Detection at ScaleJeff Henrikson
 
GDG-Shanghai 2017 TensorFlow Summit Recap
GDG-Shanghai 2017 TensorFlow Summit RecapGDG-Shanghai 2017 TensorFlow Summit Recap
GDG-Shanghai 2017 TensorFlow Summit RecapJiang Jun
 
Threading Successes 05 Smoke
Threading Successes 05   SmokeThreading Successes 05   Smoke
Threading Successes 05 Smokeguest40fc7cd
 
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU Selection
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU SelectionMachine-Learning-based Performance Heuristics for Runtime CPU/GPU Selection
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU SelectionAkihiro Hayashi
 

What's hot (20)

Cloud Computing
Cloud ComputingCloud Computing
Cloud Computing
 
IT talk: Как я перестал бояться и полюбил TestNG
IT talk: Как я перестал бояться и полюбил TestNGIT talk: Как я перестал бояться и полюбил TestNG
IT talk: Как я перестал бояться и полюбил TestNG
 
"Эффективность и оптимизация кода в Java 8" Сергей Моренец
"Эффективность и оптимизация кода в Java 8" Сергей Моренец"Эффективность и оптимизация кода в Java 8" Сергей Моренец
"Эффективность и оптимизация кода в Java 8" Сергей Моренец
 
Introduction to Chainer
Introduction to ChainerIntroduction to Chainer
Introduction to Chainer
 
Multithreading to Construct Neural Networks
Multithreading to Construct Neural NetworksMultithreading to Construct Neural Networks
Multithreading to Construct Neural Networks
 
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16
Dr. Erin LeDell, Machine Learning Scientist, H2O.ai at MLconf SEA - 5/20/16
 
TensorFlow in Context
TensorFlow in ContextTensorFlow in Context
TensorFlow in Context
 
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
 
Applying your Convolutional Neural Networks
Applying your Convolutional Neural NetworksApplying your Convolutional Neural Networks
Applying your Convolutional Neural Networks
 
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016
Rajat Monga, Engineering Director, TensorFlow, Google at MLconf 2016
 
Python library
Python libraryPython library
Python library
 
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...
Le Song, Assistant Professor, College of Computing, Georgia Institute of Tech...
 
multithreading
multithreadingmultithreading
multithreading
 
Multithreading
MultithreadingMultithreading
Multithreading
 
Anomaly Detection at Scale
Anomaly Detection at ScaleAnomaly Detection at Scale
Anomaly Detection at Scale
 
GDG-Shanghai 2017 TensorFlow Summit Recap
GDG-Shanghai 2017 TensorFlow Summit RecapGDG-Shanghai 2017 TensorFlow Summit Recap
GDG-Shanghai 2017 TensorFlow Summit Recap
 
Java
JavaJava
Java
 
Storm
StormStorm
Storm
 
Threading Successes 05 Smoke
Threading Successes 05   SmokeThreading Successes 05   Smoke
Threading Successes 05 Smoke
 
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU Selection
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU SelectionMachine-Learning-based Performance Heuristics for Runtime CPU/GPU Selection
Machine-Learning-based Performance Heuristics for Runtime CPU/GPU Selection
 

Similar to PipeTune: Pipeline Parallelism of Hyper and System Parameters Tuning for Deep Learning Clusters

How to win data science competitions with Deep Learning
How to win data science competitions with Deep LearningHow to win data science competitions with Deep Learning
How to win data science competitions with Deep LearningSri Ambati
 
StackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkStackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkSri Ambati
 
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...Fisnik Kraja
 
H2O World - Top 10 Deep Learning Tips & Tricks - Arno Candel
H2O World - Top 10 Deep Learning Tips & Tricks - Arno CandelH2O World - Top 10 Deep Learning Tips & Tricks - Arno Candel
H2O World - Top 10 Deep Learning Tips & Tricks - Arno CandelSri Ambati
 
Separating Hype from Reality in Deep Learning with Sameer Farooqui
 Separating Hype from Reality in Deep Learning with Sameer Farooqui Separating Hype from Reality in Deep Learning with Sameer Farooqui
Separating Hype from Reality in Deep Learning with Sameer FarooquiDatabricks
 
Understand and Harness the Capabilities of Intel® Xeon Phi™ Processors
Understand and Harness the Capabilities of Intel® Xeon Phi™ ProcessorsUnderstand and Harness the Capabilities of Intel® Xeon Phi™ Processors
Understand and Harness the Capabilities of Intel® Xeon Phi™ ProcessorsIntel® Software
 
Seven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingSeven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingFan Robbin
 
Josh Patterson MLconf slides
Josh Patterson MLconf slidesJosh Patterson MLconf slides
Josh Patterson MLconf slidesMLconf
 
08 neural networks
08 neural networks08 neural networks
08 neural networksankit_ppt
 
Ann model and its application
Ann model and its applicationAnn model and its application
Ann model and its applicationmilan107
 
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...James Salter
 
Convolutional neural network
Convolutional neural networkConvolutional neural network
Convolutional neural networkItachi SK
 
Parallelism in a NumPy-based program
Parallelism in a NumPy-based programParallelism in a NumPy-based program
Parallelism in a NumPy-based programRalf Gommers
 
Python for Image Understanding: Deep Learning with Convolutional Neural Nets
Python for Image Understanding: Deep Learning with Convolutional Neural NetsPython for Image Understanding: Deep Learning with Convolutional Neural Nets
Python for Image Understanding: Deep Learning with Convolutional Neural NetsRoelof Pieters
 
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB Puppet
 
IRJET - Implementation of Neural Network on FPGA
IRJET - Implementation of Neural Network on FPGAIRJET - Implementation of Neural Network on FPGA
IRJET - Implementation of Neural Network on FPGAIRJET Journal
 
Making fitting in RooFit faster
Making fitting in RooFit fasterMaking fitting in RooFit faster
Making fitting in RooFit fasterPatrick Bos
 

Similar to PipeTune: Pipeline Parallelism of Hyper and System Parameters Tuning for Deep Learning Clusters (20)

How to win data science competitions with Deep Learning
How to win data science competitions with Deep LearningHow to win data science competitions with Deep Learning
How to win data science competitions with Deep Learning
 
Effective Benchmarks
Effective BenchmarksEffective Benchmarks
Effective Benchmarks
 
StackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkStackNet Meta-Modelling framework
StackNet Meta-Modelling framework
 
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...
Performance Analysis and Optimizations of CAE Applications (Case Study: STAR_...
 
H2O World - Top 10 Deep Learning Tips & Tricks - Arno Candel
H2O World - Top 10 Deep Learning Tips & Tricks - Arno CandelH2O World - Top 10 Deep Learning Tips & Tricks - Arno Candel
H2O World - Top 10 Deep Learning Tips & Tricks - Arno Candel
 
Deep learning
Deep learningDeep learning
Deep learning
 
Separating Hype from Reality in Deep Learning with Sameer Farooqui
 Separating Hype from Reality in Deep Learning with Sameer Farooqui Separating Hype from Reality in Deep Learning with Sameer Farooqui
Separating Hype from Reality in Deep Learning with Sameer Farooqui
 
Understand and Harness the Capabilities of Intel® Xeon Phi™ Processors
Understand and Harness the Capabilities of Intel® Xeon Phi™ ProcessorsUnderstand and Harness the Capabilities of Intel® Xeon Phi™ Processors
Understand and Harness the Capabilities of Intel® Xeon Phi™ Processors
 
Seven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingSeven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch Benchmarking
 
Josh Patterson MLconf slides
Josh Patterson MLconf slidesJosh Patterson MLconf slides
Josh Patterson MLconf slides
 
08 neural networks
08 neural networks08 neural networks
08 neural networks
 
Ann model and its application
Ann model and its applicationAnn model and its application
Ann model and its application
 
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...
An Efficient Reactive Model for Resource Discovery in DHT-Based Peer-to-Peer ...
 
Convolutional neural network
Convolutional neural networkConvolutional neural network
Convolutional neural network
 
Parallelism in a NumPy-based program
Parallelism in a NumPy-based programParallelism in a NumPy-based program
Parallelism in a NumPy-based program
 
Python for Image Understanding: Deep Learning with Convolutional Neural Nets
Python for Image Understanding: Deep Learning with Convolutional Neural NetsPython for Image Understanding: Deep Learning with Convolutional Neural Nets
Python for Image Understanding: Deep Learning with Convolutional Neural Nets
 
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB
Puppet Camp Melbourne 2014: Node Collaboration with PuppetDB
 
eam2
eam2eam2
eam2
 
IRJET - Implementation of Neural Network on FPGA
IRJET - Implementation of Neural Network on FPGAIRJET - Implementation of Neural Network on FPGA
IRJET - Implementation of Neural Network on FPGA
 
Making fitting in RooFit faster
Making fitting in RooFit fasterMaking fitting in RooFit faster
Making fitting in RooFit faster
 

More from LEGATO project

Scrooge Attack: Undervolting ARM Processors for Profit
Scrooge Attack: Undervolting ARM Processors for ProfitScrooge Attack: Undervolting ARM Processors for Profit
Scrooge Attack: Undervolting ARM Processors for ProfitLEGATO project
 
A practical approach for updating an integrity-enforced operating system
A practical approach for updating an integrity-enforced operating systemA practical approach for updating an integrity-enforced operating system
A practical approach for updating an integrity-enforced operating systemLEGATO project
 
TEEMon: A continuous performance monitoring framework for TEEs
TEEMon: A continuous performance monitoring framework for TEEsTEEMon: A continuous performance monitoring framework for TEEs
TEEMon: A continuous performance monitoring framework for TEEsLEGATO project
 
secureTF: A Secure TensorFlow Framework
secureTF: A Secure TensorFlow FrameworksecureTF: A Secure TensorFlow Framework
secureTF: A Secure TensorFlow FrameworkLEGATO project
 
LEGaTO: Machine Learning Use Case
LEGaTO: Machine Learning Use CaseLEGaTO: Machine Learning Use Case
LEGaTO: Machine Learning Use CaseLEGATO project
 
Smart Home AI at the edge
Smart Home AI at the edgeSmart Home AI at the edge
Smart Home AI at the edgeLEGATO project
 
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the project
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the projectLEGaTO: Low-Energy Heterogeneous Computing Use of AI in the project
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the projectLEGATO project
 
LEGaTO: Software Stack Programming Models
LEGaTO: Software Stack Programming ModelsLEGaTO: Software Stack Programming Models
LEGaTO: Software Stack Programming ModelsLEGATO project
 
LEGaTO: Software Stack Runtimes
LEGaTO: Software Stack RuntimesLEGaTO: Software Stack Runtimes
LEGaTO: Software Stack RuntimesLEGATO project
 
LEGaTO Heterogeneous Hardware
LEGaTO Heterogeneous HardwareLEGaTO Heterogeneous Hardware
LEGaTO Heterogeneous HardwareLEGATO project
 
LEGaTO: Low-Energy Heterogeneous Computing Workshop
LEGaTO: Low-Energy Heterogeneous Computing WorkshopLEGaTO: Low-Energy Heterogeneous Computing Workshop
LEGaTO: Low-Energy Heterogeneous Computing WorkshopLEGATO project
 
TZ4Fabric: Executing Smart Contracts with ARM TrustZone
TZ4Fabric: Executing Smart Contracts with ARM TrustZoneTZ4Fabric: Executing Smart Contracts with ARM TrustZone
TZ4Fabric: Executing Smart Contracts with ARM TrustZoneLEGATO project
 
Infection Research with Maxeler Dataflow Computing
Infection Research with Maxeler Dataflow ComputingInfection Research with Maxeler Dataflow Computing
Infection Research with Maxeler Dataflow ComputingLEGATO project
 
Smart Home - AI at the edge
Smart Home - AI at the edgeSmart Home - AI at the edge
Smart Home - AI at the edgeLEGATO project
 
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-Resiliency
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-ResiliencyFPGA Undervolting and Checkpointing for Energy-Efficiency and Error-Resiliency
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-ResiliencyLEGATO project
 
Device Data Directory and Asynchronous execution: A path to heterogeneous com...
Device Data Directory and Asynchronous execution: A path to heterogeneous com...Device Data Directory and Asynchronous execution: A path to heterogeneous com...
Device Data Directory and Asynchronous execution: A path to heterogeneous com...LEGATO project
 
Scheduling Task-parallel Applications in Dynamically Asymmetric Environments
Scheduling Task-parallel Applications in Dynamically Asymmetric EnvironmentsScheduling Task-parallel Applications in Dynamically Asymmetric Environments
Scheduling Task-parallel Applications in Dynamically Asymmetric EnvironmentsLEGATO project
 
RECS – Cloud to Edge Microserver Platform for Energy-Efficient Computing
RECS – Cloud to Edge Microserver Platform for Energy-Efficient ComputingRECS – Cloud to Edge Microserver Platform for Energy-Efficient Computing
RECS – Cloud to Edge Microserver Platform for Energy-Efficient ComputingLEGATO project
 

More from LEGATO project (20)

Scrooge Attack: Undervolting ARM Processors for Profit
Scrooge Attack: Undervolting ARM Processors for ProfitScrooge Attack: Undervolting ARM Processors for Profit
Scrooge Attack: Undervolting ARM Processors for Profit
 
A practical approach for updating an integrity-enforced operating system
A practical approach for updating an integrity-enforced operating systemA practical approach for updating an integrity-enforced operating system
A practical approach for updating an integrity-enforced operating system
 
TEEMon: A continuous performance monitoring framework for TEEs
TEEMon: A continuous performance monitoring framework for TEEsTEEMon: A continuous performance monitoring framework for TEEs
TEEMon: A continuous performance monitoring framework for TEEs
 
secureTF: A Secure TensorFlow Framework
secureTF: A Secure TensorFlow FrameworksecureTF: A Secure TensorFlow Framework
secureTF: A Secure TensorFlow Framework
 
LEGaTO: Machine Learning Use Case
LEGaTO: Machine Learning Use CaseLEGaTO: Machine Learning Use Case
LEGaTO: Machine Learning Use Case
 
Smart Home AI at the edge
Smart Home AI at the edgeSmart Home AI at the edge
Smart Home AI at the edge
 
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the project
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the projectLEGaTO: Low-Energy Heterogeneous Computing Use of AI in the project
LEGaTO: Low-Energy Heterogeneous Computing Use of AI in the project
 
LEGaTO Integration
LEGaTO IntegrationLEGaTO Integration
LEGaTO Integration
 
LEGaTO: Use cases
LEGaTO: Use casesLEGaTO: Use cases
LEGaTO: Use cases
 
LEGaTO: Software Stack Programming Models
LEGaTO: Software Stack Programming ModelsLEGaTO: Software Stack Programming Models
LEGaTO: Software Stack Programming Models
 
LEGaTO: Software Stack Runtimes
LEGaTO: Software Stack RuntimesLEGaTO: Software Stack Runtimes
LEGaTO: Software Stack Runtimes
 
LEGaTO Heterogeneous Hardware
LEGaTO Heterogeneous HardwareLEGaTO Heterogeneous Hardware
LEGaTO Heterogeneous Hardware
 
LEGaTO: Low-Energy Heterogeneous Computing Workshop
LEGaTO: Low-Energy Heterogeneous Computing WorkshopLEGaTO: Low-Energy Heterogeneous Computing Workshop
LEGaTO: Low-Energy Heterogeneous Computing Workshop
 
TZ4Fabric: Executing Smart Contracts with ARM TrustZone
TZ4Fabric: Executing Smart Contracts with ARM TrustZoneTZ4Fabric: Executing Smart Contracts with ARM TrustZone
TZ4Fabric: Executing Smart Contracts with ARM TrustZone
 
Infection Research with Maxeler Dataflow Computing
Infection Research with Maxeler Dataflow ComputingInfection Research with Maxeler Dataflow Computing
Infection Research with Maxeler Dataflow Computing
 
Smart Home - AI at the edge
Smart Home - AI at the edgeSmart Home - AI at the edge
Smart Home - AI at the edge
 
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-Resiliency
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-ResiliencyFPGA Undervolting and Checkpointing for Energy-Efficiency and Error-Resiliency
FPGA Undervolting and Checkpointing for Energy-Efficiency and Error-Resiliency
 
Device Data Directory and Asynchronous execution: A path to heterogeneous com...
Device Data Directory and Asynchronous execution: A path to heterogeneous com...Device Data Directory and Asynchronous execution: A path to heterogeneous com...
Device Data Directory and Asynchronous execution: A path to heterogeneous com...
 
Scheduling Task-parallel Applications in Dynamically Asymmetric Environments
Scheduling Task-parallel Applications in Dynamically Asymmetric EnvironmentsScheduling Task-parallel Applications in Dynamically Asymmetric Environments
Scheduling Task-parallel Applications in Dynamically Asymmetric Environments
 
RECS – Cloud to Edge Microserver Platform for Energy-Efficient Computing
RECS – Cloud to Edge Microserver Platform for Energy-Efficient ComputingRECS – Cloud to Edge Microserver Platform for Energy-Efficient Computing
RECS – Cloud to Edge Microserver Platform for Energy-Efficient Computing
 

Recently uploaded

User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationColumbia Weather Systems
 
Davis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologyDavis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologycaarthichand2003
 
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxTHE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxNandakishor Bhaurao Deshmukh
 
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingBase editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingNetHelix
 
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdf
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdfBUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdf
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdfWildaNurAmalia2
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPirithiRaju
 
Harmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms PresentationHarmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms Presentationtahreemzahra82
 
Transposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptTransposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptArshadWarsi13
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...Universidade Federal de Sergipe - UFS
 
The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxEran Akiva Sinbar
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptJoemSTuliba
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensorsonawaneprad
 
Bioteknologi kelas 10 kumer smapsa .pptx
Bioteknologi kelas 10 kumer smapsa .pptxBioteknologi kelas 10 kumer smapsa .pptx
Bioteknologi kelas 10 kumer smapsa .pptx023NiWayanAnggiSriWa
 
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPirithiRaju
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》rnrncn29
 
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfBehavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfSELF-EXPLANATORY
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)riyaescorts54
 
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In DubaiDubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubaikojalkojal131
 
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxGood agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxSimeonChristian
 

Recently uploaded (20)

Volatile Oils Pharmacognosy And Phytochemistry -I
Volatile Oils Pharmacognosy And Phytochemistry -IVolatile Oils Pharmacognosy And Phytochemistry -I
Volatile Oils Pharmacognosy And Phytochemistry -I
 
User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather Station
 
Davis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technologyDavis plaque method.pptx recombinant DNA technology
Davis plaque method.pptx recombinant DNA technology
 
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxTHE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
 
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editingBase editing, prime editing, Cas13 & RNA editing and organelle base editing
Base editing, prime editing, Cas13 & RNA editing and organelle base editing
 
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdf
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdfBUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdf
BUMI DAN ANTARIKSA PROJEK IPAS SMK KELAS X.pdf
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdf
 
Harmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms PresentationHarmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms Presentation
 
Transposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptTransposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.ppt
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
 
The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptx
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.ppt
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensor
 
Bioteknologi kelas 10 kumer smapsa .pptx
Bioteknologi kelas 10 kumer smapsa .pptxBioteknologi kelas 10 kumer smapsa .pptx
Bioteknologi kelas 10 kumer smapsa .pptx
 
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
 
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfBehavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
 
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In DubaiDubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
 
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxGood agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
 

PipeTune: Pipeline Parallelism of Hyper and System Parameters Tuning for Deep Learning Clusters

  • 1. 21st International Middleware Conference December 7 - 11, 2020, Delft, Netherlands PipeTune: Pipeline Parallelism of Hyper and System Parameters Tuning for Deep Learning Clusters Isabelly Rocha1, Nathaniel Morris2, Lydia Y. Chen3, Pascal Felber1, Robert Birke4, Valerio Schiavoni1 1University of Neuchâtel, 2The Ohio State University, 3TU Delft, 4ABB Research
  • 2. Deep Learning How many neurons should each layer have? How many epochs should it run? Which learning rate to define? How many layers to use?
  • 3. Hyperparameters Tuning run optimize() Hyperparameters Model Parameters Score n_layers = 3 n_neuros = 1024 learning_rate = 0.1 Weights Optimization 60% n_layers = 5 n_neuros = 512 learning_rate = 0.1 Weights Optimization 75%
  • 4. Hyperparameters Tuning Hey, you going to sleep? Yes, now shut up What if you try 0.01 as a learning rate?
  • 5. Hyperparameters Autotuning model dataset hyper- parameters ranges metric optimization function Hyperparameter Tuner trained model optimal hyper- parameters Google Vizier user
  • 6. Auto-tuning: What is the problem? Estimated Cost of Tuning 6 Parameters Cost[$] 0 22,5 45 67,5 90 EC2 Instances m4.4xlarge m4.8xlarge m5.12xlarge m5.16xlarge m5.24xlarge Tuning Time by Number of Parameters TuningTime [hours] 0 1 2 3 4 Number of Parameters 1 2 3 4 5 6 The user can define only one objective function in the existing auto-tuning tools. The chosen function is typically accuracy and the tuning performance is ignored. Tuning duration grows exponentially with the number of parameters to be tuned. Using more resources to improve the tuning performance is an expensive solution.
  • 7. Auto-tuning: How to improve it? 1. Hyperparameters not only impact accuracy but also tuning duration and energy. 2. The optimal system parameters depend on the chosen hyperparameters. Batch Size Impact Difference[%] -70 -60 -50 -40 -30 -20 -10 0 Batch Size 64 256 1024 Accuracy Duration Energy Cores Impact on Duration DurationDifference[%] -45 -30 -15 0 15 30 45 60 Number of Cores 2 4 8 Batch 64 Batch 256 Batch 1024 Baseline: batch size = 32. Baseline: number of cores = 1.
  • 10. Evaluation: Setup Baseline Tune: Hyperparameter tuning only (i.e., no system parameter considered) Workloads Scenarios Environment I. Single Node (Intel E5-2620 with 8 cores) Implemented on top of Keras and TensorFlow II. Distributed Cluster (4x Intel E3-1275 with 8 cores) Implemented on top of Spark using BigDL I. Single-Tenancy “Offline mode” showing results of running an independent unseen HPT Job. II. Multi-Tenancy “Online mode” showing the averaged response time of a synthetic trace with 90% load.
  • 11. Evaluation Scenario I (Single-Node) Model AccuracyAccuracy[%] 0 20 40 60 80 jacobi spkmeans bfs Tune PipeTune Tuning Duration Time[s] 0 750 1500 2250 3000 jacobi spkmeans bfs Tune PipeTune Training Duration Time[s] 0 10 20 30 40 jacobi spkmeans bfs Tune PipeTune Tuning Energy Energy[kJ] 0 0,225 0,45 0,675 0,9 jacobi spkmeans bfs Tune PipeTune
  • 12. Evaluation Scenario II Averaged Response Time Time[s] 0 3000 6000 9000 12000 jacobi spkmeans bfs Tune PipeTune Single Node Distributed Cluster Averaged Response Time Time[s] 0 2375 4750 7125 9500 mnist new20 all Tune PipeTune
  • 13. Summary • PipeTune is a novel approach for DNN tuning jobs; • Leverages the combination of hyper with system parameter tuning to achieve high model accuracy under low runtime and energy consumption; • Experimental evaluation performed under various scenarios and using different state- of-the-art workloads indicates promising results; • Reduces the tuning time up to 23%; • Speeds up the training time by up to 1.7x; • Lowers energy consumption up to 20%; • Refer to the paper for: more detailed evaluation and intermediate solution; • Source code available in: https://github.com/isabellyrocha/pipetune.