: A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO...
Transcript of : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO...
![Page 1: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/1.jpg)
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
21.07.2015:: :: Dr. Alexey CheptsovSEMAPRO 2015
Dr.-Ing. Alexey Cheptsov
: A new version of Supercomputing
…or life after the end of the Moore’s Law…
![Page 2: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/2.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 2
OUTLINE
� Semantic Web as a new HPC domain?
� DreamCloud project approach
� About us
� Mission Data-Centric Parallel Programming Models
� Convergence of Supercomputing into Big Data
![Page 3: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/3.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 3
Baden-Württemberg
Turnover: ~14 bil. €Staff: ~19.000
Turnover: ~1-2 bil. €Staff: ~8000
Turnover: ~100 bil. €
Staff: 260.100
Turnover: ~50 bil. €Staff: ~300.000
Turnover: ~2,5 bil. €Staff: ~11.00
Turnover: ~5 bil. €Staff: ~50.000
About HLRS1
![Page 4: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/4.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 4
• First Cray system in Europe
(Cray-2, 1986, 4 CPUs, 2GB RAM, approx. 2 GFLOPS)
• National HPC infrastructure provider since 1995
• EU infrastructure provider since 2005
• 110M core hours delivered to industry in 2014
- 4.000 nodes (24 cores)- 4 PFLOPs performance- 128 GB RAM per node- 7,8 PB Disc- 1512 KW power consumption / 1.5M Euro
HORNET (Cray XC40, Intel Haswell CPU, Aries network)
High Performance Computing Center Stuttgart
About HLRS1
![Page 5: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/5.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 5
TOP500
Total:
- USA – 233
- Japan - 39
- Germany – 37
- China - 37
Newcomers in
2015:
- USA – 34
- Germany – 12
- Japan - 11
About HLRS1
![Page 6: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/6.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 6
• Climate simulation on a high (a few kms) resolution
– University of Hohenheim
– 84.000 cores, 84 hours, 450 TB data
• Turbulent flow simulation for air- and gas
dynamic analysis
– RWTH Aachen University
– 92.000 cores, 110 hours, 80 TB data
XXL Application Portfolio
See more at:http://www.gauss-centre.eu/gauss-centre/EN/Projects/XXL_Projects_Hornet/XXL_Projects_Hornet.html?nn=1236240
About HLRS1
![Page 7: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/7.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 7
About HLRS1
Maya, the bee (みつばちマーヤの冒険)
• 115.000 pictures
• 2 hours each
• 200 pictures in
parallel
• From 9600 days
to 48 days
• 400.000 viewers
in 2 weeks
![Page 8: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/8.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 8
• Computation power
– Exascale is on his way, perhaps by 2020
– More performance for less money
– Hazelhehn - approx. 8 PTFLOPs / 3M Euro power costs, approx. 1 PB RAM
– Sustainable performance - 1 PTFLOP
• Storage
– Spinning devices will go away by 2020
– Flash is the core technology (aka NVM)
– Tapes will remain for data persistency
• Memory
– both high memory capacity and high memory bandwidth are required
– cannot guarantee same growth as for FLOPS
– extremely deep hierarchy (at the cache level)
– programming ease
– 3D stacked memory, NVM
Challenges for New -Generation Systems
About HLRS1
Source:
http://insidehpc.com/2014/05/nvm-will-shake-supercomputing/
![Page 9: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/9.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 9
OUTLINE
� Semantic Web as a new HPC domain?
� DreamCloud project approach
� About us
� Mission Data-Centric Parallel Programming Models
� Convergence of Supercomputing into Big Data
![Page 10: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/10.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 10
The modern HPC have to address a new class of computing-intensive applicationsfrom data-intensive domains in the internet, media, business, science, etc.
Convergence of Big Data into Supercomputing2
ArithmeticalComplexity
Volume
Structure
Data-IntensiveSciences
Traditional Computational Sciences
static dynamic
Evolution of
ComputationalApplications
Localityspatially and
temporally localno or little
fit into memorydo not fit
into memory
High Precision Arithmetic
variable precision or integer based
![Page 11: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/11.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 11Data
Facts
Info
rmati
on
Knowledge
Har
dwar
eH
ardw
are
e-S
ervi
ces
HPC
Infr
astr
uktu
res
App
licat
ions
Cloud Data Center
IntranetInternetSemantic WebLinkedData
Convergence of Big Data into Supercomputing2
![Page 12: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/12.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 12
OUTLINE
� Semantic Web as a new HPC domain?
� DreamCloud project approach
� About us
� Mission Data-Centric Parallel Programming Models
� Convergence of Supercomputing into Big Data
![Page 13: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/13.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 13
What are the challenges
� Infrastructure „on demand“
- distributed memory parallel clusters with low-latency intercon.
- multicore machines with shared memory
- GPGPU devices
- alltogether?
Semantic Web as a New HPC Domain ?3
![Page 14: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/14.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 14
What are the challenges
A Semantic Web Integration Platform for Large-Scale Reasoning
ReasonerReasoner
Query (SPARQL)
Answer(RDF)
IdentificationIdentificationSelectionSelection
Semantic Web as a New HPC Domain ?3
![Page 15: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/15.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 15
� The idea of LarKC
15
Development Platforms
Identifier Identifier
Selecter 1Selecter 1
ReasonerReasoner
DeciderDecider
Selecter 2Selecter 2
Query
Transformer
Query
Transformer
Flexibility, ModularityScalability, Performance
LarKC = an infrastructure for large scale, high performanceincomplete reasoning
Semantic Web as a New HPC Domain ?3
![Page 16: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/16.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 16
� LarKC architecture: High-Level Overview
Development Platforms
Pro
cess
1
Workflow branching
Workflow branching
Identifier Identifier
SelecterSelecter
ReasonerReasoner
DeciderDecider
SelecterSelecter
QueryQuery
Plug-in parallelisation
Multi-ThreadingMPI
Map-Reduce
Pro
cess
2
Pro
cess
4
Pro
cess
3
Semantic Web as a New HPC Domain ?3
![Page 17: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/17.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 17
� LarKC architecture: High-Level Overview
Development Platforms
Infrastructure
LarKC platform
StorageStorage ComputationComputation
High-performance
computer
Semantic Web Service Farm / Plug-in Marketplace
Semantic Web Service Farm / Plug-in Marketplace
DecidersDeciders IdentifiersIdentifiers
ReasonersReasonersTransformersTransformers
SelectorsSelectors
RDF data baseRDF data base
MonitoringMonitoring
DecidersDeciders IdentifiersIdentifiers SelectorsSelectors
TransformersTransformers ReasonersReasoners
Plug-indevelopers
Plug-inPlug-in Plug-inPlug-in
Plug-inPlug-in
…
Workflowdesigners
Applicationend-users
Plug-in Registry
Plug-in Managers
Workflow SupportSystem
Data Layer
(OWLIM)
Remote InvocationFramework
(GAT)
Monitoring ServiceEnd-points
Execution FrameworkExecution
Framework
Semantic Web as a New HPC Domain ?3
![Page 18: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/18.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 18
OUTLINE
� Semantic Web as a new HPC domain?
� DreamCloud project approach
� About us
� Mission Data-Centric Parallel Programming Models
� Convergence of Supercomputing into Big Data
![Page 19: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/19.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 19
State-Of-The-Art: Distributed Memory Processing over FS
� Programming models:
MapReduce
data-centric
poor sustainable performance
fault-tolerant
restrictive key/value model
Data-Centric Parallel Programming Models4
![Page 20: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/20.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 20
The Message-Passing Interface
� Data-driven scenarios with MPI
easy to integrate
high performance
Data-Centric Parallel Programming Models4
![Page 21: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/21.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 21
The Message-Passing Interface
� Issues
lack of implementations for the programming languages typically used in thedata-centric communities, such as Java
• Java implementations (MPJExpress)
• full MPI-2 standard implementation• issues with supporting new high-speed interconnects (e.g. Cray), related to the JVM• scalability to a peta/exaflop?• support of native tools for parallel computing, i.e. debuggers, error detectors,etc.
• native implementations (mpiJava)
• JNI is used for calling communicationlibraries that are available in nativecodes (i.e. highly optimized MPI comm.)• integration with a native MPI libraryis not easy• …but if you got it running, very enjoyable performance
Data-Centric Parallel Programming Models4
![Page 22: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/22.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 22
The Message-Passing Interface
� mpiJava
Open MPI MPICH Native (C) MPIimplementation
• Architecture
JNI C Interface
Java MPI bindings
Java wrappers
import mpi.*; Applications
Data-Centric Parallel Programming Models4
![Page 23: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/23.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 23
The Message-Passing Interface
�Java bindings for Open MPI (ompiJava)
in the Open MPI’strunk!
Data-Centric Parallel Programming Models4
![Page 24: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/24.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 24
Developments @ HLRS
� ompiJava performance
• P2P communication
Data-Centric Parallel Programming Models4
![Page 25: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/25.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 25
Application Scenarios
� Random Indexing for Large Texts
A Statistical Distribution technique for word/text similarity analysis
Terms
Docs
Occurancevector
Occurancevector
Similarityindex
Subsetting Query Expansion
Data-Centric Parallel Programming Models4
![Page 26: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/26.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 26
OUTLINE
� Semantic Web as a new HPC domain?
� DreamCloud project approach
� About us
� Mission Data-Centric Parallel Programming Models
� Convergence of Supercomputing into Big Data
![Page 27: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/27.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 27
DreamCloud Project Approach5
ccNUMA CPU
RAM
I/O
NUMA node NUMA node
• Problem statement
Infrastructure
Application
HPCResourceManager
• I/O bound
• Communication-bound
![Page 28: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/28.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 28
DreamCloud Project Approach5
• DreamCloud solution
– Scalable
– Low overhead
– Flexible
– Unified
Infrastructure
Application
Dynamic SchedulerDynamic
SchedulerPerformance Monitoring Framework
Performance Monitoring Framework
![Page 29: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/29.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 29
DreamCloud Project Approach5
• Available application profiling tools
– Valgrind – general analysis
– Likwid – energy/power consumption
– Vampir/Paraver – communication
– Wireshark – networking infrastructure
– …
• Infrastructure Monitoring Frameworks
– Zabbix
– Nagious
– Excess
– …
• Problem: A very high user‘s involvement
• Key solution: Consolidation and Integration!
![Page 30: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/30.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 30
DreamCloud Project Approach5
Total energy consumption
Total energy consumption
Total execution time
Total execution time
1 2
3 DiskDisk
4
Communication-intensive tasks
I/O-intensive tasks
5
Memory-intensive tasks
RAMRAM
High-levelview
Pattern-basedview
Detailedview
Legenda:communicationstreams
to be targetedby D3.2
• Consolidated View
![Page 31: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/31.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 31
DreamCloud Project Approach5
• Architecture
Monitoring Server (Elastic Search)Monitoring Server (Elastic Search)
PAPIPAPI RAPLRAPL 3rd-partytools
3rd-partytools
. . .
EXCESS monitoring agentEXCESS monitoring agent
Node 1 Node N
. . .
Querying Interface (RESTful)Querying Interface (RESTful)
![Page 32: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/32.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 32
DreamCloud Project Approach5
• Performance Profiles
– Static (prior to the execution)
– Dynamic (collected at runtime)
• Performance Counters (PAPI etc.)
• Energy Counters (RAPL etc.)
• User-defined ones
(e.g. progress tracker)
![Page 33: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/33.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 33
DreamCloud Project Approach5
• Basic Metrics
![Page 34: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/34.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 34
DreamCloud Project Approach5
• Power consumption
![Page 35: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/35.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 35
DreamCloud Project Approach5
• Performance Profiles Representation in DreamCloud
• Visualization
Performance Profiles (Example in CSV)
timestamp;node;task;PAPI_TOT_INS:CPU0;PAPI_TOT_INS:CPU1
1418120193.4880380;node_1[perf];task_1;516680185;663789
1418120194.4996216;node_2[perf];task_1;663789;663714
1418120196.5121722;node_3[balanced];task_1;3805840;663405
1418120193.4880380;node_3[energy];task_2_1;516680185;623687
1418120194.4996216;node_9[perf];task_2_1;663789;655418
1418120196.5121722;node_3[perf];task_2_2;3805840;663789
1418120194.4996216;node_9[perf];task_2_2;663789;699712
![Page 36: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/36.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 36
DreamCloud Project Approach5
• Molecular Dynamics (MD) Simulation Code – MS2
– Simulation of movement of atoms and molecules
– Massively parallel, on a per-particle basis
MS2
AdaptationPhase
AdaptationPhase
GenerationPhase
GenerationPhase
AdaptState
CheckTermination
Generate Individuals
CalculateFitness Values
EvaluationPhase
EvaluationPhase
Generate Input Data Sets
MS2
RankIndividuals
MS2
MS2
Compute Intensive
(MPI)
PersistentTask
Exit
![Page 37: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/37.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 37
DreamCloud Project Approach5
Workflow Manager
Node Node
jdl
Subm
issionInterface
Sch
edul
ing
Inte
rfac
e
SubmissionRequest
Deployment
Plan
Scheduling Advisor
Scheduling
Request
Tasks DescriptionWorkflow
DescriptionOptimization
Criteria
Application
RTE
HeuristicsModule
Node Node
Infrastructure
Resource Manager
MonitoringFramework
Node Node
Node Node
Node
Node
Deployment Progress Tracker
1
2
4
5 6
7
98 10
3
Performance Profiles
![Page 38: : A new version of Supercomputing …or life after the end of the … · 2016. 9. 11. · SEMAPRO 2015:: ::21.07.2015 Dr. Alexey Cheptsov Dr.-Ing. Alexey Cheptsov: A new version of](https://reader034.fdocuments.us/reader034/viewer/2022052022/6036876ba660ea25266bac20/html5/thumbnails/38.jpg)
21.07.2015::
::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: ::::: :::::
::
DreamCloud: A new version of Supercomputing
Dr. Alexey CheptsovSEMAPRO 2015 38
Main Results
� HPC is going to face new challenges related to data-centric
application expansion.
� Parallel programming models (mainly MapReduce and MPI) are
the key enablers of HPC to data-centric applications
� Reaching near-peak performance is going to be the major
challenge
Future Work
� Promote existing technologies, such as MPI, to solving new
challenges, such as Big Data.
� Making existing framework more data-centric.
Conclusion6