Challenges in Developing Highly Reliable HPC systems

Size: px
Start display at page:

Download "Challenges in Developing Highly Reliable HPC systems"

Transcription

1 Dec. 1, 2012 JS International Symopsium on DVLSI Systems 2012 hallenges in Developing Highly Reliable HP systems Koichiro akayama Fujitsu Limited

2 K computer Developed jointly by RIKEN and Fujitsu First computer to achieve 10PFlops 88,128 SPAR64 M VIIIfx PUs 6-dimensional mesh/torus interconnect ofu Ran for 29.5 hours continuously for LINPAK benchmark Equivalent to MBF > 296 node-years System Board Rack K computer PU 4 = = Interconnect + IO-node 6 = 1

3 Outline of the talk Dependability of SPAR64VIIIfx (PU), Interconnect, Board/Rack, System Summary oward exa-scale computer Attributes of dependability Availability Reliability Maintainability onfiguration PU Interconnect Board/Rack System opics in this area will be presented. Safety Integrity 2

4 DDR3 interface Availability Instruction-set extension (HP-AE) Virtual single processor (VISIMPA) Reliability Soft/hard-error resiliency HSIO Maintainability Error reporting ore5 L2$ Data ore7 SPAR64 VIIIfx 8 cores 6MB shared L2$ lock 2GHz DDR3 interface ore4 MA ore1 ore0 L2$ ontrol L2$ Data ore6 MA ore3 ore2 3

5 HP-AE and VISIMPA HP-AE Extended instruction set to accelerate scientific calculation VISIMPA Efficient multi-threading on multi-core PU VISIMPA ore Arithmetic unit Register PU Hardware Barrier Synch. ore Arithmetic unit Register ore Arithmetic SIMD unit Floating-point Register register ext. HP-AE Shared L2 cache Sector cache Memory 4

6 HP-AE: Sector ache Software-controlled virtual local memory Application program can direct data stored to sector 1 by using compiler directives. Reused data Use sector 1 Others Use sector 0 Data in sector1 cannot be replaced by other data Before Other data Sector0 Array A Sector 0 A1 B1 A2 L2 cache Array B Sector 0 Frequently reused data Sector1 Array Sector 1 Sector 0 Sector 1 B2 1 2 ache miss on B3 access After B3 A1 B1 B2 1 2 A2 Sector 0 is replaceable Not replace 5

7 VISIMPA Virtual Single Processor by Integrated Multi-core Parallel Architecture A mechanism to reduce the number of processes by handling a multicore PU as 1 Processor for highly efficient parallel execution. Hardware barrier: 10 times faster than Soft Barrier Shared L2 ache: prevent false sharing of data between cores ompiler technique: flat MPI model to hybrid model ore PU ore ore PU ore Process Process Inter-core multi-threaded process L2$ L2$ L2$ Memory VISIMPA Memory 6

8 VISIMPA: Hybrid parallel model Inter-PU parallelization: process parallel execution Intra-PU parallelization: thread parallel execution Flat-MPI Model Hybrid Model Interconnect Interconnect PU PU PU PU P P P P P P P P P P P: process, : thread, : core 7

9 Soft/Hard-Error Resiliency Failure Rate Early failure hree levels of production test: chip test, board test, and rack test Early failure Random failure Wear-out failure Random failure Process shrinking and low voltage increase soft errors. Protection methods: E, redundancy, hardware-retry, etc. Water cooling reduces hard-error and delays ware-out failure. ime 8

10 Highly Reliable PU Data change due to collision with cosmic radiation (ex. 0 1) SPAR64 M VIIIfx Wrong result SPAR64 M VIIIfx can detect errors at the hardware level and self-recover from it. Error protection methods: Arithmetic Unit Register ache Parity, Residue Instruction pipeline retry Parity E E Parity+redundancy 1-bit error correctable 1-bit error detectable (Stop the System) 1-bit error harmless 9

11 Production est haracteristics Detected Undetected hip est Short-time Self-loopback onsistent Errors Defective HSIO Random failure Marginal HSIO Board est A JAG (IEEE ) Bad Solder Joints Ground open failure Imperfect shorts Rack est Inter-chip communication test Random failuer Marginal HSIO Ground open failure Imperfect shorts 10

12 Availability Flexible resource mapping Reliability Fault-tolerant routing Maintainability Y+ X+ B+ X- ofu Interconnect 6 dimensional mesh/torus network Y- Interonnect ontroller (I) B- Z+ Z- A AB 1 ofu-unit = 12nodes SPAR64 VIIIfx XYZ 11

13 Failure Bypass by Extended Dimension Order Default dimension order routing X Y Z A B Extended dimension order routing B A X Y Z A B Adding the first B--A order, we have 12 routes between two nodes. Failure 12

14 Flexible Resource Allocation Interconnect structure aware resource allocation is to suppress mutual interference between jobs Allocation unit is ofu-unit (12 nodes) Allocation area is in cuboid shape of the network Rotate the job allocation area and fit it in the system empty space. Job D Job B Job Job A 13

15 Availability Reliability Water cooling Maintainability ooling plate for I ooling plate for PU Liquid coupler Power I PU/I System Board 14

16 Water ooling for PU, I, and power I Life-time (relative) Lowering the semiconductor junction temperature extends the life-time of components. Arrhenius law: Lowering the junction temperature from 85 to 30 gives 60 to 100 times longer life-time. PU, I, Power(POL)I is water cooled Secure system reliability Also reduces leakage current of LSIs. Arrhenius law 1.0E E E+02 L=A exp(ea/k B ) L : Life-time A : onstant Ea: Energy k B : Boltzmann constant : Absolute temp. 1.0E E E E For Ea = 0.7~0.8 x61 to 85 Junction temp.( ) 15

17 Availability Hot swapping Reliability System-level redundancy for I/O network and system control Maintainability 16

18 Redundant Structure and Hot Swap Redundant structure By the n+1 redundant structure, even in case of component failure, operation can continue without any functional degradation. Hot swap Without turning off the rack, defective components can be replaced. he impact of maintenance is localized and thus system availability is preserved. 17

19 Redundant Structure and Hot Swap omponent Redundant structure Hot swap Rack power source Yes Yes ooling Fan Yes Yes Service Processor(SP) Yes (Duplicated) Yes System Board(SB/IOSB) No Bypass routing on B- axis port PU/I on SB/IOSB No Water cooled, error resiliency Yes (SB hot swap) POL on SB/IOSB No Water cooled (SB hot swap) Other power sources on SB/IOSB Yes (SB hot swap) DIMM on SB/IOSB No E (SB hot swap) System volume RAID ontroller, power source are redundant HDD is in RAID Yes(Module) 18

20 Redundant I/O Path Redundant I/O path from IOSB to RAID for the system availability Rack-0 Rack-1 19

21 System-Level Redundancy Frontend server OFU IO Network IO Node Local Disk IB SW Server Global Disk Redundant management network Job management node ontrol node System Integration node Redundant ontrol Network Maintenance servers 20

22 System-Level Redundancy Linux based OS High performance file system management OFU IO Network High performance file system management ompiler Parallel language ools/library Virtual tools Frontend server IO Node Local Disk System operation and job management Job management node System Management ontrol node IB SW Server Global Disk System management ofu partitioning management Redundant management network System Integration node Hard maintenance Redundant ontrol Network Maintenance servers 21

23 System-Level Redundancy Redundant parts are shown in red. All operation related components are redundant! Frontend server OFU IO Network IO Node Local Disk IB SW Server Global Disk Redundant management network Job management node ontrol node System Integration node Redundant ontrol Network Maintenance servers 22

24 PU Interconnect Board / Rack System A high performance, low power, multicore, scalar processor which can realize a large scale system while preserving reliability A high speed, high reliability, low latency, low power interconnect (inter PU network) Water cooling enabling both longer life-time and low power consumption A compiler able to bring out the full potential of the processor by parallelizing hundreds of thousands of processes. Management middleware for high availability over a system operating hundreds of thousands of nodes 23

25 oward Exa-Scale omputer Power gap In 2011~2012 the power efficiency trend changed. An advanced low-power technique will be required. Power efficiency 10GF/W Improve power efficiency by 60 Exa-scale supercomputer (2018~20) 1EF, 20-30MW hange in the trend 1GF/W 1000 GF: GigaFlops PF: PetaFlops EF: ExaFLops K computer (2011) 10PF, 12.7MW E+14 1.E+15 1PF 1.E+16 10PF 1.E PF 1.E+18 1EF Performance Power efficiency of the successive supercomputers 24

26 oward Exa-Scale omputer In order to achieve the requirement of power, performance, and reliability, system-wide co-design will be needed. Hardware architecture [PU, Network, ooling, Memory, storage, servers] Software architecture [application, OS, driver, compiler] System-wide co-design for dependability 25

27 26

Fujitsu s Approach to Application Centric Petascale Computing

Fujitsu s Approach to Application Centric Petascale Computing Fujitsu s Approach to Application Centric Petascale Computing 2 nd Nov. 2010 Motoi Okuda Fujitsu Ltd. Agenda Japanese Next-Generation Supercomputer, K Computer Project Overview Design Targets System Overview

More information

White paper Advanced Technologies of the Supercomputer PRIMEHPC FX10

White paper Advanced Technologies of the Supercomputer PRIMEHPC FX10 White paper Advanced Technologies of the Supercomputer PRIMEHPC FX10 Next Generation Technical Computing Unit Fujitsu Limited Contents Overview of the PRIMEHPC FX10 Supercomputer 2 SPARC64 TM IXfx: Fujitsu-Developed

More information

The way toward peta-flops

The way toward peta-flops The way toward peta-flops ISC-2011 Dr. Pierre Lagier Chief Technology Officer Fujitsu Systems Europe Where things started from DESIGN CONCEPTS 2 New challenges and requirements! Optimal sustained flops

More information

Current Status of the Next- Generation Supercomputer in Japan. YOKOKAWA, Mitsuo Next-Generation Supercomputer R&D Center RIKEN

Current Status of the Next- Generation Supercomputer in Japan. YOKOKAWA, Mitsuo Next-Generation Supercomputer R&D Center RIKEN Current Status of the Next- Generation Supercomputer in Japan YOKOKAWA, Mitsuo Next-Generation Supercomputer R&D Center RIKEN International Workshop on Peta-Scale Computing Programming Environment, Languages

More information

Advanced Software for the Supercomputer PRIMEHPC FX10. Copyright 2011 FUJITSU LIMITED

Advanced Software for the Supercomputer PRIMEHPC FX10. Copyright 2011 FUJITSU LIMITED Advanced Software for the Supercomputer PRIMEHPC FX10 System Configuration of PRIMEHPC FX10 nodes Login Compilation Job submission 6D mesh/torus Interconnect Local file system (Temporary area occupied

More information

Compiler Technology That Demonstrates Ability of the K computer

Compiler Technology That Demonstrates Ability of the K computer ompiler echnology hat Demonstrates Ability of the K computer Koutarou aki Manabu Matsuyama Hitoshi Murai Kazuo Minami We developed SAR64 VIIIfx, a new U for constructing a huge computing system on a scale

More information

Fujitsu Petascale Supercomputer PRIMEHPC FX10. 4x2 racks (768 compute nodes) configuration. Copyright 2011 FUJITSU LIMITED

Fujitsu Petascale Supercomputer PRIMEHPC FX10. 4x2 racks (768 compute nodes) configuration. Copyright 2011 FUJITSU LIMITED Fujitsu Petascale Supercomputer PRIMEHPC FX10 4x2 racks (768 compute nodes) configuration PRIMEHPC FX10 Highlights Scales up to 23.2 PFLOPS Improves Fujitsu s supercomputer technology employed in the FX1

More information

Blue Gene/Q. Hardware Overview Michael Stephan. Mitglied der Helmholtz-Gemeinschaft

Blue Gene/Q. Hardware Overview Michael Stephan. Mitglied der Helmholtz-Gemeinschaft Blue Gene/Q Hardware Overview 02.02.2015 Michael Stephan Blue Gene/Q: Design goals System-on-Chip (SoC) design Processor comprises both processing cores and network Optimal performance / watt ratio Small

More information

Technical Computing Suite supporting the hybrid system

Technical Computing Suite supporting the hybrid system Technical Computing Suite supporting the hybrid system Supercomputer PRIMEHPC FX10 PRIMERGY x86 cluster Hybrid System Configuration Supercomputer PRIMEHPC FX10 PRIMERGY x86 cluster 6D mesh/torus Interconnect

More information

White paper FUJITSU Supercomputer PRIMEHPC FX100 Evolution to the Next Generation

White paper FUJITSU Supercomputer PRIMEHPC FX100 Evolution to the Next Generation White paper FUJITSU Supercomputer PRIMEHPC FX100 Evolution to the Next Generation Next Generation Technical Computing Unit Fujitsu Limited Contents FUJITSU Supercomputer PRIMEHPC FX100 System Overview

More information

Carlo Cavazzoni, HPC department, CINECA

Carlo Cavazzoni, HPC department, CINECA Introduction to Shared memory architectures Carlo Cavazzoni, HPC department, CINECA Modern Parallel Architectures Two basic architectural scheme: Distributed Memory Shared Memory Now most computers have

More information

PRIMEHPC FX10: Advanced Software

PRIMEHPC FX10: Advanced Software PRIMEHPC FX10: Advanced Software Koh Hotta Fujitsu Limited System Software supports --- Stable/Robust & Low Overhead Execution of Large Scale Programs Operating System File System Program Development for

More information

The Future of SPARC64 TM

The Future of SPARC64 TM The Future of SPARC64 TM Kevin Oltendorf VP, R&D Fujitsu Management Services of America 0 Raffle at the Conclusion Your benefits of Grandstand Suite Seats Free Beer, Wine and Fried snacks Early access

More information

Fujitsu s Technologies to the K Computer

Fujitsu s Technologies to the K Computer Fujitsu s Technologies to the K Computer - a journey to practical Petascale computing platform - June 21 nd, 2011 Motoi Okuda FUJITSU Ltd. Agenda The Next generation supercomputer project of Japan The

More information

Lecture 9: MIMD Architectures

Lecture 9: MIMD Architectures Lecture 9: MIMD Architectures Introduction and classification Symmetric multiprocessors NUMA architecture Clusters Zebo Peng, IDA, LiTH 1 Introduction MIMD: a set of general purpose processors is connected

More information

Virtual Memory. Reading. Sections 5.4, 5.5, 5.6, 5.8, 5.10 (2) Lecture notes from MKP and S. Yalamanchili

Virtual Memory. Reading. Sections 5.4, 5.5, 5.6, 5.8, 5.10 (2) Lecture notes from MKP and S. Yalamanchili Virtual Memory Lecture notes from MKP and S. Yalamanchili Sections 5.4, 5.5, 5.6, 5.8, 5.10 Reading (2) 1 The Memory Hierarchy ALU registers Cache Memory Memory Memory Managed by the compiler Memory Managed

More information

An Overview of Fujitsu s Lustre Based File System

An Overview of Fujitsu s Lustre Based File System An Overview of Fujitsu s Lustre Based File System Shinji Sumimoto Fujitsu Limited Apr.12 2011 For Maximizing CPU Utilization by Minimizing File IO Overhead Outline Target System Overview Goals of Fujitsu

More information

Multi-core Programming Evolution

Multi-core Programming Evolution Multi-core Programming Evolution Based on slides from Intel Software ollege and Multi-ore Programming increasing performance through software multi-threading by Shameem Akhter and Jason Roberts, Evolution

More information

SPARC64 X: Fujitsu s New Generation 16 core Processor for UNIX Server

SPARC64 X: Fujitsu s New Generation 16 core Processor for UNIX Server SPARC64 X: Fujitsu s New Generation 16 core Processor for UNIX Server 19 th April 2013 Toshio Yoshida Processor Development Division Enterprise Server Business Unit Fujitsu Limited SPARC64 TM SPARC64 TM

More information

Fujitsu HPC Roadmap Beyond Petascale Computing. Toshiyuki Shimizu Fujitsu Limited

Fujitsu HPC Roadmap Beyond Petascale Computing. Toshiyuki Shimizu Fujitsu Limited Fujitsu HPC Roadmap Beyond Petascale Computing Toshiyuki Shimizu Fujitsu Limited Outline Mission and HPC product portfolio K computer*, Fujitsu PRIMEHPC, and the future K computer and PRIMEHPC FX10 Post-FX10,

More information

Introduction to the K computer

Introduction to the K computer Introduction to the K computer Fumiyoshi Shoji Deputy Director Operations and Computer Technologies Div. Advanced Institute for Computational Science RIKEN Outline ü Overview of the K

More information

Fujitsu s new supercomputer, delivering the next step in Exascale capability

Fujitsu s new supercomputer, delivering the next step in Exascale capability Fujitsu s new supercomputer, delivering the next step in Exascale capability Toshiyuki Shimizu November 19th, 2014 0 Past, PRIMEHPC FX100, and roadmap for Exascale 2011 2012 2013 2014 2015 2016 2017 2018

More information

NAS System. User s Manual. Revision 1.0

NAS System. User s Manual. Revision 1.0 User s Manual Revision 1.0 Before You Begin efore going through with this manual, you should read and focus on the following safety guidelines. Information about the NAS system s packaging and delivery

More information

Memory technology and optimizations ( 2.3) Main Memory

Memory technology and optimizations ( 2.3) Main Memory Memory technology and optimizations ( 2.3) 47 Main Memory Performance of Main Memory: Latency: affects Cache Miss Penalty» Access Time: time between request and word arrival» Cycle Time: minimum time between

More information

Post-K Supercomputer Overview. Copyright 2016 FUJITSU LIMITED

Post-K Supercomputer Overview. Copyright 2016 FUJITSU LIMITED Post-K Supercomputer Overview 1 Post-K supercomputer overview Developing Post-K as the successor to the K computer with RIKEN Developing HPC-optimized high performance CPU and system software Selected

More information

The End of Redundancy. Alan Wood Sun Microsystems May 8, 2009

The End of Redundancy. Alan Wood Sun Microsystems May 8, 2009 The End of Redundancy Alan Wood Sun Microsystems May 8, 2009 Growing Demand, Shrinking Resources By 2008, 50% of current data centers will have insufficient power and cooling capacity to meet the demands

More information

Overview of Tianhe-2

Overview of Tianhe-2 Overview of Tianhe-2 (MilkyWay-2) Supercomputer Yutong Lu School of Computer Science, National University of Defense Technology; State Key Laboratory of High Performance Computing, China ytlu@nudt.edu.cn

More information

Introduction of Fujitsu s next-generation supercomputer

Introduction of Fujitsu s next-generation supercomputer Introduction of Fujitsu s next-generation supercomputer MATSUMOTO Takayuki July 16, 2014 HPC Platform Solutions Fujitsu has a long history of supercomputing over 30 years Technologies and experience of

More information

SPARC64 X: Fujitsu s New Generation 16 Core Processor for the next generation UNIX servers

SPARC64 X: Fujitsu s New Generation 16 Core Processor for the next generation UNIX servers X: Fujitsu s New Generation 16 Processor for the next generation UNIX servers August 29, 2012 Takumi Maruyama Processor Development Division Enterprise Server Business Unit Fujitsu Limited All Rights Reserved,Copyright

More information

Japan s post K Computer Yutaka Ishikawa Project Leader RIKEN AICS

Japan s post K Computer Yutaka Ishikawa Project Leader RIKEN AICS Japan s post K Computer Yutaka Ishikawa Project Leader RIKEN AICS HPC User Forum, 7 th September, 2016 Outline of Talk Introduction of FLAGSHIP2020 project An Overview of post K system Concluding Remarks

More information

Energy and Thermal Aware Buffer Cache Replacement Algorithm

Energy and Thermal Aware Buffer Cache Replacement Algorithm MSST 2010 nergy and Thermal Aware uffer ache Replacement Algorithm ianhui Yue, Yifeng Zhu, Zhao ai, and Lin Lin lectrical and omputer ngineering University of Maine Research Summary Memory power consumption

More information

Lecture 9: MIMD Architectures

Lecture 9: MIMD Architectures Lecture 9: MIMD Architectures Introduction and classification Symmetric multiprocessors NUMA architecture Clusters Zebo Peng, IDA, LiTH 1 Introduction A set of general purpose processors is connected together.

More information

Brand-New Vector Supercomputer

Brand-New Vector Supercomputer Brand-New Vector Supercomputer NEC Corporation IT Platform Division Shintaro MOMOSE SC13 1 New Product NEC Released A Brand-New Vector Supercomputer, SX-ACE Just Now. Vector Supercomputer for Memory Bandwidth

More information

User Training Cray XC40 IITM, Pune

User Training Cray XC40 IITM, Pune User Training Cray XC40 IITM, Pune Sudhakar Yerneni, Raviteja K, Nachiket Manapragada, etc. 1 Cray XC40 Architecture & Packaging 3 Cray XC Series Building Blocks XC40 System Compute Blade 4 Compute Nodes

More information

Introduction CPS343. Spring Parallel and High Performance Computing. CPS343 (Parallel and HPC) Introduction Spring / 29

Introduction CPS343. Spring Parallel and High Performance Computing. CPS343 (Parallel and HPC) Introduction Spring / 29 Introduction CPS343 Parallel and High Performance Computing Spring 2018 CPS343 (Parallel and HPC) Introduction Spring 2018 1 / 29 Outline 1 Preface Course Details Course Requirements 2 Background Definitions

More information

Experiences of the Development of the Supercomputers

Experiences of the Development of the Supercomputers Experiences of the Development of the Supercomputers - Earth Simulator and K Computer YOKOKAWA, Mitsuo Kobe University/RIKEN AICS Application Oriented Systems Developed in Japan No.1 systems in TOP500

More information

Programming for Fujitsu Supercomputers

Programming for Fujitsu Supercomputers Programming for Fujitsu Supercomputers Koh Hotta The Next Generation Technical Computing Fujitsu Limited To Programmers who are busy on their own research, Fujitsu provides environments for Parallel Programming

More information

VX3000-E Unified Network Storage

VX3000-E Unified Network Storage Datasheet VX3000-E Unified Network Storage Overview VX3000-E storage, with high performance, high reliability, high available, high density, high scalability and high usability, is a new-generation unified

More information

Non-uniform memory access machine or (NUMA) is a system where the memory access time to any region of memory is not the same for all processors.

Non-uniform memory access machine or (NUMA) is a system where the memory access time to any region of memory is not the same for all processors. CS 320 Ch. 17 Parallel Processing Multiple Processor Organization The author makes the statement: "Processors execute programs by executing machine instructions in a sequence one at a time." He also says

More information

Exascale: Parallelism gone wild!

Exascale: Parallelism gone wild! IPDPS TCPP meeting, April 2010 Exascale: Parallelism gone wild! Craig Stunkel, Outline Why are we talking about Exascale? Why will it be fundamentally different? How will we attack the challenges? In particular,

More information

Getting the best performance from massively parallel computer

Getting the best performance from massively parallel computer Getting the best performance from massively parallel computer June 6 th, 2013 Takashi Aoki Next Generation Technical Computing Unit Fujitsu Limited Agenda Second generation petascale supercomputer PRIMEHPC

More information

Storage. Hwansoo Han

Storage. Hwansoo Han Storage Hwansoo Han I/O Devices I/O devices can be characterized by Behavior: input, out, storage Partner: human or machine Data rate: bytes/sec, transfers/sec I/O bus connections 2 I/O System Characteristics

More information

The Road from Peta to ExaFlop

The Road from Peta to ExaFlop The Road from Peta to ExaFlop Andreas Bechtolsheim June 23, 2009 HPC Driving the Computer Business Server Unit Mix (IDC 2008) Enterprise HPC Web 100 75 50 25 0 2003 2008 2013 HPC grew from 13% of units

More information

Lecture 7: PCM, Cache coherence. Topics: handling PCM errors and writes, cache coherence intro

Lecture 7: PCM, Cache coherence. Topics: handling PCM errors and writes, cache coherence intro Lecture 7: M, ache coherence Topics: handling M errors and writes, cache coherence intro 1 hase hange Memory Emerging NVM technology that can replace Flash and DRAM Much higher density; much better scalability;

More information

Fujitsu M10/ SPARC M10 Systems. Quick Guide

Fujitsu M10/ SPARC M10 Systems. Quick Guide Fujitsu M10/ SPARC M10 Systems Quick Guide Manual Code: C120-E677-12EN June 2016 Preface This document describes the basic specifications and system configurations that users need to be familiar with when

More information

OpenFOAM Performance Testing and Profiling. October 2017

OpenFOAM Performance Testing and Profiling. October 2017 OpenFOAM Performance Testing and Profiling October 2017 Note The following research was performed under the HPC Advisory Council activities Participating vendors: Huawei, Mellanox Compute resource - HPC

More information

Overview of the SPARC Enterprise Servers

Overview of the SPARC Enterprise Servers Overview of the SPARC Enterprise Servers SPARC Enterprise Technologies for the Datacenter Ideal for Enterprise Application Deployments System Overview Virtualization technologies > Maximize system utilization

More information

Supercomputers. Alex Reid & James O'Donoghue

Supercomputers. Alex Reid & James O'Donoghue Supercomputers Alex Reid & James O'Donoghue The Need for Supercomputers Supercomputers allow large amounts of processing to be dedicated to calculation-heavy problems Supercomputers are centralized in

More information

CS370 Operating Systems

CS370 Operating Systems CS370 Operating Systems Colorado State University Yashwant K Malaiya Spring 2018 Lecture 24 Mass Storage, HDFS/Hadoop Slides based on Text by Silberschatz, Galvin, Gagne Various sources 1 1 FAQ What 2

More information

White paper PRIMEQUEST 2800E and 2400E Enterprise Server What s Inside The High Reliability Platform

White paper PRIMEQUEST 2800E and 2400E Enterprise Server What s Inside The High Reliability Platform White paper PRIMEQUEST 2800E What s Inside the High Reliability Platform White paper PRIMEQUEST 2800E and 2400E Enterprise Server What s Inside The High Reliability Platform Business continuity and high

More information

NEC Express5800/1000 Series

NEC Express5800/1000 Series NEC Enterprise Server NEC Express5800/1000 Series NEC Express5800/1000 Technology Guide Vol.1 Powered by the Dual-Core Intel Itanium NEC Express5800/1000 Series Reliability and Performance through the

More information

VX1800 Series Unified Network Storage

VX1800 Series Unified Network Storage Datasheet VX1800 Series Unified Network Storage Overview VX1800 series storage, with high performance, high reliability, high density, high scalability and high usability, is a new-generation unified network

More information

Cray XC Scalability and the Aries Network Tony Ford

Cray XC Scalability and the Aries Network Tony Ford Cray XC Scalability and the Aries Network Tony Ford June 29, 2017 Exascale Scalability Which scalability metrics are important for Exascale? Performance (obviously!) What are the contributing factors?

More information

Post-K Development and Introducing DLU. Copyright 2017 FUJITSU LIMITED

Post-K Development and Introducing DLU. Copyright 2017 FUJITSU LIMITED Post-K Development and Introducing DLU 0 Fujitsu s HPC Development Timeline K computer The K computer is still competitive in various fields; from advanced research to manufacturing. Deep Learning Unit

More information

POWER4 Systems: Design for Reliability. Douglas Bossen, Joel Tendler, Kevin Reick IBM Server Group, Austin, TX

POWER4 Systems: Design for Reliability. Douglas Bossen, Joel Tendler, Kevin Reick IBM Server Group, Austin, TX Systems: Design for Reliability Douglas Bossen, Joel Tendler, Kevin Reick IBM Server Group, Austin, TX Microprocessor 2-way SMP system on a chip > 1 GHz processor frequency >1GHz Core Shared L2 >1GHz Core

More information

IBM Power AC922 Server

IBM Power AC922 Server IBM Power AC922 Server The Best Server for Enterprise AI Highlights More accuracy - GPUs access system RAM for larger models Faster insights - significant deep learning speedups Rapid deployment - integrated

More information

1. NoCs: What s the point?

1. NoCs: What s the point? 1. Nos: What s the point? What is the role of networks-on-chip in future many-core systems? What topologies are most promising for performance? What about for energy scaling? How heavily utilized are Nos

More information

eslim SV Xeon 2U Server

eslim SV Xeon 2U Server eslim SV7-2250 Xeon 2U Server www.eslim.co.kr Dual and Quad-Core Server Computing Leader!! ELSIM KOREA INC. 1. Overview Hyper-Threading eslim SV7-2250 Server Outstanding computing powered by 64-bit Intel

More information

MAHA. - Supercomputing System for Bioinformatics

MAHA. - Supercomputing System for Bioinformatics MAHA - Supercomputing System for Bioinformatics - 2013.01.29 Outline 1. MAHA HW 2. MAHA SW 3. MAHA Storage System 2 ETRI HPC R&D Area - Overview Research area Computing HW MAHA System HW - Rpeak : 0.3

More information

ANNEXES. to the. COMMISSION REGULATION (EU) No /.. of XXX

ANNEXES. to the. COMMISSION REGULATION (EU) No /.. of XXX EUROPEAN COMMISSION Brussels, XXX [ ](2018) XXX draft ANNEXES 1 to 6 ANNEXES to the COMMISSION REGULATION (EU) No /.. of XXX implementing Directive 2009/125/EC of the European Parliament and of the Council

More information

Chapter 6. Storage and Other I/O Topics

Chapter 6. Storage and Other I/O Topics Chapter 6 Storage and Other I/O Topics Introduction I/O devices can be characterized by Behaviour: input, output, storage Partner: human or machine Data rate: bytes/sec, transfers/sec I/O bus connections

More information

Sugon TC6600 blade server

Sugon TC6600 blade server Sugon TC6600 blade server The converged-architecture blade server The TC6600 is a new generation, multi-node and high density blade server with shared power, cooling, networking and management infrastructure

More information

Efficient Evaluation and Management of Temperature and Reliability for Multiprocessor Systems

Efficient Evaluation and Management of Temperature and Reliability for Multiprocessor Systems Efficient Evaluation and Management of Temperature and Reliability for Multiprocessor Systems Ayse K. Coskun Electrical and Computer Engineering Department Boston University http://people.bu.edu/acoskun

More information

Research on the Implementation of MPI on Multicore Architectures

Research on the Implementation of MPI on Multicore Architectures Research on the Implementation of MPI on Multicore Architectures Pengqi Cheng Department of Computer Science & Technology, Tshinghua University, Beijing, China chengpq@gmail.com Yan Gu Department of Computer

More information

LS-DYNA Performance Benchmark and Profiling. October 2017

LS-DYNA Performance Benchmark and Profiling. October 2017 LS-DYNA Performance Benchmark and Profiling October 2017 2 Note The following research was performed under the HPC Advisory Council activities Participating vendors: LSTC, Huawei, Mellanox Compute resource

More information

Findings from real petascale computer systems with meteorological applications

Findings from real petascale computer systems with meteorological applications 15 th ECMWF Workshop Findings from real petascale computer systems with meteorological applications Toshiyuki Shimizu Next Generation Technical Computing Unit FUJITSU LIMITED October 2nd, 2012 Outline

More information

Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming ( )

Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming ( ) Systems Group Department of Computer Science ETH Zürich Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming (252-0061-00) Timothy Roscoe Herbstsemester 2012 Today Non-Uniform

More information

Lecture 7: PCM Wrap-Up, Cache coherence. Topics: handling PCM errors and writes, cache coherence intro

Lecture 7: PCM Wrap-Up, Cache coherence. Topics: handling PCM errors and writes, cache coherence intro Lecture 7: M Wrap-Up, ache coherence Topics: handling M errors and writes, cache coherence intro 1 Optimizations for Writes (Energy, Lifetime) Read a line before writing and only write the modified bits

More information

Today. SMP architecture. SMP architecture. Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming ( )

Today. SMP architecture. SMP architecture. Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming ( ) Lecture 26: Multiprocessing continued Computer Architecture and Systems Programming (252-0061-00) Timothy Roscoe Herbstsemester 2012 Systems Group Department of Computer Science ETH Zürich SMP architecture

More information

Altair OptiStruct 13.0 Performance Benchmark and Profiling. May 2015

Altair OptiStruct 13.0 Performance Benchmark and Profiling. May 2015 Altair OptiStruct 13.0 Performance Benchmark and Profiling May 2015 Note The following research was performed under the HPC Advisory Council activities Participating vendors: Intel, Dell, Mellanox Compute

More information

Chapter 6 Storage and Other I/O Topics

Chapter 6 Storage and Other I/O Topics Department of Electr rical Eng ineering, Chapter 6 Storage and Other I/O Topics 王振傑 (Chen-Chieh Wang) ccwang@mail.ee.ncku.edu.tw ncku edu Feng-Chia Unive ersity Outline 6.1 Introduction 6.2 Dependability,

More information

High Performance Computing with Fujitsu

High Performance Computing with Fujitsu High Performance Computing with Fujitsu Ivo Doležel 0 2017 FUJITSU FUJITSU Software HPC Cluster Suite A complete HPC software stack solution HPC cluster general characteristics HPC clusters consist primarily

More information

CAS 2K13 Sept Jean-Pierre Panziera Chief Technology Director

CAS 2K13 Sept Jean-Pierre Panziera Chief Technology Director CAS 2K13 Sept. 2013 Jean-Pierre Panziera Chief Technology Director 1 personal note 2 Complete solutions for Extreme Computing b ubullx ssupercomputer u p e r c o p u t e r suite s u e Production ready

More information

Performance and Optimization Issues in Multicore Computing

Performance and Optimization Issues in Multicore Computing Performance and Optimization Issues in Multicore Computing Minsoo Ryu Department of Computer Science and Engineering 2 Multicore Computing Challenges It is not easy to develop an efficient multicore program

More information

HPC Architectures. Types of resource currently in use

HPC Architectures. Types of resource currently in use HPC Architectures Types of resource currently in use Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us

More information

Chapter 6 - External Memory

Chapter 6 - External Memory Chapter 6 - External Memory Luis Tarrataca luis.tarrataca@gmail.com CEFET-RJ L. Tarrataca Chapter 6 - External Memory 1 / 66 Table of Contents I 1 Motivation 2 Magnetic Disks Write Mechanism Read Mechanism

More information

LECTURE 5: MEMORY HIERARCHY DESIGN

LECTURE 5: MEMORY HIERARCHY DESIGN LECTURE 5: MEMORY HIERARCHY DESIGN Abridged version of Hennessy & Patterson (2012):Ch.2 Introduction Programmers want unlimited amounts of memory with low latency Fast memory technology is more expensive

More information

Computer Organization and Structure. Bing-Yu Chen National Taiwan University

Computer Organization and Structure. Bing-Yu Chen National Taiwan University Computer Organization and Structure Bing-Yu Chen National Taiwan University Storage and Other I/O Topics I/O Performance Measures Types and Characteristics of I/O Devices Buses Interfacing I/O Devices

More information

eslim SV Xeon 1U Server

eslim SV Xeon 1U Server eslim SV7-2100 Xeon 1U Server www.eslim.co.kr Dual and Quad-Core Server Computing Leader!! ELSIM KOREA INC. 1. Overview Hyper-Threading eslim SV7-2100 Server Outstanding computing powered by 64-bit Intel

More information

Steve Scott, Tesla CTO SC 11 November 15, 2011

Steve Scott, Tesla CTO SC 11 November 15, 2011 Steve Scott, Tesla CTO SC 11 November 15, 2011 What goal do these products have in common? Performance / W Exaflop Expectations First Exaflop Computer K Computer ~10 MW CM5 ~200 KW Not constant size, cost

More information

THE PATH TO EXASCALE COMPUTING. Bill Dally Chief Scientist and Senior Vice President of Research

THE PATH TO EXASCALE COMPUTING. Bill Dally Chief Scientist and Senior Vice President of Research THE PATH TO EXASCALE COMPUTING Bill Dally Chief Scientist and Senior Vice President of Research The Goal: Sustained ExaFLOPs on problems of interest 2 Exascale Challenges Energy efficiency Programmability

More information

Computer Architecture A Quantitative Approach, Fifth Edition. Chapter 2. Memory Hierarchy Design. Copyright 2012, Elsevier Inc. All rights reserved.

Computer Architecture A Quantitative Approach, Fifth Edition. Chapter 2. Memory Hierarchy Design. Copyright 2012, Elsevier Inc. All rights reserved. Computer Architecture A Quantitative Approach, Fifth Edition Chapter 2 Memory Hierarchy Design 1 Introduction Programmers want unlimited amounts of memory with low latency Fast memory technology is more

More information

Models Smart Array 6402/128 Controller B21 Smart Array 6404/256 Controller B21

Models Smart Array 6402/128 Controller B21 Smart Array 6404/256 Controller B21 Overview The Smart Array 6400 high performance Ultra320, PCI-X controller family provides maximum performance, flexibility, and reliable data protection for HP ProLiant servers, through its unique modular

More information

Post-K: Building the Arm HPC Ecosystem

Post-K: Building the Arm HPC Ecosystem Post-K: Building the Arm HPC Ecosystem Toshiyuki Shimizu FUJITSU LIMITED Nov. 14th, 2017 Exhibitor Forum, SC17, Nov. 14, 2017 0 Post-K: Building up Arm HPC Ecosystem Fujitsu s approach for HPC Approach

More information

Fujitsu High Performance CPU for the Post-K Computer

Fujitsu High Performance CPU for the Post-K Computer Fujitsu High Performance CPU for the Post-K Computer August 21 st, 2018 Toshio Yoshida FUJITSU LIMITED 0 Key Message A64FX is the new Fujitsu-designed Arm processor It is used in the post-k computer A64FX

More information

Course II Parallel Computer Architecture. Week 2-3 by Dr. Putu Harry Gunawan

Course II Parallel Computer Architecture. Week 2-3 by Dr. Putu Harry Gunawan Course II Parallel Computer Architecture Week 2-3 by Dr. Putu Harry Gunawan www.phg-simulation-laboratory.com Review Review Review Review Review Review Review Review Review Review Review Review Processor

More information

4.1 Introduction 4.3 Datapath 4.4 Control 4.5 Pipeline overview 4.6 Pipeline control * 4.7 Data hazard & forwarding * 4.

4.1 Introduction 4.3 Datapath 4.4 Control 4.5 Pipeline overview 4.6 Pipeline control * 4.7 Data hazard & forwarding * 4. Chapter 4: CPU 4.1 Introduction 4.3 Datapath 4.4 Control 4.5 Pipeline overview 4.6 Pipeline control * 4.7 Data hazard & forwarding * 4.8 Control hazard 4.14 Concluding Rem marks Hazards Situations that

More information

Overview. Energy-Efficient and Power-Constrained Techniques for ExascaleComputing. Motivation: Power is becoming a leading design constraint in HPC

Overview. Energy-Efficient and Power-Constrained Techniques for ExascaleComputing. Motivation: Power is becoming a leading design constraint in HPC Energy-Efficient and Power-Constrained Techniques for ExascaleComputing Stephanie Labasan Computer and Information Science University of Oregon 17 October 2016 Overview Motivation: Power is becoming a

More information

Database Systems II. Secondary Storage

Database Systems II. Secondary Storage Database Systems II Secondary Storage CMPT 454, Simon Fraser University, Fall 2009, Martin Ester 29 The Memory Hierarchy Swapping, Main-memory DBMS s Tertiary Storage: Tape, Network Backup 3,200 MB/s (DDR-SDRAM

More information

Copyright 2012, Elsevier Inc. All rights reserved.

Copyright 2012, Elsevier Inc. All rights reserved. Computer Architecture A Quantitative Approach, Fifth Edition Chapter 2 Memory Hierarchy Design 1 Introduction Introduction Programmers want unlimited amounts of memory with low latency Fast memory technology

More information

Computer Architecture. A Quantitative Approach, Fifth Edition. Chapter 2. Memory Hierarchy Design. Copyright 2012, Elsevier Inc. All rights reserved.

Computer Architecture. A Quantitative Approach, Fifth Edition. Chapter 2. Memory Hierarchy Design. Copyright 2012, Elsevier Inc. All rights reserved. Computer Architecture A Quantitative Approach, Fifth Edition Chapter 2 Memory Hierarchy Design 1 Programmers want unlimited amounts of memory with low latency Fast memory technology is more expensive per

More information

Lecture 9: MIMD Architecture

Lecture 9: MIMD Architecture Lecture 9: MIMD Architecture Introduction and classification Symmetric multiprocessors NUMA architecture Cluster machines Zebo Peng, IDA, LiTH 1 Introduction MIMD: a set of general purpose processors is

More information

Understanding Sources of Inefficiency in General-Purpose Chips. Hameed, Rehan, et al. PRESENTED BY: XIAOMING GUO SIJIA HE

Understanding Sources of Inefficiency in General-Purpose Chips. Hameed, Rehan, et al. PRESENTED BY: XIAOMING GUO SIJIA HE Understanding Sources of Inefficiency in General-Purpose Chips Hameed, Rehan, et al. PRESENTED BY: XIAOMING GUO SIJIA HE 1 Outline Motivation H.264 Basics Key ideas Implementation & Evaluation Summary

More information

White paper PRIMEQUEST 1000 series high availability realized by Fujitsu s quality assurance

White paper PRIMEQUEST 1000 series high availability realized by Fujitsu s quality assurance White paper PRIMEQUEST 1000 series high availability realized by Fujitsu s quality assurance PRIMEQUEST is an open enterprise server platform that fully maximizes uptime. This whitepaper explains how Fujitsu

More information

LQCD Facilities at Jefferson Lab. Chip Watson May 6, 2011

LQCD Facilities at Jefferson Lab. Chip Watson May 6, 2011 LQCD Facilities at Jefferson Lab Chip Watson May 6, 2011 Page 1 Outline Overview of hardware at Jefferson Lab GPU evolution since last year GPU R&D activities Operations Summary Page 2 CPU + Infiniband

More information

RAID Rack (RR2035RPHMS)

RAID Rack (RR2035RPHMS) RAID Rack (RR2035RPHMS) www.addonics.com Technical Support If you need any assistance to get your unit functioning properly, please have your product information ready and contact Addonics Technical Support

More information

Mapping MPI+X Applications to Multi-GPU Architectures

Mapping MPI+X Applications to Multi-GPU Architectures Mapping MPI+X Applications to Multi-GPU Architectures A Performance-Portable Approach Edgar A. León Computer Scientist San Jose, CA March 28, 2018 GPU Technology Conference This work was performed under

More information

Power Systems AC922 Overview. Chris Mann IBM Distinguished Engineer Chief System Architect, Power HPC Systems December 11, 2017

Power Systems AC922 Overview. Chris Mann IBM Distinguished Engineer Chief System Architect, Power HPC Systems December 11, 2017 Power Systems AC922 Overview Chris Mann IBM Distinguished Engineer Chief System Architect, Power HPC Systems December 11, 2017 IBM POWER HPC Platform Strategy High-performance computer and high-performance

More information

The Cisco MCS 7835-H2 can run any of the following Cisco applications:

The Cisco MCS 7835-H2 can run any of the following Cisco applications: Cisco MCS 7835-H2 Cisco Unified Communications Solutions unify voice, video, data, and mobile applications on fixed and mobile networks enabling easy collaboration every time from any workspace. Product

More information

Cray XD1 Supercomputer Release 1.3 CRAY XD1 DATASHEET

Cray XD1 Supercomputer Release 1.3 CRAY XD1 DATASHEET CRAY XD1 DATASHEET Cray XD1 Supercomputer Release 1.3 Purpose-built for HPC delivers exceptional application performance Affordable power designed for a broad range of HPC workloads and budgets Linux,

More information