A Strategy for the Future of High Performance Computing? Advanced Computing Laboratory Los Alamos National Laboratory Pete Beckman.

Slides:



Advertisements
Similar presentations
NAS vs. SAN 10/2010 Palestinian Land Authority IT Department By Nahreen Ameen 1.
Advertisements

♦ Commodity processor with commodity inter- processor connection Clusters Pentium, Itanium, Opteron, Alpha GigE, Infiniband, Myrinet, Quadrics, SCI NEC.
Ver 0.1 Page 1 SGI Proprietary Introducing the CRAY SV1 CRAY SV1-128 SuperCluster.
Beowulf Supercomputer System Lee, Jung won CS843.
Types of Parallel Computers
Information Technology Center Introduction to High Performance Computing at KFUPM.
Managing Linux Clusters with Rocks Tim Carlson - PNNL
The road to reliable, autonomous distributed systems
Presented by Scalable Systems Software Project Al Geist Computer Science Research Group Computer Science and Mathematics Division Research supported by.
IBM RS6000/SP Overview Advanced IBM Unix computers series Multiple different configurations Available from entry level to high-end machines. POWER (1,2,3,4)
HELICS Petteri Johansson & Ilkka Uuhiniemi. HELICS COW –AMD Athlon MP 1.4Ghz –512 (2 in same computing node) –35 at top500.org –Linpack Benchmark 825.
A Comparative Study of Network Protocols & Interconnect for Cluster Computing Performance Evaluation of Fast Ethernet, Gigabit Ethernet and Myrinet.
Hitachi SR8000 Supercomputer LAPPEENRANTA UNIVERSITY OF TECHNOLOGY Department of Information Technology Introduction to Parallel Computing Group.
CENG334 Introduction to Operating Systems Erol Sahin Dept of Computer Eng. Middle East Technical University Ankara, TURKEY URL:
Why Linux is a Bad Idea as a Compute Node OS (for Balanced Systems) Ron Brightwell Sandia National Labs Scalable Computing Systems Department
IBM RS/6000 SP POWER3 SMP Jari Jokinen Pekka Laurila.
High Performance Communication using MPJ Express 1 Presented by Jawad Manzoor National University of Sciences and Technology, Pakistan 29 June 2015.
Chiba City: A Testbed for Scalablity and Development FAST-OS Workshop July 10, 2002 Rémy Evard Mathematics.
NPACI: National Partnership for Advanced Computational Infrastructure August 17-21, 1998 NPACI Parallel Computing Institute 1 Cluster Archtectures and.
1b.1 Types of Parallel Computers Two principal approaches: Shared memory multiprocessor Distributed memory multicomputer ITCS 4/5145 Parallel Programming,
Building a High-performance Computing Cluster Using FreeBSD BSDCon '03 September 10, 2003 Brooks Davis, Michael AuYeung, Gary Green, Craig Lee The Aerospace.
Parallel Computing The Bad News –Hardware is not getting faster fast enough –Too many architectures –Existing architectures are too specific –Programs.
SSI-OSCAR A Single System Image for OSCAR Clusters Geoffroy Vallée INRIA – PARIS project team COSET-1 June 26th, 2004.
Bring Consolidation Into Focus The Value of Compaq AlphaServer and Storage Consolidation Solutions Joseph Batista Director Enterprise & Internet Initiatives.
Tools and Utilities for parallel and serial codes in ENEA-GRID environment CRESCO Project: Salvatore Raia SubProject I.2 C.R. ENEA-Portici. 11/12/2007.
Silberschatz, Galvin and Gagne  Operating System Concepts Chapter 1: Introduction What is an Operating System? Mainframe Systems Desktop Systems.
University of Illinois at Urbana-Champaign NCSA Supercluster Administration NT Cluster Group Computing and Communications Division NCSA Avneesh Pant
CLUSTER COMPUTING STIMI K.O. ROLL NO:53 MCA B-5. INTRODUCTION  A computer cluster is a group of tightly coupled computers that work together closely.
1b.1 Types of Parallel Computers Two principal approaches: Shared memory multiprocessor Distributed memory multicomputer ITCS 4/5145 Parallel Programming,
March 3rd, 2006 Chen Peng, Lilly System Biology1 Cluster and SGE.
Guide to Linux Installation and Administration, 2e1 Chapter 2 Planning Your System.
Cluster Workstations. Recently the distinction between parallel and distributed computers has become blurred with the advent of the network of workstations.
Distributed Resource Management and Parallel Computation Dr Michael Rudgyard Streamline Computing Ltd.
Taking the Complexity out of Cluster Computing Vendor Update HPC User Forum Arend Dittmer Director Product Management HPC April,
Loosely Coupled Parallelism: Clusters. Context We have studied older archictures for loosely coupled parallelism, such as mesh’s, hypercubes etc, which.
Introduction to dCache Zhenping (Jane) Liu ATLAS Computing Facility, Physics Department Brookhaven National Lab 09/12 – 09/13, 2005 USATLAS Tier-1 & Tier-2.
Parallel and Grid I/O Infrastructure W. Gropp, R. Ross, R. Thakur Argonne National Lab A. Choudhary, W. Liao Northwestern University G. Abdulla, T. Eliassi-Rad.
Issues Autonomic operation (fault tolerance) Minimize interference to applications Hardware support for new operating systems Resource management (global.
CHEP04 Performance Analysis of Cluster File System on Linux Yaodong CHENG IHEP, CAS
CSC 7600 Lecture 28 : Final Exam Review Spring 2010 HIGH PERFORMANCE COMPUTING: MODELS, METHODS, & MEANS FINAL EXAM REVIEW Daniel Kogler, Chirag Dekate.
Test Results of the EuroStore Mass Storage System Ingo Augustin CERNIT-PDP/DM Padova.
Nanco: a large HPC cluster for RBNI (Russell Berrie Nanotechnology Institute) Anne Weill – Zrahia Technion,Computer Center October 2008.
Scalable Systems Software for Terascale Computer Centers Coordinator: Al Geist Participating Organizations ORNL ANL LBNL.
Welcome to the PVFS BOF! Rob Ross, Rob Latham, Neill Miller Argonne National Laboratory Walt Ligon, Phil Carns Clemson University.
Cluster Software Overview
Chapter 9: Networking with Unix and Linux. Objectives: Describe the origins and history of the UNIX operating system Identify similarities and differences.
Comprehensive Scientific Support Of Large Scale Parallel Computation David Skinner, NERSC.
COMP381 by M. Hamdi 1 Clusters: Networks of WS/PC.
HPC HPC-5 Systems Integration High Performance Computing 1 Application Resilience: Making Progress in Spite of Failure Nathan A. DeBardeleben and John.
3/12/2013Computer Engg, IIT(BHU)1 PARALLEL COMPUTERS- 2.
Pathway to Petaflops A vendor contribution Philippe Trautmann Business Development Manager HPC & Grid Global Education, Government & Healthcare.
Background Computer System Architectures Computer System Software.
HPC University Requirements Analysis Team Training Analysis Summary Meeting at PSC September Mary Ann Leung, Ph.D.
SYSTEM MODELS FOR ADVANCED COMPUTING Jhashuva. U 1 Asst. Prof CSE
Page : 1 SC2004 Pittsburgh, November 12, 2004 DEISA : integrating HPC infrastructures in Europe DEISA : integrating HPC infrastructures in Europe Victor.
Towards a High Performance Extensible Grid Architecture Klaus Krauter Muthucumaru Maheswaran {krauter,
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING CLOUD COMPUTING
Operating System & Application Software
Chapter 1: Introduction
OpenMosix, Open SSI, and LinuxPMI
Evolution of Operating Systems
Grid Computing.
CRESCO Project: Salvatore Raia
Is System X for Me? Cal Ribbens Computer Science Department
University of Technology
חוברת שקפים להרצאות של ד"ר יאיר ויסמן מבוססת על אתר האינטרנט:
Pluggable Architecture for Java HPC Messaging
CLUSTER COMPUTING.
SiCortex Update IDC HPC User Forum
MPJ: A Java-based Parallel Computing System
Presentation transcript:

A Strategy for the Future of High Performance Computing? Advanced Computing Laboratory Los Alamos National Laboratory Pete Beckman

2 Advanced Computing Laboratory Los Alamos National Laboratory Observations: The US Supercomputing Industry l All US high-performance vendors are building clusters of SMPs (with the exception of Tera) l Each company, IBM, SGI, Compaq, HP, and SUN has a different version of Unix l Each company attempts to scale system software designed for database, internet, technical and servers l This fractured market forces 5 different parallel file systems, fast message implementations, etc l Supercomputer companies tend to go out of business

Pete Beckman3 Advanced Computing Laboratory Los Alamos National Laboratory New Limitations People used to say: “The number of Tflops available is limited only by the amount of money you wish to spend” The Reality: We are at a point where our ability to build machines from components exceeds our ability to admin, program and run them But we do it anyway. Many large clusters are being installed...

Pete Beckman4 Advanced Computing Laboratory Los Alamos National Laboratory Scalable System Software is currently the weak link Software for Tflop clusters of SMPs is hard l System administration, configuration, booting, management, & monitoring l Scalable smart NIC messaging (zero copy) l Cluster/Global/Parallel File System l Job queuing and running l I/O (scratch, prefetch, NASD) l Fault tolerance and on-the-fly reconfiguration

Pete Beckman5 Advanced Computing Laboratory Los Alamos National Laboratory Why use Linux for clusters of SMPs, and as a basis for system software research? l The OS for scalable clusters needs more research l Open Source! (it’s more then just geek chic) n No lawyers, no NDAs, no worries mate! n Visible code improves faster n The whole environment, or just the mods can be distributed n Scientific collaboration is just an URL away... l Small, well designed, stable, mature, kernel n ~240K lines of code without device drivers /proc filesystem and dynamically loadable modules n The OS is extendable, optimizable, tunable Linux is a lot of fun (Shagadelic, Baby!) Did I mention no lawyers?

Pete Beckman6 Advanced Computing Laboratory Los Alamos National Laboratory Isn’t Open Source hype? Do you really need it? A very quick example Supermon and Superview: High Performance Cluster Monitoring Tools Ron Minnich, Karen Reid, Matt Sottile

Pete Beckman7 Advanced Computing Laboratory Los Alamos National Laboratory The problem: get really fast stats from a very large cluster l Monitor hundreds of nodes at rates up to 100 Hz l Monitor at 10Hz without significant impact on the application l Monitor hardware performance counters l Collect a wide range of kernel information (disk blocks, memory, interrupts, etc)

Pete Beckman8 Advanced Computing Laboratory Los Alamos National Laboratory Solution l Modify the kernel so all the parameters can be grabbed without going though /proc l Tightly coupled clusters can get real- time monitoring stats. l This is not of general use to the desktop, and web server markets l Stats for 100 nodes takes about 20 ms

Pete Beckman9 Advanced Computing Laboratory Los Alamos National Laboratory Superview: the Java tool for Supermon

Pete Beckman10 Advanced Computing Laboratory Los Alamos National Laboratory Where should we concentrate our efforts? Some areas for improvement…. Scalable Linux System Software

Pete Beckman11 Advanced Computing Laboratory Los Alamos National Laboratory Software: The hard part Linux environments (page 1) l Compilers n F90 (PGI, Absoft, Compaq) n F77 (GNU, PGI, Absoft, Compaq, Fujitsu) n HPF (PGI, Compaq?) n C/C++ (PGI, KAI, GNU, Compaq, Fujitsu) n OpenMP (PGI) n Metrowerks Code Warrior for C, C++, (Fortran?) l Debuggers n Totalview… maybe, real soon now, almost? n gdb, DDD, etc.

Pete Beckman12 Advanced Computing Laboratory Los Alamos National Laboratory Software: The hard part Linux environments (page 2) l Message Passing n MPICH, PVM, MPI MSTI, Nexus n OS Bypass: –ST, FM, AM, PM, GM, VIA, Portals, etc –Fast Interconnects: Myrinet, GigE, HiPPI, SCI l Shared Memory Programming n Pthreads, Tulip-Threads, etc. l Parallel Performance Tools n TAU, Vampir, PGI PGProf, Jumpshot, etc

Pete Beckman13 Advanced Computing Laboratory Los Alamos National Laboratory Software: The hard part Linux environments (page 3) l File Systems & I/O n e2fs (native), NFS n PVFS, Coda, GFS n MPI-IO, ROMIO l Archival Storage n HPSS & ADSM clients l Job Control n LSF, PBS, Maui

Pete Beckman14 Advanced Computing Laboratory Los Alamos National Laboratory Software: The hard part Linux environments (page 4) l Libraries and Frameworks n BLAS, OVERTURE, POOMA, Atlas n Alpha math libraries (Compaq) l System Administration n Building and booting tools n Cfengine n Monitoring and management tools n Configuration database n SGI Project Accounting

Pete Beckman15 Advanced Computing Laboratory Los Alamos National Laboratory Software for Linux clusters A report card (current status) Compilers Parallel debuggers Message passing Shared memory prog. Parallel performance tools File Systems Archival Storage Job Control Math Libraries ………………………………...………..A …………………………..……...I ………………………….….…..A- ……….…………..……..A …………...………..C+ ………………….………...………..D …………………………….……..C ………………………………...……..B- ……………………………………..B

Pete Beckman16 Advanced Computing Laboratory Los Alamos National Laboratory Summary of the most important areas l First Priority n Cluster management, administration, images, monitoring, etc n Cluster/parallel/global file systems n Continued work on scalable messaging n Faster, more scalable SMP n Virtual memory optimized for HPC n TCP/IP improvements Wish List NIC boot, BIOS NVRAM, Serial console OS bypass standards in the kernel Tightly-coupled scheduling, accounting Newest Drivers

Pete Beckman17 Advanced Computing Laboratory Los Alamos National Laboratory Honest cluster costs: publish the numbers l How many sysadmins and programmers are we required for support? l What are the service and replacement costs? l How much was hardware integration? l How many users can you support and at what levels? l How much was the hardware?

Pete Beckman18 Advanced Computing Laboratory Los Alamos National Laboratory Compute Nodes Gigabit Ethernet Control Node Control Node Gigabit Multistage Interconnection Fabric Network Attached Secure Disks Unit Tera-Scale SMP Cluster Architecture

Pete Beckman21 Advanced Computing Laboratory Los Alamos National Laboratory Let someone else put it together l Compaq l Dell l Penguin Computing l Alta Tech l VA Linux l DCG l Paralogic l Microway Ask about support

Pete Beckman22 Advanced Computing Laboratory Los Alamos National Laboratory Cluster Benchmarking Lies, Damn Lies, and the Top500 l Make MPI zero-byte messaging a special case (improves latency numbers) l Convert multiply flops to addition, recount flops l Hire a Linpack consultant to help you achieve “the number” the vendor promised l “We unloaded the trucks, and 24hrs later, we calculated the size of the galaxy in acres.” l For $15K and 3 rolls of duct tape I built a supercomputer in my cubicle…. Vendor Published Linpack, Latency, and Bandwidth numbers are worthless

Pete Beckman23 Advanced Computing Laboratory Los Alamos National Laboratory Plug-in Framework for Cluster Benchmarks

Pete Beckman24 Advanced Computing Laboratory Los Alamos National Laboratory MPI Message Matching

Pete Beckman25 Advanced Computing Laboratory Los Alamos National Laboratory

Pete Beckman26 Advanced Computing Laboratory Los Alamos National Laboratory

Pete Beckman27 Advanced Computing Laboratory Los Alamos National Laboratory Conclusions l Lots of Linux clusters will be at SC99 l The Big 5 vendors do not have the critical mass to develop the system software for multi-teraflop clusters l The HPC community (labs, vendors, universities, etc.) needs to work together l The hardware consolidation is nearly over, the software consolidation is on its way l A Linux-based “commodity” Open Source strategy could provide a mechanism for: n open vendor collaboration n academic and laboratory participation n one Open Source software environment

Pete Beckman28 Advanced Computing Laboratory Los Alamos National Laboratory News and Announcements: l The next Extreme Linux conference will be in Williamsburg in October. The call for papers will be out soon, start preparing those technical papers… l There will be several cluster tutorials at SC99. Remy Evard, Bill Saphir, and Pete Beckman will be running one focused on system administration and user environment for large clusters.