HS06 on new CPU, KVM virtual machines and commercial cloud Michele Michelotto 1.

Slides:



Advertisements
Similar presentations
Intel Multi-Core Technology. New Energy Efficiency by Parallel Processing – Multi cores in a single package – Second generation high k + metal gate 32nm.
Advertisements

Cloud Computing Mick Watson Director of ARK-Genomics The Roslin Institute.
An evaluation of the Intel Xeon E5 Processor Series Zurich Launch Event 8 March 2012 Sverre Jarp, CERN openlab CTO Technical team: A.Lazzaro, J.Leduc,
Microprocessor Microarchitecture Multithreading Lynn Choi School of Electrical Engineering.
Hepmark project Evaluation of HEP worker nodes Michele Michelotto at pd.infn.it.
OPNET Technologies, Inc. Performance versus Cost in a Cloud Computing Environment Yiping Ding OPNET Technologies, Inc. © 2009 OPNET Technologies, Inc.
Nov COMP60621 Concurrent Programming for Numerical Applications Lecture 6 Chronos – a Dell Multicore Computer Len Freeman, Graham Riley Centre for.
OPTERON (Advanced Micro Devices). History of the Opteron AMD's server & workstation processor line 2003: Original Opteron released o 32 & 64 bit processing.
1 Lecture 26: Case Studies Topics: processor case studies, Flash memory Final exam stats:  Highest 83, median 67  70+: 16 students, 60-69: 20 students.
Performance benchmark of LHCb code on state-of-the-art x86 architectures Daniel Hugo Campora Perez, Niko Neufled, Rainer Schwemmer CHEP Okinawa.
1 Chapter 01 Authors: John Hennessy & David Patterson.
HS06 on the last generation of CPU for HEP server farm Michele Michelotto 1.
Moving out of SI2K How INFN is moving out of SI2K as a benchmark for Worker Nodes performance evaluation Michele Michelotto at pd.infn.it.
COMPUTER ARCHITECTURE
Alleviating Constraints with Resource Pools & Live Migration with Enhanced VMotion* Breakout Session# 2823 Raghu Yeluri Sr. Architect Intel Corporation.
SUMMER VACATION SCHOLARSHIP | IM&T Scientific Computing in the Cloud.
Enabling Technologies for Distributed and Cloud Computing Dr. Sanjay P. Ahuja, Ph.D FIS Distinguished Professor of Computer Science School of.
Different CPUs CLICK THE SPINNING COMPUTER TO MOVE ON.
Computer Performance Computer Engineering Department.
Multi-core architectures. Single-core computer Single-core CPU chip.
Intel’s Penryn Sima Dezső Fall 2007 Version nm quad-core -
High Performance Computing Processors Felix Noble Mirayma V. Rodriguez Agnes Velez Electric and Computer Engineer Department August 25, 2004.
Comparing Intel’s Core with AMD's K8 Microarchitecture IS 3313 December 14 th.
CSE 451: Operating Systems Spring 2013 Module 26 Cloud Computing Ed Lazowska Allen Center 570 © 2013 Gribble, Lazowska, Levy,
Fast Benchmark Michele Michelotto – INFN Padova Manfred Alef – GridKa Karlsruhe 1.
1 Latest Generations of Multi Core Processors
Virtualisation Front Side Buses SMP systems COMP Jamie Curtis.
Hyper Threading Technology. Introduction Hyper-threading is a technology developed by Intel Corporation for it’s Xeon processors with a 533 MHz system.
Cloud Computing Ed Lazowska Bill & Melinda Gates Chair in Computer Science & Engineering University of Washington.
CSE 451: Operating Systems Autumn 2010 Module 25 Cloud Computing Ed Lazowska Allen Center 570.
Enabling Technologies for Distributed Computing Dr. Sanjay P. Ahuja, Ph.D. Fidelity National Financial Distinguished Professor of CIS School of Computing,
Ian Gable HEPiX Spring 2009, Umeå 1 VM CPU Benchmarking the HEPiX Way Manfred Alef, Ian Gable FZK Karlsruhe University of Victoria May 28, 2009.
HS06 on last generation of HEP worker nodes Berkeley, Hepix Fall ‘09 INFN - Padova michele.michelotto at pd.infn.it.
Multi-core CPU’s April 9, Multi-Core at BNL First purchase of AMD dual-core in 2006 First purchase of Intel multi-core in 2007 –dual-core in early.
HS06 performance per watt and transition to SL6 Michele Michelotto – INFN Padova 1.
HEPMARK2 Consiglio di Sezione 9 Luglio 2012 Michele Michelotto - Padova.
Industry Standard Servers Infrastructure Refresh Why?
From Westmere to Magny-cours: Hep-Spec06 Cornell U. - Hepix Fall‘10 INFN - Padova michele.michelotto at pd.infn.it.
Parallel Computers Today Oak Ridge / Cray Jaguar > 1.75 PFLOPS Two Nvidia 8800 GPUs > 1 TFLOPS Intel 80- core chip > 1 TFLOPS  TFLOPS = floating.
New CPU, new arch, KVM and commercial cloud Michele Michelotto 1.
Performance analysis comparison Andrea Chierici Virtualization tutorial Catania 1-3 dicember 2010.
The last generation of CPU processor for server farm. New challenges Michele Michelotto 1.
© 2008 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice ProLiant G5 to G6 Processor Positioning.
PASTA 2010 CPU, Disk in 2010 and beyond m. michelotto.
Hardware Architecture
HPC/HTC vs. Cloud Benchmarking An empirical evaluation of the performance and cost implications Kashif Iqbal - PhD ICHEC, NUI Galway,
Moving out of SI2K How INFN is moving out of SI2K as a benchmark for Worker Nodes performance evaluation Michele Michelotto at pd.infn.it.
Multi-Core CPUs Matt Kuehn. Roadmap ► Intel vs AMD ► Early multi-core processors ► Threads vs Physical Cores ► Multithreading and Multi-core processing.
SI2K and beyond Michele Michelotto – INFN Padova CCR – Frascati 2007, May 30th.
Benchmarking of CPU models for HEP application
Intel and AMD processors
Brief introduction about “Grid at LNS”
Manycore processors Sima Dezső October Version 6.2.
Evaluation of HEP worker nodes Michele Michelotto at pd.infn.it
CCR Autunno 2008 Gruppo Server
Hardware 201: Selecting Database Hardware
Gruppo Server CCR michele.michelotto at pd.infn.it
Solid State Disks Testing with PROOF
How to benchmark an HEP worker node
Low Power processors in HEP
Gruppo Server CCR michele.michelotto at pd.infn.it
How INFN is moving out of SI2K has a benchmark for Worker Nodes
Passive benchmarking of ATLAS Tier-0 CPUs
Chapter 6 Warehouse-Scale Computers to Exploit Request-Level and Data-Level Parallelism Topic 11 Amazon Web Services Prof. Zhang Gang
Comparing dual- and quad-core performance
CERN Benchmarking Cluster
INFN - Padova michele.michelotto at pd.infn.it
Multi-Core Computing Osama Awwad Department of Computer Science
HEPiX Spring 2009 Highlights
Expanded CPU resource pool with
Presentation transcript:

HS06 on new CPU, KVM virtual machines and commercial cloud Michele Michelotto 1

The HEP server for CPU farm 2  Two socket  Rack mountable: 1U, 2U, dual twin, blade  Multicore  About 2GB per logical cpu  x86-64  Intel or AMD

AMD 3

4

Interlagos 5

Intel roadmap 6

New Intel Naming After Several generation of Xeon 5n xx 51xx (Woodcrest /Core 2c 65nm) 53xx (Clovertown / Core 4c 65nm) 54xx (Harpertown / Penryn 4c 45nm) 55xx (Gainestown / Nehalem 4c/8t 45nm) 56xx (aka Gulftown / Nehalem 6c/12t 45) Now Xeon E5 26xx “Sandy Bridge” EP 8c/16t nm ) 7

The dual proc Xeon E5 26xx 8

Configuration Software 9  Operating System: SL release 5.7 (Boron)  Compiler: gcc version (Red Hat )  HEP-SPEC06 based on SPEC CPU 1.2 (32bit)  HEP-SPEC06 64 bit (default config + remove “–m32”)  2GB per core unless explicitly stated

AMD x16core 64GB at 2.1( up to 2.6) GHz 10

AMD x16core 64GB at 2.1 GHz in 64bit mode 11

Piledriver core 12  Substitute the “Bulldozer” core  Smarter prefetching  A perceptron branch predictor that supplements the primary BPU  Larger L1 TLB  Schedulers that free up tokens more quickly  Faster FP and integer dividers and SYSCALL/RET (kernel/System call instructions)  Faster Store-to-Load forwarding

Abu Dhabi Family 13

AMD x16core 64GB at 2.4GHz vs GHz – 32bit 14

AMD x16core 64GB at 2.4GHz vs GHz – 64bit 15

Better architecture 16

Intel vs AMD 17

Intel Xeon E5 18

After E5 19  E5 version 2?  Ivy Bridge core with 22nm tri-gate process  Reduce TDP  Up to 10 cores with 25MB (or 30?) L3 cache  15% - 20% clock update

Mining in the SPEC site 20  Selected CPU INT RATE 2006 as a proxy of HS06  Kept only measurements on dual proc configuration  Only x86-64 architecture intel or AMD

Clock range is stable 21

Since Aug

Elastic Cloud 23  European Project e-Fiscal  Sergio Andreozzi (EGI)  Kashif Iqbal (ICHEC, NUI Galway, Ireland)  HS06 is the main benchmark in EGI  How we compare it with Amazon Elastic Cloud (EC2)  Create Virtual Machines environment on SL  Map to EC2 type of computing node

24

Xeon E vs EC2 25

Opteron 6272 vs EC2 26

HS06 for Medium (SBridge) SPEC score no. of VMs Virtualisation + Multi-tenancy effect on performance ~ 3.28% to 58.48% More realistic figure ~ to HPC/HTC vs. Cloud Benchmarking – eFiscal EGI TF 2012, Prague 27 HS06 Score

HS06 for Large (SBridge) Virtualisation + MT effect on performance ~ 9.49% to 57.47% Note the minimal effect of > no. of VMs HPC/HTC vs. Cloud Benchmarking – eFiscal EGI TF 2012, Prague 28 HS06 Score

HS06 for Xlarge (SBridge) Virtualisation + MT effect on performance ~ 8.14% to 55.84% Note the minimal effect of > no. of VMs HPC/HTC vs. Cloud Benchmarking – eFiscal EGI TF 2012, Prague 29 HS06 Score

HS06 for Medium (Opteron) Virtualisation + MT effect on performance ~ 3.77% to 47.89% HPC/HTC vs. Cloud Benchmarking – eFiscal EGI TF 2012, Prague 30 HS06 Score

HS06 for Large (Opteron) HPC/HTC vs. Cloud Benchmarking – eFiscal EGI TF 2012, Prague 31 Virtualisation + MT effect on performance ~ 9.04% to 48.88% HS06 Score