Page 1
©2011 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice ©2011 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice
[email protected]
HPC Architect
HP Russia
BEST PRACTICES FOR ACCELERATING INNOVATION WITH HPC
HPC DAY-2013 Novosibirsk
Page 2
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 2
HPC is all around you
Media & Entertainment
•Rendering •Gaming •Big Data
GeoSciences
•Seismic •Reservoir Modeling •Big Data
Engineering/ Manufacturing
•Structural Analysis •Fluid Dynamics •Impact Modeling •Big Data
Government /
Classified
•Cryptography •Military/Security •Nuclear Safety •Big Data
Financial Services
•Financial Analytics •High Freq Trading
Life Sciences
•Drug design •Next gen sequencing •Bio Informatics •Big Data
Government/ Academic Research
•Particle Physics •Life Sciences •Climate Modeling •Big Data
Engineering/ Manufacturing
•Electrical Design •Circuit Verification •Board Layout
Page 3
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 3
IDC HPC Segmentation, 2011
Who are the players
Workgroup
19%
Departmental
32% Divisional
12%
Super-
computer
37%
Supercomputing: >$500K 1. IBM 2. HP 3. Cray 4. Other
Divisional: $250k-500K 1. HP 2. IBM 3. Dell 4. Other
Departmental: $100k-250K 1. HP 2. Dell 3. IBM 4. Other
Workgroup: <$100k 1. HP 2. Dell 3. IBM 4. Other
Other includes Niche players – SGI, Cray
Regional players – T-Platforms, Fujitsu White box players – Appro, Supermicro Worldwide Technical Computing Server
2011–2015 Forecast, IDC, May, 2011
Page 4
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Affordable performance and technology transforming HPC
4
HP leading this transition
Per
form
ance
• Early community and customer engagement
• 1st GPU-enabled servers
• SL’s integrated support
• Catalyst partner program for multi-core software
• 1st IB support for Blades,
• Focused MPI optimization for IB
• Extensive benchmarking
• Unified Cluster Portfolio
• HP Cluster Platforms
• Insight Cluster Management Utility
x86 clusters, with Linux
Multi-core, and
InfiniBand
GPGPUs/co-processors
Low power computing
• Discovery lab
• Ecosystem development
• Commercialization
Page 5
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 5
Main Stream
Flexible HPC
HPC Technology is Specializing
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
MOD2MOD1
10/100/1000 Base-T 1000 Base-X
RPSPWR49/46 52/4751/4550/48
Console Unit5 6 15 1613 1411 129 107 81 2 3 4 21 22 31 3229 3027 2825 2623 2417 18 19 20 37 38 47 4845 4643 4441 4239 4033 34 35 36
H3C S5500 SeriesSpeed:Green = 1000Mbps,Yellow = 10/100Mbps Duplex:Green = Full Duplex,Yellow = Half Duplex
Mode Green=SpeedYellow=Duplex
HP is Converging Infrastructure
Providing Flexibility and Optimization
Page 6
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
HPC solutions
Mostly clusters (70%+ of market)
Leveraging our Unified Cluster Portfolio and Cluster Platforms
Flexible design allows us to configure to meet specific workload needs
Standalone systems
Large SMP systems, used for large memory applications (e.g., pre-processing)
Workstations
Services
Factory integration, on-site installation and start-up services, outsourcing
Data Center infrastructure including PODs
Emerging – HPC cloud capability
6
Page 7
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 7
Tackle any challenge, with HPC solutions from HP
Accelerate your innovation
Faster
Speed advancements with a converged infrastructure, purpose-built for scale.
7
Better
Optimize your performance footprint with the world’s most efficient systems.
Smarter
Deploy easily, adapt quickly to change, and improve quality of service.
Overcome barriers to Innovation and Scale • Realized system performance and throughput
• Power capacity and cost
• Infrastructure complexity and inflexibility
Page 8
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
HP Unified Cluster Portfolio
Simplified cluster design and deployment • Base of HP Cluster Platforms with system nodes
and networking, and choice of software
• Storage options for scalable I/O
• Springboard for new technologies
• Reference platform for ISV qualification
8
HP Cluster Platforms
Operating systems and extensions
Cluster management layer
Scalable data management
Advanced and specialty options
HPC application, development and cloud software portfolio
HP Datacenter Products and Services
HP Technical and Enterprise Services
Page 9
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 9
Unified Cluster Portfolio: The world’s first Converged Infrastructure for HPC
We win with complete solutions
HP Cluster Platform Rapid deployment &
high quality
Cluster/Resource/Cloud Management
Scalable provisioning and management
Application Portfolio & Expertise
Trusted advisor, confident deployment
Rack and Power Efficiency and management
HP POD Efficiency and modular
deployment
HP ProLiant DL-series Maximum flexibility
HP BladeSystem Mission-critical manageability
Scale-out storage HPC performance
and scale
HP ProLiant SL Scalable Systems Optimized for HPC/hyperscale
Accelerators/GPUs HPC performance
and efficiency
High-performance InfiniBand/10GbE HPC performance
and scalability
HP POD Efficiency and modular
deployment
Innovation
Partnerships
Efficiency
Agility
Rap
id d
eplo
ymen
t th
ru
Fact
ory
Exp
ress
Page 10
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 10
HP ProLiant Gen8 systems driven new levels of performance, efficiency and agility
HP eliminates the barriers to scale
Next gen performance with the HP ProLiant SL6500 Gen8 portfolio
The ProLiant SL6500 Gen8 portfolio, purpose-built for HPC, enables scientific and engineering innovation
Integrated accelerators boost performance
Family of integrated accelerator offerings enables explosive growth in performance and efficiency
New levels of scalability with FDR InfiniBand
Mellanox 56 Gb/s FDR InfiniBand establishes the basis for new levels of performance and scalability
Intel Xeon Phi
NVIDIA Tesla GPUs
ProLiant SL230s Gen8 SL250s Gen8 and
SL270s Gen8
ProLiant SL6500
Mellanox 56 Gb/s FDR IB
Page 11
Confidential and Proprietary
ScaleMP’s solutions start where
the hardware end
5/28/2013 - 11 -
Page 12
Confidential and Proprietary
Server Virtualization
Hypervisor or VMM
Virtual Machines App OS
App OS
App OS
Virtual Machine App OS
Hypervisor or VMM
Hypervisor or VMM
Hypervisor or VMM
Hypervisor or VMM
AGGREGATION Concatenation of physical resources
PARTITIONING Subset of the physical resources
- 12 -
Page 13
Confidential and Proprietary
Over 4,000 systems in hundreds of deployment across 30 countries
Israel
Qatar India
China
Brazil
South Africa
Canada
US
Russia
Spain
France
Belgium
Denmark
United Kingdom
S. Korea
Hong Kong
Vietnam
Brunei
Taiwan
Poland
Italy
Germany Netherlands
Australia
Norway
Japan
Malaysia
Singapore
Creating the Power of One 1 VM 1 OS N x Servers N x OS
ScaleMP Fact-sheet Founded in 2003
Proven and mature technology
Scalable Virtual Machine (VM)
SMP replacement for application capability and system manageability
World’s largest SMP
32,768 CPUs & 256 TB RAM
Standards-based
x86 and InfiniBand (Intel/Mellanox)
Product shipping since 2006
Over 4000 systems in hundreds of deployment across 30 countries
Channel-only business model
5/28/2013 - 13 -
Page 14
Confidential and Proprietary 5/28/2013 - 14 -
Page 15
Confidential and Proprietary 28-May-13 - 15 -
…and many more homegrown applications
Life Sciences Comp. Chemistry
•AMBER
•CFOUR
•DOCK
•GAMESS
•Gaussian
•GOLD
•NWChem
•Octopus
•OpenEye FRED
•OpenEye OMEGA
•Schrödinger Jaguar
•Schrödinger Glide
•SCM ADF
•VASP
Molecular Dynamics
•GROMACS
•MOLPRO
•NAMD
•OpenEye ROCS
•Schrödinger Desmond
•Turbomole
Manufacturing Structural Mechanics
•ABAQUS/Explicit
•ABAQUS/Standard
•ALTAIR Radioss
•ANSYS Mechanical
•LSTC LS-DYNA
•NASTRAN
•TNO Diana
Fluid Dynamics
•ANSYS CFX
•ANSYS Fluent
•ANSYS TGrid
•AVL FIRE
•EXA PowerFlow
•EZNSS
•GeoDict
•MHD3D
•NASA Cart3D
•STAR-CCM+
•STAR-CD
Other
•Comsol
• inTrace OpenRT
Energy • IMEX
•Norsar 3D
•Paradigm GeoDepth
•Schlumberger ECLIPSE
Bio-informatics •454/Newbler
•Abyss
•Bowtie
•CLC Bio
•FASTA
•HMMER
• Illumina
•mpiBLAST
•SOAPDenovo
•Velvet
Numerical Simulations
•Octave
•R
•MathWorks MATLAB
•Wolfram Mathematica
Finance •KX
•Wombat
EDA •Cadence
•HSPICE
•Mentor
•Quartus
•Silvaco SmartSpice
•Synopsys Weather Forecasting
•MITgcm
•MM5 (MPI & OpenMP)
•MOM4
•WRF
RDBMS & Analytics •Actian VectorWise
•Oracle
•SAP HANA
•Sybase
Selected applications used by ScaleMP’s end-users…
Page 16
Confidential and Proprietary
vSMP Foundation - Solutions Single
Operating System Single
(large) System Single
Infrastructure
Cluster management and server
consolidation
vSMP Foundation for Cluster
Compute and memory demanding
applications
vSMP Foundation for SMP
Cloud enabler – on-demand infrastructure
vSMP Foundation for Cloud
Single OS Single Infrastructure Single System 5/28/2013 - 16 -
Application Driven
Compute Intensive
CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory
Operating System App.
Large Memory
CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory
Operating System App.
I/O Intensive
CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory
Operating System App.
IT Driven
Consolidation
CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory CPU CPU
I/O
Memory
Operating System App. App. App. App. App. App. App. App.
Page 17
Confidential and Proprietary
Memory Expansion Performance
5/28/2013 - 17 -
300
650
1 150
1 800
2 250 100%
92%
78%
67% 67%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0
500
1 000
1 500
2 000
2 500
100 200 300 400 500
Effi
cie
ncy
Ru
nti
me
(se
c.)
Raw Data Size (GB)
In-memory Database Performance
Runtime
Efficiency
Max. capability with single VM
Large VM contribution
Single System
Page 18
Confidential and Proprietary 5/28/2013 - 18 -
29 118 43 753 12 844 36 482
-33%
-65%
-100%
-90%
-80%
-70%
-60%
-50%
-40%
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
-50 000
-40 000
-30 000
-20 000
-10 000
0
10 000
20 000
30 000
40 000
50 000
vSMP FoundationIntel Xeon (64 cores)
E5-2670 @ 2.6GHz64GB/socket
8TB total
SGI UV1Intel Xeon (64 cores)
X7560 @ 2.27GHz64GB/socket
16TB total
vSMP FoundationIntel Xeon (64 cores)
E5-2670 @ 2.6GHz64GB/socket
8TB total
SGI UV1Intel Xeon (64 cores)
X7560 @ 2.27GHz64GB/socket
16TB total
Foxglove Diporphyrin
Gaussian Performance Comparison - 64 cores
RuntimePerformance Difference
Single System
Page 19
Confidential and Proprietary 5/28/2013 - 19 -
2 410 3 096 3 675 3 670 4 328 7 965
18 000 21 778
28 229 25 380
29 006
37 501
-18%
-36%
-13%
-36%
-100%
-90%
-80%
-70%
-60%
-50%
-40%
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
-50 000
-40 000
-30 000
-20 000
-10 000
0
10 000
20 000
30 000
40 000
50 000
vSMP FoundationIntel Xeon
E5-2670 @ 2.6GHz64GB/socket
8TB total
Sun X4600M2AMD Opteron
8380 @ 2.5GHz64GB/socket512GB total
SGI UV1Intel Xeon
X7560 @ 2.27GHz64GB/socket
16TB total
vSMP FoundationIntel Xeon
E5-2670 @ 2.6GHz64GB/socket
8TB total
Sun X4600M2AMD Opteron
8380 @ 2.5GHz64GB/socket512GB total
SGI UV1Intel Xeon
X7560 @ 2.27GHz64GB/socket
16TB total
brain.chr1.k29 (Hash=118GB, Graph=344GB) blood.chr1.k29 (Hash=164GB, Graph=459GB)
Velvet Performance Comparison - 16 cores
Graph
Hash
Single System
Page 20
Confidential and Proprietary
The Virtualized Datacenter
5/28/2013 Single Infrastructure
• Single infrastructure, many workloads. Allows for big-data problems in cloud environment
• Integrated with CMU and other provisioning software. On-demand SMP.
- 20 -
Big Data VMs
Compute Resource
Big Data Resource Compute VM
Page 21
Confidential and Proprietary
Doing more –
Big-data and analytics - handle larger data volumes, faster
Technical computing - larger, more refined models
Hardware utilization - different workloads = single hardware platform
Paying less –
System and cluster management - free IT resources for other tasks
Software licensing - reduce application license cost, more RAM means better throughput
Optimize hardware investment - single hardware platform = different workloads
End to end solutions offering lower CAPEX and OPEX
5/28/2013
What's in it for our end-customers ?
- 21 -
Page 22
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Mellanox 56Gb/s FDR* InfiniBand supported at full speed
New levels of fabric performance
22
Next generation ConnectX-3 Flexible LOM • The industry’s first FDR 56Gb/s InfiniBand and 10/40
gigabit Ethernet multi-protocol adapter, with PCIe-gen3 for full bandwidth
End-to-End FDR solutions • A complete solution for FDR 56Gb/s InfiniBand consisting
of adapter cards, switch systems, software and cables
Largest FDR system on the Nov’12 TOP500 list • “Carter” cluster at Purdue University
• based on 648 SL230s Gen8 servers Mellanox FDR IB Flexible LOM
* Fourteen Data Rate
Page 23
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Qualified range of options, available installed on HP Clusters
Best of breed software to support your needs
23
Operating systems: Red Hat, SUSE, or customer-supported community distributions –
as well as Windows HPC
Cluster management: HP Insight Cluster Management Utility (CMU), or third party, via
HP Software and Licensing Management Solutions (SLMS) or customer installed
MPI: HP-MPI, or third party/open source
Workload manager: Platform LSF (HP SKU), Altair PBS Professional (HP SKU), Adaptive
Computing Moab (HP user unit SKU)
ScaleMP – Virtual SMP software – for large memory, large SMP
HP HPC Linux Value Pack: HP-MPI, Platform HPC (LSF, Application Portal)
Page 24
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
X9000 Network Storage System and Fusion File System • Shared datacenter multipurpose storage
− Desktops, clusters, farms, clouds
− Windows, Linux; CIFS/NFS
• Shared datacenter multipurpose storage
• High performance and scale with distributed metadata
• Data tiering
• Will address most HPC needs, including; bioInformatics, FSI, GeoScience apps with many small files
Scalable storage solutions for HPC
Cluster File System with DataDirect Networks/Xyratex • Tightly coupled HPC storage
− For large HPC Linux clusters with large single files/single stream requirements (traditional HPC, such as CAE)
• High parallel bandwidth
• High capacity
• High scalability and reliability
• Lustre-open source technology
24
Page 25
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Control Monitor
Hyperscale cluster lifecycle management software
HP Insight Cluster Management Utility (CMU)
Provision
• Simplified discovery, firmware audits
• Fast and scalable cloning
• ‘At a glance’ view of entire system; zoom to component
• Customizable
• Lightweight and efficient
• GUI and CLI options
• Easy, friction-less control of remote servers
25
• 10 years+ in deployment, included Top500 sites with1000’s of nodes
• Built for Linux, with support for multiple Linux distributions
• HP supported, available as factory-integrated cluster option
Page 26
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Vendor Product Key features of Insight CMU
Integration
Adaptive Moab HPC Suite
with dynamic provisioning capability
Moab scheduler integrated with CMU
imaging for dynamic provisioning
Platform LSF
Scheduler with application portal
CMU compute nodes are added to LSF
hosts database
LSF job metrics are displayed by CMU
real-time monitor
Altair PBS Professional
Scheduler with dynamic provisioning
capability
PBS scheduler integrated with CMU
imaging for dynamic provisioning
ScaleMP vSMP
Virtual SMP operating system for
multiple servers
Virtual SMP systems can be imaged by
CMU
StackIQ
Rocks+
HPC stack and cluster provisioning
system and package manager
One-button installation of CMU as part of a
Rocks roll
CMU monitoring and control of Rocks-
provisioned systems
The Insight CMU Ecosystem: Third-Party Options
Page 27
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
HP = choice without the pain and risk
HP Cluster Platforms installed HP Cluster Platforms delivered
HP Cluster Platforms: Simple and robust
Typical cluster delivery
Lots of choice but:
• Work • Time • Risk
Not a reference design
27
Page 28
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 28
HP industry leadership required to unlock savings
Project Moonshot – Extreme Low Energy Servers
By 2015, extreme low-energy servers could grow to 10-15% of the global server market
Space Energy Costs
HP Project Moonshot
• 1,000s of servers per rack
• Workload - tuned servers
• Federated infrastructure scales seamlessly with additional servers
+
Extreme Low Energy Servers
Page 29
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 29
A new era of extreme scale computing Complete solution for breakthrough efficiencies in density, energy, cost and simplicity
HP Project Moonshot Infrastructure
Federated System Architecture
Customer Validation Lab
HP Discovery Lab HP Pathfinder Program
Partner Collaboration
Open for Business Focused Workloads & Solution Ecosystem
“Redstone” “Gemini”
Server Development Platform
The first production Moonshot system
Page 30
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 30
HP’s first compute cartridge for the Moonshot Project Architecture
‘Redstone’ Development Platform—Calxeda EnergyCore™
X 72
HP ‘Redstone’ Development Platform Compute Cartridge
HP ‘Redstone’ Development Platform
Calxeda EnergyCore™ SoCs
I/O Controllers SATA, PCIe, Ethernet, SD/eMMC
Processor Complex
L2 C
ach
e
Mem
ory
C
on
tro
ller
s
ARM
ARM
ARM ARM
EnergyCore Fabric Switch
EnergyCore Management
Engine
0
20
40
60
80
100
120
140
x86 (Single CPU w/12GB
DDR)
Calxeda (4 SOC's w/16GB
DDR)
Wa
tts
Processor Memory Chipsets
~123 Watts
~20 Watts
• Robust—4 ARM cores integrated with 4MB L2 cache and high bandwidth memory controllers
• Efficient—Total power: 5 watts; idle power: < ½ watt
• Integrated—Systems management and high performance fabric
• Simple—Standard drivers, standard interfaces, no surprises
X 4
Page 31
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 31
Breakthrough savings and simplicity Radically changing economics of IT
89% less energy 94% less space 63% less cost 93% less HW complexity
400 servers in 10 racks 20 switches 1,600 cables 91 kilowatts
1,600 servers in 1/2 rack 2 switches 41 cables
9.9 kilowatts
Based on weighted average performance projections for workloads such as web serving, memcached, and Data Analytics. Cost estimates include infrastructure, space, and power and cooling costs over three years.
Traditional x86
$3.3M
Example: “Redstone” Server
$1.2M
Page 32
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 32
Moonshot: Gemini 101 Quad-Node cartridge =180 nodes per chassis Single-Node cartridge = 45 nodes per chassis
: Compute, Storage, or Both,
X86 and ARM
Integrated A & B Switches:, 180x10G downlinks,
6 x10G Stackable Uplinks
Hot Plug Fans
Federated iLO Chassis Management Module
Common Slot Power Supplies
A & B Switch Uplink Personality Modules
Page 33
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 33
Recent References/Success Stories •Airbus –The largest industrial supercomputer in the world (as of June’12)
–2048 BL280c servers, with storage and QDR InfiniBand
–Deployed in only 4 months in 40’ PODs in Toulouse and Hamburg
•University Purdue –Fastest campus system in the US, based on future SL6500 systems
–Deployed in 3 weeks, Top500 #54, 86.87% efficiency
•ENI Italy –Large Oil & Gas reference using HP’s SL6500/SL390s servers
–1247 SL390s servers, implemented in less than 2 months
How we win – performance, efficiency and agility
Photos courtesy of Airbus
Page 34
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 34
•The U.S. Department of Energy's National Renewable Energy Laboratory (NREL) has selected HP and Intel to provide a new energy-efficient high performance computer (HPC) system dedicated to energy systems integration, renewable energy research, and energy efficiency technologies.
Net Zero Datacenter – truth or myth? •NREL to work with HP and Intel to create one of the world’s most energy efficient data centers
•The new center will provide additional computing resources to support the breadth of research at NREL, leading to increased efficiency and lower costs for research into clean energy technologies including solar photovoltaics, wind energy, electric vehicles, buildings technologies, and renewable fuels
•The $10 million HPC system will reside at the Energy Systems Integration Facility (ESIF), under construction on the Golden, Colorado, campus. It will also support research into fully integrated energy systems that would otherwise be too expensive, or even impossible, to study directly. The HPC’s petascale computing capability (1 million billion calculations per second) is the world’s largest computing capability dedicated solely to renewable energy and energy efficiency research
Page 35
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 35
Only HP has the technology and capability to deliver the future
HP Leadership—Today and Tomorrow
Change ready architectures & systems
Modular data centers
Industry standard servers
38% x86 market share 61 qtrs of leadership
50% x86 blade share— more than IBM, Dell & Cisco combined
43% of virtualized x86 servers
up to… 88% faster deployment 75% capex savings 95% less facilities energy
21 years of leadership - shipping a server every 11 seconds
Extreme low-energy servers
89% less energy 94% less space 97% less complexity
Page 36
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 36
HP delivers high-performance innovation at any scale.
Accelerate innovation with HP
36
Faster
Speed advancements with a converged
infrastructure, purpose-built for scale.
Better
Optimize your performance footprint with
the world’s most efficient systems.
Smarter
Deploy easily, adapt quickly to change, and
improve quality of service.
HP Converged
Infrastructure
Page 37
© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Thank you