Innovation without Limits SGI ® High Performance Computing Silicon Graphics, Inc. Nyuli Gábor Silicon Computers Kft.
Dec 26, 2015
Innovation without LimitsSGI® High Performance Computing
Silicon Graphics, Inc.
Nyuli GáborSilicon Computers Kft.
2
Silicon Graphics
• Focus on the technical computing market
• Over 20 years of experience in high performance computing innovation
• Technology designed to enable the most significant scientific and creative breakthroughs of the 21st century
• Products and services are mission critical to government and defense, science and research, manufacturing, energy and media industries
Industry Leading Compute, Storage and Visualization Solutions
Images courtesy of Dr. Arthur W. Toga, The Laboratory of Neuro Imaging; American Museum of Natural History; Industrial Light and Magic, Lucasfilm Ltd; Image courtesy of Parametric Technology Corporation.
3
Science35%
Manufacturing20%
Defense and Security
30%
Media10%Energy
5%
Media
Focus Is on Engineers, Scientists and Creative Professionals
ScienceDefense & Security
Manufacturing
Images courtesy of EPL Productions, Leonard Wikberg III, EnSight image, Landmark, and Georgia Public Broadcasting
Energy
4
SGI® Core Technology Offerings
High-PerformanceComputing
• Linux® OS and Intel® Itanium® 2 Processor-based platform
• Scales up and out• Programming models and libraries• System partitioning and
resource management
AdvancedVisualization • Unmatched visualization
• Leverages industry and open standard components
• Single-user and collaborative • Large-model visualization• Real-time and distributed visualization
Storage
• I/O performance• File systems• Data management• Networking• Distributed data access
Delivering technology that enables significant scientific and creative breakthroughs
5
SGI in High Performance Computing
and BEYOND2000s1990s1980s
1982Jim Clark
founds SGI
1984IRIS®
Workstations
1994Challenge® XL server -
Steven Spielberg’s Shoah project
1988Power Series™ multiprocessing
1995Origin® 2000 and Indigo2™
help Team New Zealand win America’s Cup
1996 First-generation NUMA
System: Origin 2000
1998DOE 6144p Origin 2000 to
simulate nuclear stockpileASCI Blue Mountain
1995SGI introduces its first 64-bit operating system
• The industry's most scalable architecture – from 2 to 10,000+ processors• Shared memory scales up to 24 terabytes• Based on Intel processors and the Linux® operating system • Flexible, modular design that independently scales CPU, memory, I/O• Complete solution stack for High Performance Computing
2001Introduced modular
NUMAflex™ architecture with Origin® 3000
2003SGI Altix first scalable
64-bit Linux Server
2004First 512, CPU SGI Altix cluster drives ocean research at NASA
Ames and 10,000 CPUs upgrade
Image courtesy of NASA
2005SGI introduces Altix
330 Server and RASC™
6
FY03 FY04
800+Processors Sold!
3,300+ Processors Sold!
62,000+ Processors Sold!
NASA
JAERI
NCSAFord
Total
GFDLSARA
SGI® Altix® Market Momentum
FY05 FY06
DoDMod
GFDL
APAC
LRZ
Dresden
CSCS
IFS
IMS Yukawa
Pro
cess
ors
So
ld
SGI ® Altix ® 3000 Server Announced 1/03
SGI ® Altix ® 350 Server Announced
SGI ® Altix ® 3000 BX2 Announced
SGI® Altix® 330, RASC™ Announced
HPCWire Awards For 2004Best Price/performanceMost Innovative Hardware Technology Most Innovative Visualization Technology Most Innovative StorageBest Govt/industry collaborationMost Innovative Implementation
7
SGI® Altix® Server High-Performance Architecture
Global Shared Memory
...CPU CPUCPU CPU CPU CPU CPU
Fast NUMAlink™ interconnect technology• All processors operate on one large
shared-memory space• Industry’s highest bandwidth
interconnect at 6.4GB/second• High performance, low cost,
easy to deploy
• NUMAflex™ global shared memory and ultra-high bandwidth interconnect
• Modular, expandable architecture–processors, I/O, memory
• SGI ProPack™ software feature for Linux® OS optimizations for HPC
SGI® Altix® Family
8
Sca
le u
p(p
roc
es
so
rs p
er
no
de
)
100s
10s
2s
10s 1,000s
• Mix of applications• Unpredictable workloads• Growing job size
ApplicationComplexity
• Number of users• Number of jobs
Single-job, single-user Processing Capacity
SGI® Altix® Servers
Scale out(Total number of processors)
SGI® Altix® 350Departmentalservers(4–32 CPUs)
SGI® Altix® ClustersLarge nodeclusters
SGI® Altix® 3700 Bx2NEW: SGI® Altix® 4000 Supercomputers
SGI® Altix® 330Low-costworkgroup(1–16 CPUs)
9
SGI® Altix® High-end Servers
• Small-footprint packaging for high density, reduced costs
• Up to 512 processors under one OS, single system image
• Up to 2,048 total processors per NUMAlink™ system
• Best-in-class NUMAlink™ 4 routers with new topologies
• Water-cooled door option for larger configurations
• NEW SGI® Altix® 4000 extends scaling up to 8192 core NUMAlink™’d systems and 128TB of global shared memory in a functional blade design
10
SGI® Altix® 3700 Bx2 Building Blocks
Linux™
Operating
Environment
M-BrickMemory
Intel® Itanium® 2 Processor CR-Brick CPU and memory
IX-Brick, IA-BrickBase I/O module
PA-Brick, PX-BrickPCI-X expansion
R-BrickRouter interconnect
D-Brick2Disk expansion
11
Altix® Customers with Memory >500GB
• 98 systems with >500GB memory– 39 with > 1TB
• Benefits of Large Shared Memory– Massive In-core Computation,
Memory-Mapped I/O Lead to Breakthrough Results
– Unmatched Ease of Scaling from Efficient MPI, Trivial Load Balancing and No Data Replication
– Up to 512P, 6TB under one OS Means Ease of Development and Administration
* Only publicly referenced customers are listed.
Customers* with >1TB Memory
Customers* with 500GB to 1TB Memory
APAC APAC
GFDL Army Research Lab
JAERI BP Amoco
Japan Inst of Statistical Mathematics
GFDL
Los Alamos National Lab LRZ
McLaren Motor Racing Ltd. NASA Goddard
NASA Ames SARA
NASA Goddard Saudi Aramco
National Center for Supercomputing Apps.
Tata Motors
Naval Research Lab Total SA
Oak Ridge National Lab U of MN Supercomputing Inst
Total SA U Montreal
US Air Force UNC Chapel Hill
U Manchester US Government
12
SGI Altix® 4000 Platform:Extending Performance Leadership
•Superior Performance Density: up to nearly 3X density improvement over Altix 3700 Bx2
•‘Plug and solve’ Versatility: 8 standard functional blade choices, high-speed NUMAlink™ backbone supports top performance for any HPC application
•Best Performance on Cluster Applications: SGI NUMAlink™ interconnect beats the competition with sub-microsecond latency
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
13
SGI® Altix® 4000 Platform:Step into Multi-Paradigm Computing
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
L 1 Display
• Multi-Paradigm Computing: Taking HPC to the next level with integrated computational resources that can be seamlessly accessed as application workload requires.
• SGI Altix 4000 Resources Tightly Integrated with Peer I/O: Direct, high-speed connection of integrated graphics, RASC/FPGA and other future processing elements to global memory address space
• Takes HPC to the Next Level – beyond the limits of Moore’s Law and parallelism
L1 Display
Fil
ler
Pan
el
Fil
ler
Pa
ne
l
Bla
de
Slo
t 1
Bla
de
Slo
t 2
Bla
de
Slo
t 3
Bla
de
Slo
t 4
Bla
de
Slo
t 5
Bla
de
Slo
t 6
Bla
de
Slo
t 7
Bla
de
Slo
t 8
Bla
de
Slo
t 9
Bla
de
Slo
t 1
0
I/O
Bla
de
s
CP
U
Bla
de
Gra
ph
ics
Bla
de
RA
SC
B
lad
e
Mem
ory
B
lad
e
14
One Giant Leap for NASA
Deployment20 systems, each:
>512p, 1 TB memory, 1 Linux kernel
>440 Tbytes CXFX SAN storage
>SAN with >1 petabyte managed by DMF
The Challenge:
> Deliver a world-class supercomputing platform to NASA in support of their world-class scientific and engineering problems
> Provide the environment needed for NASA to perform simulations to support Space Shuttle Return to Flight
The Solution:
>Deploy a 10,240 processor supercomputer in a record 15 weeks. Ranks as 3rd most powerful computer in world.
15
SGI® Altix® 350: Powerhouse Midrange Server
• Provides superior performance and scalability from 1 to 32 processors
• Best price/performance midrange server
• Ideal for departmental application servers, technical database, throughput clusters
• One Linux® instance to manage
• Cluster to thousands of processors using industry-standard interconnects
• Independently scale CPU, memory, I/O– Investment protection and leverage current
assets
– Allocate budget and resources to ongoing needs
– Right-size systems for the ultimate price/performance
16
Ultra-dense,
Ultra-affordable,
Ultra-powerful
SGI® Altix® 330: Entry Midrange Price/Performance Leader
NUMAflex™Architecture
Linux® OS
Intel® Itanium® 2Processor
HPC Heritage
SGI HPC Expertise
Advanced Developer Tools
Scalable Architecture
SGI® Altix® Family:
Ultra-high Density Package
Great Price Point
Combined With: Results:
SGI® Altix® 330 Server
17
SGI® Altix® Midrange: Powerful, Affordable, Flexible
• Advanced scalability:– Altix 350 scales to 32 processors,
384GB of memory– Altix 330 scales to 16 processors,
128GB of memory
• Expand-on-demand modularity for perfect right-sizing & cost effectiveness
• Best in class price/performance
• Altix 330 provides affordable entry point, development platform
• Based on award-winning SGI® NUMAflex™ architecture for superior performance and scalability from 1 to
“Expand-on-Demand” Growth Path
1-32P/2-384GB
1-12P/2-144GB
Base Unit: 1 or 2P/2-24GB
1- 4P/2-48GB
1-8P/2-96GB
CPU Expansion Module
Memory Expansion Module
I/O Expansion (4-32) Module
18
Accelerating Research at the University of Minnesota Supercomputing Institute
DeploymentAdded a 256P Altix 3700 system and 18P SGI Altix 3700 system featuring Intel® Itanium™ 2 Processors with 512GB of memory to existing SGI Altix infrastructureAdded a 128P (8 x 16P) SGI Altix 350 cluster with Voltaire InfiniBand interconnectIntegrated new cluster with SGI InfiniteStorage SAN Solution with CXFS™ Shared Filesystem software and 3TB fibre channel RAID
The Challenge:
>Provide computational power for researchers in a broad range of disciplines, including computational genetics, biology, and atmospheric sciencesCost efficiently acquire a supercomputing class, shared memory computing solution
The Solution:
> The 256P SGI Altix system is helping to break new ground in a broad range of disciplinesThe University obtained a 128P cluster with lower interconnect and fabric management costs than with other 2P or 4P node systems
Image Courtesy of Dr. Shuxia Zhang at the Minnesota Supercomputer Institute.
“Shared memory architecture is widely acknowledged as ideal for large-scale scientific computing tasks…It undoubtedly will
lead to significant improvements in our already high productivity.DONALD TRUHLAR, SUPERCOMPUTING INSTITUTE DIRECTOR
AND PROFESSOR OF CHEMISTRY
19
SGI® RASC™ Technology
•Reconfigurable Application Specific Computing
•Extends SGI’s leadership in high performance computing
•Delivers unmatched performance, scalability and bandwidth for data-intensive applications
• Over 100X Improvements in mission-critical application acceleration
•Solution stack comprised of industry-leading tools and libraries
20
RASC™ Technology — Demonstrated Application Speed-up
Bit Manipulation (Cryptography)1
• 79x 1.5GHz Intel® Itanium® 2 Processor (single RASC Unit)• 119x 1.5GHz Intel® Itanium® 2 Processor (dual RASC Unit)
Customer Application• 20,000x speedup on scalar microprocessor
Graphics Edge Detection1
• 7.4x 1.5GHz Intel® Itanium® 2 Processor (single RASC Unit)
1 Based on internal testing
21
SGI® Altix® and DatabasesA Wealth of Options
SGI Altix Database Options• MySQL®
• Empress RDBMS• Sybase® Adaptive Server® Enterprise• Oracle9i™• IBM® DB2®
• IBM® Informix®
• Objectivity/DB™
• Eliminates data bottlenecks with blazing I/O throughput (6.4GB/second)
• Faster, more complex analyses through scalable database processing
• Full database workload placement in shared memory within a cluster system
• Cost-efficient database-in-memory by allowing user to add incremental memory as required
• Based on the industry-leading database performance processor (Intel® Itanium® 2 Processor)
1-32 Intel Itanium 2 Processors1 Linux®
Global Shared Memory
SGI ® NUMAlink ™
Shared Storage
22
Altix-ok itthon
• Országos Meteorológiai Szolgálat (Altix 3700, 144 CPU)• Szegedi Tudományegyetem (Altix 3700, 48 CPU)• ELTE Szerves Kémia Tsz. (Altix 350, 16 CPU)• Collegium Budapest (Altix 350, 16 CPU)• MTA Atomenergia Kutatóintézet (Altix 350, 12 CPU, 96GB RAM)• MTA SZBK Enzimológia Intézet (Altix 350, 12 CPU)• Semmelweis Egyetem Biofizika Intézet (Altix 350, 6 CPU)• Duna Televízió (Altix 3300)• Nemzeti AudioVizuális Archívum (Altix 350)• Silihost (Altix 350)
23
Summary
• SGI® proven leader in High Performance Computing, with industry-leading technology and expertise
• SGI® Altix® platform based on SGI® NUMAflex™ architecture, with ability to scale from 1 to 10,000+ processors and terabytes of memory
• Unified Altix family of servers, supercomputers, and clusters provides the ability to "start small" and grow to meet any processing requirements
• Investment protection, with an open standards-based system that supports all major HPC applications
Thank You
Thank You