presented by leadership computing facility (lcf) roadmap buddy bland center for computational...
TRANSCRIPT
Presented by
Leadership Computing Facility (LCF) Roadmap
Buddy BlandCenter for Computational Sciences
Leadership Computing Facility Project
2 Bland_JaguarRoadmap_0611
Outline
Systems
Facilities upgrade
Systems infrastructure Overview Networking Storage
Software and science
2 Bland_JaguarRoadmap_0611
3 Bland_JaguarRoadmap_0611
SystemsCCS firsts (1991–2008)
4 Bland_JaguarRoadmap_0611
Facilities upgradePreparing computer center for next generation of Jaguar
Floor raised by 1 ft
New chilled water pipes to support up to 7 MW power
New air-handling system
New power distribution units
5 Bland_JaguarRoadmap_0611
Systems infrastructure—overviewCurrent and projected
External B/W (GB/s)
LAN B/W (GB/s)
Networking
5
240
FY 2009FY 2008FY 2007
4
140
3
60
Capacity (PB)
Bandwidth (GB/s)
Archival storage
18
19
FY 2009FY 2008FY 2007
10
10
4
4
Central storage
10.0
240
FY 2009FY 2008FY 2007
1.0
60
0.22
10
Capacity (PB)
Bandwidth (GB/s)
6 Bland_JaguarRoadmap_0611
Shifting to a hybrid InfiniBand/Ethernet network
InfiniBand-based network helps meet the bandwidth and scaling needs for the center
Wide-area network will scale to meet user demand using currently deployed routers and switches
Systems infrastructure—network1000 TF
240 GB/s LAN5 GB/s WAN
2007
2009
2008
1000 TF240 GB/s LAN5 GB/s WAN
100 TF60 GB/s LAN3 GB/s WAN
7 Bland_JaguarRoadmap_0611
Consistent planned growth in ORNL external network bandwidth
Systems infrastructure—network
ORNL and LCF Backbone Connectivity
8 Bland_JaguarRoadmap_0611
HPSS software has already demonstrated ability to scale to many PB
Add two silos/year
Tape capacity and bandwidth, disk capacity and bandwidth are all scaled to maintain a balanced system
Use new methods to improve data transfer speeds between parallel file systems and archival system
Systems infrastructure—storageArchival storage
1000 TF18 PB
19 GB/s
2007
2009
2008
1000 TF10 PB
10 GB/s
100 TF4 PB
4 GB/s
9 Bland_JaguarRoadmap_0611
Increase scientific productivity by providing single repository for simulation data
Connect to all major LCF resources
Connect to both InfiniBand and Ethernet networks
Potentially becomes the primary file system for the 1000 TF system
Systems infrastructure—storageArchival storage
1000 TF10 PB
240 GB/s
2007
2009
2008
1000 TF1 PB
60 GB/s
100 TF250 TB10 GB/s
10 Bland_JaguarRoadmap_0611
Software and science—overview Cutting-edge hardware lays the foundation for
science at the petascale—scientists using a production petascale system with petascale application development software Establishing fully integrated computing environment Developing software infrastructure to enable productive
utilization and system management Empowering scientific and engineering progress and
allied educational activities using petascale system Developing and educating the next generation of
computational scientists to use petascale system
CCS Management Plan coordinates the transition to petascale production
10 Bland_JaguarRoadmap_0611
11 Bland_JaguarRoadmap_0611
Software and scienceRoadmap to deliver Science Day 1
2006 2007 2008 2009
50 T 100 T 250 T 1000 T
expand quad-core
Establish CFS Lustre Center of excellence at ORNL
quad-core Linux LWK
quad-core Catamount
Baker LWK
petascale LWK
dual-core LWK
SIO Lustre clients and external cluster of Lustre servers
XT4 LW IO interfaceMPI-Lustre failover LCF-wide file system
Hardware supervisory system Mazama system
Scale tests using Jaguar
Scale tests using Jaguar expand
Fault monitor and prediction system
C/R on XT3
dual-core
on 100 TF on 250 TF on 1 PFScience day 1
Testbeds
OS (key releases)
File system
RAS
HW
12 Bland_JaguarRoadmap_0611
Science–drivers
Advanced energy systems (e.g., fuel cells, fusion)
Biotechnology (e.g., genomics, cellular dynamics)
Environmental modeling (e.g., climate prediction, pollution remediation)
Nanotechnology (e.g., sensors, storage devices)
“Computational simulation offers to enhance, as well as leapfrog, theoretical
and experimental progress in many areas of science and engineering…”
— [A Science-Based Case for Large-Scale Simulation (SCaLeS Report), Office of Science, U.S. DOE, July 2003]
— [A Science-Based Case for Large-Scale Simulation (SCaLeS Report), Office of Science, U.S. DOE, July 2003]
12 Bland_JaguarRoadmap_0611
13 Bland_JaguarRoadmap_0611
Software and science—fusion
14 Bland_JaguarRoadmap_0611
Software and science—biology
15 Bland_JaguarRoadmap_0611
Software and science—climate
16 Bland_JaguarRoadmap_0611
Software and science—nanoscience
17 Bland_JaguarRoadmap_0611
Contacts
Arthur S. BlandLeadership Computing Facility Project DirectorCenter for Computational Sciences(865) [email protected]
17 Presenter_date