Date post: | 17-Jan-2016 |
Category: |
Documents |
Upload: | louisa-harrison |
View: | 213 times |
Download: | 1 times |
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
High Performance Computing
at NCAR
Tom BettgeDeputy Director
Scientific Computing DivisionNational Center for Atmospheric Research
Boulder, CO [email protected]
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Outline
• Current Events / News
• Current Computing Capacity at NCAR
• Future Computing Capacity at NCAR
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Current Events / News
• IBM Power3 blackforest decommissioned Jan 10 (yesterday!)
• IBM e325 Linux Cluster lightning begins production Feb 1
• Machine Room Shutdowns:– Feb 24-27: Chiller Upgrade Phase II– May (1 day): Chiller Upgrade Phase III
• Introduction of LSF to manage batch submissions, scheduling, and accounting (not bluesky).
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Current HPC Environment….
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Peak TFLOPs at NCAR
0
2
4
6
8
10
12
Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05
IBM Opteron/Linux
IBM POWER4/Federation(thunder)
IBM POWER4/Colony(bluesky)
IBM POWER4 (bluedawn)
SGI Origin3800/128
IBM POWER3(blackforest)
IBM POWER3 (babyblue)
Compaq ES40/32(prospect)
SGI Origin2000/128 (ute)
HP SPP-2000/64 (sioux)
CRI Cray C90/16 (antero)
CRI Cray J90 series
Cray C90/16
HP SPP2000
SGI Origin2000
blackforestWH-1
blackforestWH-2
ARCS Phase 1blackforest upgrade SGI Origin3800
ARCS Phase 2bluesky
ARCS Phase 3bluesky expansion
IBM Linux
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
New Linux Cluster: lightning
• Linux Cluster– 256 processors (128 dual node configuration)– 2.2 GHz AMD Opteron processors– 4 GB/node– Myricom Myrinet interconnect– 6 TByte FastT500 RAID with GPFS
• Performance Characteristics– 40% faster than bluesky (1.3 GHz POWER4) cluster on
parallel POP and CAM simulations– 75 Gflops on WRF benchmark (full system)
• Accounts– email [email protected]– provide short description of tasks, codes, job sizes
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Computing Demand
• Science Driving Demand for Scientific Computing
Summer 2004: CSL Requests 1.5x Availability
Sept 2004: NCAR Requests 2x Availability
Sept 2004: University Requests 3x Availability
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
• Supercomputers are well utilized ...
• ... yet average job queue-wait times* are measured in minutes, not hours or days
Sep’04 FY04
Bluesky 8-way LPARs
91% 88%
Bluesky 32-way LPARs
98% 93%
(Regular Queue) CSL Community
Bluesky 8-way
86m 31m
Bluesky 32-way
40m 34m
Servicing the Demand
* September 2004 average
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Future HPC at NCAR……
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
NCAR/SCD
1990 1995 2000 2005 2010
1
50
100
200
250
300
350
150
Posit
ion
Year1996
Procurement
IBMPower3
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
SCD Strategic Plan:High-End Computing
Within the current funding envelop, achieve a 25-fold increase over current sustained computingcapacity in five years.
SCD intends as well to pursue opportunitiesfor substantial additional funding for computationalequipment and infrastructure to support therealization of demanding institutional scienceobjectives.
SCD will continue to investigate and acquireexperimental hardware and software systems.
•IBM Linux Cluster •IBM BlueGene/L
(~ 4+ fold in 1Q2006)
1Q2005
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
SCD Target Capacity
Target Sustained Computing Capacity at NCAR
0
2
4
6
8
10
12
Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10
Su
sta
ined
Tera
FL
OP
s
Moore's Law
SCD Target
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Mass Storage Archival…..
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
NCAR MSS - Data Holdings
0
500
1000
1500
2000
2500
Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04
Ter
abyt
es ~18 years for1st Petabyte
Nov '02
18 months for2nd Petabyte
Jul '04
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Scientific Computing DivisionStrategic Plan
2005-2009
www.scd.ucar.edu
to serve the computing, research and data management needs of atmospheric and related sciences.
Supercomputing • Communications • Data
NCAR Scientific Computing Division
11 January 2005
Questions