Upload
dokhanh
View
217
Download
4
Embed Size (px)
Citation preview
Building structures we live in
Designing aircrafts that we fly in Designing vehicles that we drive
Launching satellites that beam our TV channels
Making movies that we enjoy Designing medicines that cure us
Devising national security mechanisms
Forecasting climate that help in many ways
And Many More…
Complementing theory and experimentation,
supercomputing now offers the third mode of
scientific discovery and innovation
Switch
RC Accelerator
Vardha Software Stack
KSHIPRA Communication Substrate
288 Nodes, 4608 PEs, 200 TB Storage
System & Application Software
PARAMNet-3 Card
15
Biogene
RCS Card
Biocrome System
• Facility exclusive for bioinformatics research
• An enabling platform for collaborative research in Life sciences
• National HPC Grid Infrastructure (total
compute power: 75TF/s)
• Connects geographically distributed
HPC Resources of the Nation
• HPC Resources include HPC Systems,
Storage, I/O devices, Instruments &
sensors
• Has more than 65 partners from
Academic & Research Institutions
spread across ~25 cities of the country
• Enables Scientists & Researchers to
work in collaboration
• Applications of GARUDA include Open
Source Drug Discovery (OSDD) and
Virtual Science Laboratory
• Life Sciences – Drug Discovery & More
• Ocean, Earth Surface, and Atmosphere Modelling with Forecasting
• Computational Fluid Dynamics
• Disaster Management
Earthquake engineering
Flood/Cyclone/Tsunami forecasting and response
Modelling of spread of fire
• Discovery of New Materials – Nano science research
• Discovery of Natural Resources – Seismic Data Processing
• Discoveries Beyond Earth – Space Applications
• The simulations have been performed using the extended conformation as the starting structure.
• Total simulation time of 18µs has been achieved.
480 processors of Yuva used for approximately 6 months
S. Koulgi, U. Sonavane & R. Joshi Insights into the folding pathway of the Engrailed Homeodomain protein using replica exchange molecular dynamics simulations, J. Mol. Graph. Model. (2010), 29, 481-491
Publications
Objective: To forecast up to village level/point locations over India
Forecast available for 3 days advance in 6 hour interval over web as well as mobile.
Users: Agriculture Institutes, KSPB, Reuters, IAF, State Irrigation Departments, Farmers.
Pilot Simulation already done
• 4000 kW/m2 intensity fire with 4.2 m/s wind @ 45 deg
• 8 million mesh (0.4 x 0.4 x 0.4 m cell size) • ~ 120 s simulated in 40 hrs @ 18 cores of Param Yuva • Planned 30 million mesh (0.25 x 0.25 x 0.25 m cell size)
Estimated time ~ 300 s simulated in 4 days on 100 cores
• Completed using Grid pre-port processor
• Simulation of 1.2 lac elements of single pier system for
earthquake data was done on OpenSees using 96 processors of
PARAM Yuva
• Simulation of entire bridge and soil system is under progress.
• In collaboration with College of Engineering, Pune (CoEP)
• Parallel, Multicore and Manycore Programming
• System Administration & Management
• Storage Management Technologies
• Facility Operations Management and Maintenance
• GPU based Programming
• HPC User Symposiums
• C-DAC Certified HPC Professional
• ONAMA integrated package aids in understanding of
engineering disciplines through High Performance
Computing.
• It comprises of a well selected set of parallel as well as serial
applications, libraries and tools across various engineering
disciplines.
• Besides, it also consists of additional accelerated processing
unit (APU) enabled applications in specific domains.
• ONAMA provides the tools that have close affinity with
practical simulation, hence making the learning process a lot
more applied.
• It provides access to various APIs from several libraries viz.
parallel libraries, performance libraries.
• HPC plays a vital role in almost all critical areas and considered as a
medium to increase economic competitiveness of the nation.
• In India, HPC is at a developing stage with an acute shortage of HPC
professionals.
• Today, there is a dearth of academic institutions teaching or using
HPC.
• Factors like prohibitive costs, lack of support for adopting HPC and
little awareness about importance of HPC plays a pivotal role here.
• Onama takes on these issues head-on. By providing flexibility in
selection of HPC configuration, the user is given a control over the
cost factor.
• This encourages institutions, which earlier avoided adopting HPC,
to embrace HPC in totality. As a direct effect, the
student/researcher is exposed to HPC at an early stage.
• Moreover, each HPC professional is capable of catalysing the
process of spreading the HPC culture.
• Thus, when an increasing number of institutions adopt Onama, the
right environment for producing HPC skilled manpower is created.
In such a scenario, Onama is at the centre stage.
• The society is then poised to reap immense benefits from the
proliferation of HPC technologies to all the strata of society.
• In short – CAPABILITY BUILDING – BUILDING OF SKILLED
HPC AWARE MANPOWER