Home Our Clients Clients case study

Sunday16 December 2018

case study

duke-nus-large


CUSTOMER BACKGROUD

The beginnings of the Duke-NUS medical school can be traced to 2000, when Singapore launched an ambitious Biomedical Sciences Initiative designed to make the country the biomedical hub of Asia. Duke-NUS Graduate Medical School was formally established in 2005 when Duke University and National University of Singapore signed an agreement under which the two institutions would partner to establish a new medical school. Since then Duke-NUS with its sophisticated state-of-the-art research facilities is building leaders in medical research, education and patient care.

Customer application: 3D reconstruction of dengue virus with different receptors and large simulation experiments. Application is time and computing power consuming and requires a lot of computing power.


BUSINESS NEED

Provide high-performance computing system that can efficiently speed up dengue virus research process and will accelerate dengue virus vaccine creation.


SOLUTION

In order to facilitate the development of new and more effective dengue vaccines and therapeutics, the research team of the Emerging Infectious Diseases (EID) program at Duke-NUS needed a high performance computing solution.

Customer compared and tested several different architectures before deciding to continue its software application development for NVIDIA Tesla GPUs architecture as Tesla GPUs gave him the superior performance per dollar invested. Customer was evaluating multiple vendors through the open tender before choosing NOVATTE as its HPC provider. NOVATTEs’ flexible local testing facilities that allowed to built PoCs and conduct initial testing, in-depth HPC knowledge and experience of NOVATTE engineers, local same-hour technical support and NOVATTEs’ direct access to the latest Intel and NVIDIA technologies that allowed customer to evaluate future technologies and plan software coding for those technologies to maximize software performance were among the major strengths NOVATTE had over the peers and positively influenced customers’ decision making process.

Highly parallelized customer software code allowed NOVATTE to use X4000 and even X8000 GPU servers with just released at that time Tesla 2050s in its G10K with unified InfiniBand architecture allowing further HPC system scale outs. Further access to the latest Tesla GPU architectures and further testing allowed the customer to provision it’s coding accordingly and later painlessly scale the HPC system with Tesla 2070, Tesla 2075 cards and most recently with Tesla K10s. NOVATTE is continuing to work closely with NVIDIA to expand the customer’s HPC cluster further with Tesla K20s.


BENEFITS

- Customer can now do more science every single day improving research efficiency and productivity
- NOVATTE unified HPC architecture allows outstanding scalability for the future performance increase
- Increased system uptime
- Peace of mind with 3 years technical support and comprehensive after-sales service

JA Minisite

2857.orig.q75.o0 - Copy IntelTPP amd    qctlogo e   Mellanox APAC Partner   1