The New York-based Institute is the internal research division of the Simons Foundation. The researchers here innovate in the field of Biological Sciences, Astrophysics, Quantum Physics and Computational Mathematics. Their contributions range from breakthroughs in genomic studies to nuances related to black holes.
Massive Data Processing and HPC Requirement
The high-intensity research here generates rich, large data files. There also arises a necessity to simulate complex physical processes. For instance, the High-Performance Computing (HPC) infrastructure has to handle giant genomic sequence files to 100,000 small files in a single directory. In short, Petabytes of data have to be transferred quickly. For this, institutes on-site cluster – a central HPC infrastructure has around 17,000 Intel computing cores in Lenovo ThinkSystem SD530 servers. They run HPC and AI works.
What Powers the System?
Researches are carried out with the help of Lenovo ThinkSystem SD530 Dense Rack servers with Lenovo Neptune Thermal Transfer Module (TTM) technology running leading-edge Intel Xeon Platinum 8268 processors, part of the 2nd Gen Intel Xeon Scalable family featuring Intel Deep Learning Boost technology.
Flatiron needed an intensively air-cooled system with vast amounts of memory as they have space constraints. “With our Lenovo HPC supercomputer cluster, we have four or five times more memory than a typical HPC farm,” Dr Ian Fisk, Scientific Computing Core Co-Director at the Flatiron Institute. He was referring to the additional RAM available with each node.
“Reading a 200 GB genomic sequencing file for a Biology project or making 11,000 connections to other nodes for an Astrophysics project requires a lot of memory. Now, when a researcher comes along with an interesting challenge that their equipment doesn’t support, we can achieve that,” he said.
How This Helps?
Prior to the installation of these systems, the scientific progress of researches was limited by system constraints. This affected the speed and scope of their research.
“Cosmological simulations like galaxy formation and black hole creation require hundreds of thousands of cores connected to each other to run. Science should dominate our researchers’ time, not computation. Our goal is that the limiting factor in research should be the carbon-based systems rather than the silicon-based systems,” said Dr Ian Fisk, Scientific Computing Core co-director at the Flatiron Institute.
Earlier, Scientists could sequence only around two per cent of genomic data. However, now they can look at the entire genomic sequence of thousands of families at the same time. This progress can be the key to a more effective discovery of genes that cause disease or the development of precision medicine.
Lenovo and Intel have recently extended their collaboration in the data centres with a multiyear commitment. They aim to play a crucial role in the fast-growing convergence of HPC and AI solutions.