ALL FOR SCIENCE, SCIENCE FOR ALL
If you’ve been following us on our travails with NVIDIA, you’ll know that nearing the end of 2015, we were given a chance to take a tour of the CAVE2 facility, a giant display built and powered by NVIDIA’s own Quadro workstation GPUs. The facility, which is housed inside the Clayton campus of Monash University in Melbourne, Australia, was designed to aid academic researchers in their pursuit of science.
As one would expect, the backbone behind the CAVE2 is, of course, powered by not one, but two supercomputers, known simply as MASSIVE (short for Multimodal Australian ScienceS Imaging and Visualization Environment). However, back in February of this year, Monash University unveiled yet another supercomputer – known simply as the MASSIVE 3 (or M3), to further aid in their scientific endeavors in the field of medicine, diseases and illnesses.
Just like its two other supercomputer sibilings, the new M3 supercomputer is a beast of a machine. Where the CAVE2 was designed to create highly
HWM detailed renders of subjects, places, and the odd anatomical structure of a biological being, the M3 was primarily designed for the purpose of High Performance Computing (HPC).
The M3 is no slouch either. Like the previous two supercomputers, the M3 is powered by NVIDIA graphics cards. More specifically, the M3 runs on a combination of: • 1,700 Intel Haswell CPU-cores, • 50 NVIDIA Tesla K80 GPU coprocessors which are used for data processing and high-end visualization, • Eight NVIDIA GRID K1 GPUs for medium-end visualization that are made to support up to 32 users concurrently. To speed things up, the M3 also houses a 1.15 Petabyte Lustre parallel file system, which is essentially a storage device that is capable of reading data at 24 gigabytes per peak second. To put that into layman’s terms: that’s approximately four times the peak transfer speed of the current M.2 medium, which most high-end notebooks and gaming PCs use.
However, even when we went back to Monash University for the announcement of its new M3 supercomputer, we never really got the chance to check out the machine or its housing facility, on the grounds that the institution did not allow photos or videos of it to be taken.
HPC and Deep Learning in the realm of science has long been a driving force for academic institutions. These highly-complex facilities and GPU-powered supercomputers have been responsible in helping scientists in medicine and forecasting weather patterns, creating and reconstructing elements within the human body, as well as animals. These include the ability to replicate – with absolute intricacy – intimate details of the human nervous system, synaptic patterns of a brain, or even a complete render of mouse’s head.
The CAVE2 facility at Monash University just amassed a new supercomputer to play with.
(From left to right) Allen Finkle, Chief Scientist of Australia; Ian Smith, Vice Provost, Monash University; and Christina Mitchell, Dean of Faculty Medical, Nursing and Health Sciences, officiating the launch of the new MASSIVE 3 facility.
Sure, NVIDIA’s GPUs are the driving force of the M3, but look at the number of those Intel CPUs!