MASSIVE M3 DGX User Community Meeting

The MASSIVE team has recently expanded the GPU capability by adding five DGX-1 servers to the resource pool. We are close to completing the integration of these servers into the M3 queue, and we will be expanding this server partition in the future.

DGX-1 is a fully integrated machine with eight Nvidia V100 GPU cards and it's powered by NVIDIA NVLink™ interconnect to accelerate data processing and high-performance training in deep learning. Our benchmarking has demonstrated that these servers are outstanding at computational imaging (including CryoEM processing) and deep neural network workloads.

We are keen to talk to future users of these servers to present how they will be accessed, discuss requirements, and capture any feedback. Therefore, next week, we will be running an information and engagement session for future users of these servers. The purpose of this session is to:

  • Provide information about the purpose and the capability of these servers;

  • Outline the way users can gain access;

  • Discuss the requirements to gain access; and

  • Gather user feedback.

A particular point of discussion will be the requirements for the access, which will favour users that are able to make good use of eight GPUs in parallel.

Thursday 25th July at 11AM
Please email gin.tan@monash.edu to register.