The Qlustar HPC Stack contains all components necessary, to render a networked server farm based on off-the-shelf hardware into a high-performance compute cluster.
Qlustar is designed to handle small, medium and even the largest clusters with minimum effort but maximum control and efficiency. It shows, that more than a decade of experience in the creation, management and operation of powerful Linux clusters has contributed to its underlying technology.
Efficient and scalable workload management is of crucial importance for an HPC cluster. Users expect their compute jobs to get executed as fast as possible and want to get a fair share of the available compute resources. They also want detailed information about their jobs status and convenient tools to manipulate them. On the other hand, cluster administrators are most interested in maximum utilization of their cluster hardware and need simple control over the status of compute nodes.
Qlustar’s workload management provides the premier open-source software Slurm to achieve all of the above goals. To complement Slurm’s superb feature set, QluMan provides a powerful GUI management front-end to Slurm.
Qlustar HPC Core stack
To achieve scalable and optimized compute performance paired with reliable operation on a compute cluster, the base HPC related software components must be well tuned and matched up. The Qlustar HPC Core Stack is comprised of a handful software packages that together with the kernel drivers provide the foundation of the HPC application stack. We take the stance that the borderline between the latter and the core stack is defined by the MPI software packages.
The use of graphical processing units (GPU) to accelerate certain compute workloads has revolutionized HPC in recent years. The performance boost from parallelization on a large numbers of GPU cores increased the achievable FLOPS per node dramatically.
Qlustar has out-of-the-box support for the most widely adopted hardware options and programming models in GPU computing. In particular, the Qlustar Nvidia image modules provide up-to-date integrated drivers for Nvidia GPUs supplemented by matching CUDA versions.
Fast networking is a must for HPC clusters. The de-facto standard technology for high-speed networks nowadays is Infiniband (IB). To optimally support IB networks, Qlustar always ships an up-to-date MLNX OFED software stack with its drivers readily built to use with Qlustar kernels. This ensures that modern IB hardware is optimally supported and that new user-space IB software modules are available for use. Wherever possible, Qlustar integrates IB support in other HPC components like MPI or workload managers.