CPU/GPU High-performance Computing Cluster

Supercomputer for Compute-intensive Workloads

Robust HPC team builds world-class, high-performance, Infiniband interconnected HPC cluster for various industry verticals: oil & gas, media & entertainment, genomics, AI/Big Data, automotive and more.

Unlike many providers who only deliver the platform, Robust HPC delivers an end-to-end solution that includes support for user applications. You can trust our solution in powering your most compute-intensive workloads.

RenderPal

A professional Render Farm Manager for Windows, Linux and MacOSX, dedicated to network rendering across small to large render farms.

Munu

A powerful (5000+ nodes), highly customizable render farm platform that support industrial design software such as VRED.

IBM Spectrum LSF

A complete workload management solution for demanding HPC environments.

SLURM

An open source, fault-tolerant, and highly scalable cluster management and job scheduling system for large and small Linux clusters.

Render farm for VFX, Animation & Arch Viz

Robust Fox, our managed cloud render farm service is used by over 80% of the animation and arch viz studios in Malaysia, remains the top render farm choice for Malaysian. We are also the sole large-scale cloud render farm service in South East Asia region.

With the experience of operating an online render farm consists of 10,000 render nodes serving global clients and having a team of 30 experienced TDs who possess real production experience span across arch viz, animation and VFX industry, we have solid understanding not only in render farm hardware components, but also inside out of every bits of the 3D software and plugins that run on a render farm.

Our strong presence and proven expertise in this industry guarantee a peace-of-minds for studios or teams who are looking for setting up a render farm.
If you’re looking for a vendor for your render farm project, look nowhere else and give us a call now.

HPC Cluster for Data Science and Research

We provide a complete solution to build and run your HPC cluster, from single to tens of thousands of server nodes.

Planning and design – Our team guides you through in the designing of the head node, compute node, resource utilization.
Server hardware – We provide a highly customizable server with tested compatibility.
Infrastructure – Multi-gigabits network with either copper or fiber.
Workload scheduler deployment – Installation, configuration and fine-tuning.
Application UI development – Create or customize application UI for an intuitive and smooth user’s job submission.
Maintenance and Support – 8×5 or 24×7 L1 & L2 local support.

We work with the following software, including but not limited to:

Work scheduler: SLURM and IBM Spectrum LSF
User applications:
- Keras
- TensorFlow
- ANSYS
- Python
- MPI

VRED Cluster for Automotive Design

Connecting a VRED workstation to a cluster not only could speed up the render time, allowing more still images to completed in less amount of time, it also made possible to run real-time presentation of your design at high FPS with Raytracing and Antialiasing turned on.

However, it might not be easy for a designer or general IT team to setup the cluster as it involves a wide range of IT + Computer Graphics knowledge. Example it would generally includes making the cluster serving render farm function to CGI artist at the same time to yield a higher ROI. In that case, system components such a high performance storage cluster, multi-gigabits networking etc would be in place which further complicated the setup.

We are ready to assist you in setting up the cluster now or in the future. Best of all, we’re not just an ordinary system integration team, we also provide consulting and planning service which could collaborate with your team as earlier as in the idea stage to help you establishing user requirements and business use cases.

A bit of achievement:
In Jan 2019, we managed to run VRED presentation on a HTC Vive Pro over a cluster with Raytracing and Antialiasing turned on @ 12 FPS. This is never possible on a single workstation regardless of single or dual processors.

On-premise HPC Cluster