GPU Servers
Our GPU servers utilise NVIDIA’s world-leading technology to achieve the pinnacle of high performance computing. Develop your own supercomputing cluster or build a virtualised VDI environment using industry-leading hardware from NVIDIA.
Massively parallel GPU dedicated server architecture and bare metal servers accelerate your CUDA applications and power some of the most advanced applications in the world to optimise workloads and save costs, or gain competitive advantages to drive revenue.
With the convenience of our automated provisioning and massively scalable, low-latency network, we provide server hosting to some of the world’s most powerful GPU clouds – choose from GRID, Tesla and GeForce GPUs.
See More
NVIDA’s GRID technology brings unprecedented performance to the delivery of graphics intensive applications, such as CAD and video transcoding, from the cloud. Designed for density, GRID is the ideal GPU for achieving supercomputer speed at the desktop, balanced with data centre economics.
Know More
When big data analysis, business intelligence or scientific modelling are the name of the game, NVIDIA’s Tesla helps you win. The world’s fastest GPU accelerators combined with high speed NVLINK interconnect technology make Tesla the ideal GPU for today’s computationally intensive workloads.
Test More
Nvidia’s GeForce range provides the same level of computing performance as the enterprise cards, but without official application support and optimised drivers. Ideal for testing and budget applications, the GeForce cards represent a fantastic opportunity to develop CUDA applications that do not require any virtualisation of the underlying physical GPUs.
Customise Your Supercomputing Cloud
Need an GPU-powered cloud environment with your specific combination of Operating System, CPU, RAM and storage tiers? No problem. We can design and build highly-customised, massively-dense GPU cloud environments for you to squeeze every last bit of performance from your hardware, and we can liaise with Nvidia directly to ensure that all aspects of the environment are balanced to deliver the best possible return on investment.
Manufacturing & Engineering
Offload computationally intensive 3D modelling and manipulation to the parallel processing power of our GPU servers. Enjoy faster simulation, rendering and iteration speeds in your CFD, CEA or EAD projects.
Medical & Life Science
Take your research data modelling to the next level with hosted GPU servers. Accelerate a wide range of iterative and analytical processes like gene sequencing, clinical results analysis and molecular dynamics to achieve quicker results and make faster discoveries.
Oil & Gas
Ensure productivity, efficiency and safety are optimised, and make better investment decisions, with the power of our GPU cloud. Use on-demand GPU resources to perform faster process engineering, simulations, seismic activity analysis and data processing.
Big Data & Analytics
Make better, real-time decisions by accelerating your big data analysis. Sorting, searching and finding the patterns in increasingly large data-sets can be performed significantly faster when supported with massively parallel GPU architecture.
Typical GPU-Ready Server Configurations
- Model
- GPU Cards
- CUDA Cores
- GPU RAM
- CPU
- RAM
- HDDs
- RAID
Entry Level
- 1x GRID K1
- 768 Cores
- 16GB GDDR5
- Dual Intel CPU
- Up to 128GB
- Up to 8x 3.5"
- Customisable
Level 2
- 2x Tesla K10
- 6,144 Cores
- 16GB GDDR5
- Dual Intel CPU
- Up to 128GB
- Up to 8x 3.5"
- Customisable
Level 3
- 4x Tesla K20X
- 10,752 Cores
- 24GB GDDR5
- Dual Intel CPU
- Up to 128GB
- Up to 8x 3.5"
- Customisable
Level 4
- 4x GRID K2
- 12,288 Cores
- 32GB GDDR5
- Dual Intel CPU
- Up to 128GB
- Up to 8x 3.5"
- Customisable
Extreme
- 8x Tesla K40M
- 23,040 Cores
- 96GB GDDR5
- Dual Intel CPU
- Up to 128GB
- Up to 8x 3.5"
- Customisable
The above table shows some options for both GRID and Tesla servers – typically the E5 Intel CPU range would provide more than sufficient computing power for any application, and DDR3 RAM is more than adequate to power the number of GPUs chosen for each chassis. All chassis chosen have been fully certified as compatible by Nvidia and if more than 8x GPU cards are provided in a server farm, we can advise on methods of clustering GPUs.