NVIDIA Data Centre GPUs

- View product

# CUDA Cores:
10752
# RT Cores:
84
# Tensor Cores:
336
Interfaces:
3x DisplayPort 1.4
Manufacturer:
PNY
Memory:
48GB GDDR6 with ECC
Passive Cooling:
Yes

View product

NVIDIA Data Centre GPUs
PNY NVIDIA A16 64GB Passive Ampere GPU

# RT Cores:
4x 10 (2nd Gen)
# Tensor Cores:
4x 40 (3rd Gen)
GPU Memory Bandwidth:
4x 232GB/s
Manufacturer:
PNY
Max Power Consumption (W):
250W
Memory:
64 GB GDDR6 ECC (16 GB per GPU)
Passive Cooling:
Yes
PCIe:
4.0 x16

£2,253.99 exc. VAT
The NVIDIA A16 GPU bring next-generation performance to your Virtual Workstations enabling your users to tackle any project from anywhere. When combined with the NVIDIA Virtual PC software, the NVIDIA A16 is at its best. This highly optimised GPU for graphics-rich virtual desktop infrastructure (VDI) is based on the latest Ampere architecture and delivers twice the user density over the…

- View product
- View product
- View product
- View product
- View product
- View product
- View product
- View product

Manufacturer:
PNY
Memory:
24 GB HBM2
Passive Cooling:
Yes
Peak FP32:
10.3 TFLOPS
Peak FP64 Tensor Core:
10.3 TFLOPS
Peak FP64:
5.2 TFLOPS

# RT Cores:
72
FP32:
31.2 TFLOPS
Manufacturer:
PNY
Memory:
24GB GDDR6
Passive Cooling:
Yes
TF32 Tensor Core:
31.2 TFLOPS

# CUDA Cores:
5120
# Tensor Cores:
640
GPU Architecture:
Ampere
GPU Slot Size:
Dual
Memory:
40GB HBM2
Passive Cooling:
Yes

# CUDA Cores:
2560
# Tensor Cores:
320
GPU Architecture:
Turing
GPU Slot Size:
Single
Memory:
16GB GDDR6
Passive Cooling:
Yes

# CUDA Cores:
5120
# Tensor Cores:
640
GPU Architecture:
Volta
GPU Slot Size:
Dual
Memory:
32GB HBM2
Passive Cooling:
Yes

# CUDA Cores:
3840
GPU Architecture:
Pascal
GPU Slot Size:
Dual
Memory:
24GB GDDR5
Passive Cooling:
Yes

# CUDA Cores:
4096
GPU Architecture:
Maxwell
GPU Slot Size:
Dual
Memory:
16GB GDDR5
Passive Cooling:
Yes

# CUDA Cores:
2560
GPU Architecture:
Maxwell
GPU Slot Size:
Dual
Memory:
32GB GDDR5
Passive Cooling:
Yes

Leverage the power of NVDIA® Data Centre GPUs to accelerate hyperscale datacentre and intensive HPC workloads. Now scientists can process extensive datasets faster than ever before, opening up a new world of possibilities for research and development in oil & gas, medical research, facial recognition, artificial intelligence and many more.

Not only do NVIDIA® professional GPUs deliver the fastest means of completing complex simulations available today. They also excel in high user density virtual desktop ecosystems when combined with NVIDIA virtual GPU software.

Designed Specifically For Server Platforms

NVIDIA® Data Centre GPUs are passively cooled designs that integrate perfectly into our high-density rackmount server solutions, enabling users to achieve supercomputer performance levels within a 1U, 2U or 4U footprint. With up to 10x GPU compute cards integrated into a single system, configurations benefiting from NVIDIA’s NVLINK technology and the ability to cluster multiple systems to work together on tasks.

Data scientists, researchers & infrastructure managers will be able to power and accelerate their workflows, whether that be Deep Learning training/inference tasks, virtualising a workforce or needing to crunch large datasets within supercomputing simulation applications.

Accelerate Deep Learning Workflows

There are two key phases to Deep Learning workflows, and each has its requirements to achieve the most efficient computation times that data scientists demand. Training is the first stage, and this highly demanding stage can make full use of multiple NVIDIA® Data Centre GPUs to reduce computation times. With this approach, a neural network is trained in a matter of hours when traditionally it took months to on a CPU.

Once training is complete, inference operations of new data points within a trained neural network enable the solution to generate real-world benefits. Artificial Intelligence (AI) services can now take video, images and speech as inputs and return valid answers or recommendations. This simple yet challenging concept is now a reality and changing the world around us. It is all made possible due to the acceleration NVIDIA® Data Centre GPUs provide.

NEW ERA IN HIGH-PERFORMANCE-COMPUTING (HPC)

HPC workloads have traditionally only been possible by utilising large CPU-based systems clusters within a Data Centre. This approach can be high in costs and, due to its inefficient design, offers diminishing returns in performance as clusters grow in size. Furthermore, due to the high costs associated with this approach, HPC clusters were very limited to only large businesses and institutions that had the resources to implement such a solution.

GPU accelerators such as those designed for Data Centres from NVIDIA® have changed the High-Performance-Computing landscape by opening up the ability to perform HPC simulation and analysis to a much larger audience. A single GPU compute server can achieve what used to require dozens or even hundreds of clustered CPUs at a fraction of the cost. With over 600 High-Performance-Compute (HPC) applications now fully supporting GPU acceleration, the possibilities are ever-growing.

Virtualise Your Workforce

How businesses operate is evolving. People are demanding more flexibility in their working lives, and they shouldn’t be held back by the need to stay in a fixed location because of computing resources. NVIDIA virtual GPU solutions powered by NVIDIA® Data Centre GPUs are the perfect solution to this issue.

NVIDIA virtual GPU (vGPU) software in conjunction with NVIDIA® Data Centre GPUs enable you to create virtualised graphics workstations that run on any device, anywhere. These solutions give IT administrators complete freedom to efficiently allocate resources for any workload, quickly expand their infrastructure to add additional users and obtain the highest user density currently possible from a VDI solution.

NVIDIA® Tesla® GPUs Are Ideal For

Why Choose Workstation Specialists

Quick Turnaround

We are able to guarantee an industry leading turnaround time of 5-7 working days on the majority of our system orders by holding stock of key components. In addition only reliable supply lines are used to obtain remaining components.

High Quality Components

We track the reliability of every component we use and if that item is not up to our high standards we will not sell it. Therefore you can be confident that buying any solution from us will meet the stability and reliability you and your business require.

Here to Help You

You will receive a dedicated account manager responsible for resolving any questions or enquiries you may have. When contacting us via phone or email you will be greeted by a real person ready to assist you.

Hand Built by Experienced Engineers

Each of our systems are hand built to order by a dedicated highly trained engineer. This enables us to guarantee the quality of each of our solution builds and allows us to supply a product you can rely on.

Rigorous Testing Procedures

Every system has to pass a minimum 48 hours of burn-in to stress components to their maximum to identify any faults. A thorough Q&A procedure is also used to independently verify build quality. As a result we are confident in the robustness of every system that leaves our workshop.

Warranty Packages Designed To Fit Around You

Support that goes above and beyond. Our standard warranty package includes 36 Months Premium RTB Hardware Warranty with Remote Engineer Diagnostics by Next Business Day. If you require something extra then labour only extensions and on-site options can be tailored around you.

Shopping Cart

NVIDIA Data Centre GPUs