${ selectedTariff.descr }
Sign up to our platform and get started training your neural networks
-
Ready-made infrastructure for working on ML, Data Science through JupyterLab and by SSH
-
Connection of the network storage and switching it between the containers
-
Dedicated team for project assistance
-
Flexible plans: per-minute billing and monthly subscription
-
Containers are provided with a ready-made software environment
-
Automated management with API
-
Frameworks for ML: PyTorch, Tensorflow
Graphics accelerator
for different needs of your business
NVIDIA TESLA A100
Multi-Instance GPU (MIG) feature allows GPU A100 to split up to seven GPU instances to solve different tasks simultaneously on one accelerator.
CUDA cores double the number of single-precision floating point operations (FP32).
40 GB of HBM2 high speed memory with the best in its class bandwidth of 1555 GB / sec.
Select configuration for needs of your business
Virtual containers configuration for rent
Total
${ formatPrice(dboard.total, 0) } ₽ / per ${ selectedTariff['totalScaleText'] }
the price is indicated including VATContainer with 1/7 of A100 (MIG) х${form.nv_a100_mig} un.
${ formatPrice(dboard.per_moment.gpu) } ₽ / ${ selectedTariff['momentScaleText'] }
${ formatPrice(dboard.gpu) } ₽ / ${ dboard.month_days } days
Container with one NVIDIA Tesla A100 accelerator х${Math.floor(form.nv_a100_mig / 7)} un.
${ formatPrice(dboard.per_moment.gpu) } ₽ / ${ selectedTariff['momentScaleText'] }
${ formatPrice(dboard.gpu) } ₽ / ${ dboard.month_days } days
Container without GPU x${form.nogpu} un.
${ formatPrice(dboard.per_moment.nogpu) } ₽ / ${ selectedTariff['momentScaleText'] }
${ formatPrice(dboard.nogpu) } ₽ / ${ dboard.month_days } days
Storage size ${form.storage} GB
${ formatPrice(dboard.per_moment.storage) } ₽ / ${ selectedTariff['momentScaleText'] }
${ formatPrice(dboard.storage) } ₽ / ${ dboard.month_days } days
Grant for the platform HPC Park Cloud Service
Free one-month trial period
Get a Grant of 30,000 rubles and try out the HPC Park Cloud Service platform's features without paying for the resources consumed.
Within the trial period you will have access to:
Containers with 1/7 of A100 (MIG) up to 28 units.
Containers with entire Nvidia A100 up to 4 units.
Containers without GPU up to 4 units.
Storage size up to 2000 Gb.
Containers with 1/7 of A100 (MIG) up to 28 units.
Containers with entire Nvidia A100 up to 4 units.
Containers without GPU up to 4 units.
Storage size up to 2000 Gb.
Сomplete the contact
form below
Our specialist will
contact you soon
Hpc Park cooperates with legal entities
Thank you, your application submitted! We will contact you soon.
Loyalty program
USE OUR SERVICE AND GET AN ADVANTAGEOUS OFFER
Discount level
Name
Discount
Deposited amount
Bronze
Silver
Gold
Platinum
Diamond
Почему стоит выбрать HPC Park Cloud Service?
Virtual servers are provided based on CPU and GPU. You can customize and change the configuration of the virtual server to suit your needs.
You can increase or decrease resources depending on the actual needs of your business at a transparent and understandable price.
It's provided a lot of security solutions such as virtual firewalls, role-based permission control, internal network isolation, virus protection, and traffic throttling.
Services can be delivered within minutes, enabling rapid deployment and faster time-to-market.
You can use high-performance resources in data centers and servers in the Russian Federation.
You can use HPC Park's Cloud Service to access other cloud services.
You don't need to purchase hardware or build data centers.
A standardized API is provided to improve ease of use and applicability.
Compliance with Russian legislation (152-FZ).
Partners
HPC PARK IS THE CURRENT PARTNER OF THE WORLD'S LEADING MANUFACTURERS
Our partners are reliable suppliers of information solutions. We are always glad to cooperate and invite you to join the community of technical specialists, who are focused on a progress and development.
Frequently asked questions (FAQ)
Given your computing requirements for a particular business task, you can increase or decrease resources. See the Guides for details.
You can create, delete and manage containers both through the web interface of the User’s Personal Account and automatically using the HPC Park Client API. Detailed instructions for working with the Account and HPC Park Client API here.
Two plans are available.
- The cost is calculated on the basis of the amount of resources consumed and the time of consumption at a minute-by-minute rate. Storage is charged separately according to the plan conditions, even when the container is stopped.
- When you buy a month subscription, you pay the full fixed price for the parameters selected in the configurator within this plan with its limits. Services are terminated only at the beginning of the new settlement period (unless auto-prolongation is enabled). No refund for unused resources.
We provide the possibility of obtaining not only a full-fledged NVIDIA Tesla A100 accelerator, but also its virtualized part, allowing to obtain a corporate-class accelerator for less demanding tasks, or the most complete utilization of computing tasks.
You can get either a pre-installed environment (Pytorch, TensorFlow) or a minimal software environment to choose from.
Scientific computing, rendering, special industrial applications. Computing related to photo and video processing, face recognition, machine translation and synthesis, predictive analytics.