High Performance Computing is more accessible than ever!
HPCsolutions allows organizations to use the power of HPC datacenter for heavy computing tasks solution without having to invest in the very expensive infrastructure
Support by HPC Experts
HPCsolutions provides with free support from our HPC experts who can assist you in setting the optimal environment for your tasks
Isolated computing environment
HPCsolutions provides an isolated bare-metal HPC environment equipped with FDR Infiniband 56Gb/s interconnect
Software
The infrastructure will be optimally configured for all your application needs
Remote visualization
Remote visualization solution is designed for HPC desktop applications -- enabling real-time interactive GUI workflows and 3D visualization using NVidia GRID and PCoIP
Security
HPC cluster nodes operation without access to the Internet environment, mutual isolation of clients resources and use of industry-standard encryption protocols for remote connections provides high security level of our service

HPCsolutions - explore the power of HPC datacenter

HPCsolutions offers High-Performance Computing as a Service (HPCaaS) for tasks related to geophysics, computer aided engineering (CAE), hydro- and gas dynamic modeling and any other task that requires high performance computing.
HPCsolutions uses the most up-to-date hardware/software solutions for high performance computing in the most efficient bare-metal environment without any virtualization of compute resources. These solutions fully comply with the strict terms of the complex computing tasks.
The company has a strong HPC expertise. Our team of highly qualified IT engineers will help you solve efficiently any computing task.
HPCaaS is an optimal solution for tasks that require supercomputing power such as geophysical / seismic processing and imaging, hydro- and gas dynamics, video rendering, automated engineering calculations and much more.

The main benefits of HPCaaS:

You get exclusive use of customized HPC environment in the required amount - from tens to thousands of CPU cores, with no capital expenses.
Support of experienced IT engineers is part of the Service. There is no need to keep additional IT personnel employed by your Company
Pay only for the time HPC resourses are used to solve your computing tasks. There are no extra costs
The HPC equipment is always new, modern and in good working order. Its maintenance, repair and upgrade is our responsibility

HPC center description

InfiniBand FDR 56 Gb/s interconnect
built with Mellanox SX6518 switch
(324 ports FDR), fat tree
topology w/o over-subscription
Drives - 600Gb SSD or 500Gb HDD.
It's possible to install up to 2х2TB HDDs
on each node
Panasas ActiveStor 12 30TB,
connected to Infiniband
network
Cooling system - free cooling chillers
Uninterruptible power supply –
APC Symmetra 250 kWA
120 compute nodes 2 x Xeon E5-2670/2680v2
2400 CPU cores (with the possibility to
extend to 5000 CPU cores),
128–256 GB RAM per node,
(6,4 – 12,8 GB/CPU core)
Т-Platforms V-class.
V5000 blade chassis, V200S, V200F2 nodes
Servers equipped with AMD Firepro / NVIDIA Tesla
GPU accelerators, Intel Xeon Phi™ processors
are offered as an option. Additional NAS with required
storage size and performance are available
Internet connection speed – 1 Gbps,
dark fiber channel to MMTS-9
We can allocate Client's equipment –
(SAN, additional servers and network
equipment) in our datacenter. Possibility to setup
VPN tunnel to Client's network
HPC center is allocated in Moscow
in a specialized datacenter
equipped with necessary infrastructure

Collaboration format

1
HPC resources are available in the required amount, priced at the rate of CPU core per hour
2
Ordered HPC resources are provided exclusively to the Client in an isolated segment for the duration of use
3
Nodes are configured according to Client’s specifications. There is a free trial period for adjustments
4
Once the testing phase is over, the cluster is transferred to the Client for the computational tasks
5
Node configurations can be saved for rapid deployment in future if additional calculations are required

Cluster configuration for software use

HPC cluster has predefined configurations for the following software:
Paradigm
Ansys
SeisSpace® ProMAX®
FlowVision
pSeven
Salome
WRF
Intel
(compilers, MPI, debuggers)

Pricing

Service plan Т1   Т2   Т3   Т4   Т5   Т6
Prepayment
per month, US$.
2 000   4 000   10 000   20 000   30 000   40 000
CPU core, US$/hour 0.060   0.050   0.045   0.040   0.035   0.030
Fast SSD storage,
1 GB, US$/month
0.13   0.11   0.09   0.08   0.07   0.06
NVIDIA Tesla K40 GPU
US$/hour
0.88   0.76   0.66   0.58   0.50   0.42
Prepaid (*) amount of
cores x hour per month
30 000   67 000   189 000   435 000   750 000   1144 000
Cores (*) at constant
use per month
40   90   260   600   1040   1600
* - rough calculation based on the prepaid amount per month
Order

News

The advantages of HPC as a Service (HPCaaS)

Purchase of HPC cluster with 1000 CPU cores

TCO for 5 years – US$ 1 700 000:
  1. Cost of the servers and storage
    including bank's interest – US$ 830 000
  2. Equipment co-location expenses over 5 years –
    US$ 300 000 (US$ 5 000 per month)
  3. Salaries of IT personnel – US$ 420 000
  4. Equipment repair costs (AFR 5%) - US$ 170 000

Keeping in mind that in 5 years the equipment becomes obsolete and must be replaced. The equipment becomes outdated even earlier - after about 3 years.
 

HPCaaS

Five-year usage cost, with 50% utilization ratio - US$ 750 000
  1. Usage cost per month – US$ 25 000

By using HPCaaS from HPCsolution for 5 years, your savings will amount to about US$ 1 million. Even with the constant use of HPC cluster for 5 years, the total cost is lower by US$ 200 000

Other benefits
  • The risks of absence of commercial projects for HPC utilization are largely reducing
  • The equipment is always new. No need to worry about its modernization and maintenance
  • Our experienced IT engineers are always there to assist you.

Contacts

Address:
Moscow, Nauchniy prd, 19