4 Tesla™ P100 GPUs
8 Tesla™ P100 GPUs
16 Tesla™ P100 GPUs
4 Tesla™ P100 GPUs
GPU ACCELERATORS
Four NVIDIA Tesla P100s with NVLink
PROCESSORS
Dual Intel Xeon E5-2690 v4 (2.6Ghz)
SYSTEM MEMORY
512GB
PRIMARY STORAGE
1.6TB NVMe SSD
WEEKLY PRICE
$1,300/wk
MONTHLY PRICE
$4,200/mo
8 Tesla™ P100 GPUs
GPU ACCELERATORS
Eight NVIDIA Tesla P100s with NVLink
PROCESSORS
Dual Intel Xeon E5-2667 v4 (3.2Ghz)
SYSTEM MEMORY
512GB
PRIMARY STORAGE
1.6TB NVMe SSD
WEEKLY PRICE
$2,600/wk
MONTHLY PRICE
$8,400/mo
16 Tesla™ P100 GPUs
GPU ACCELERATORS
Sixteen NVIDIA Tesla P100s
PROCESSORS
Dual Intel Xeon E5-2667 v4 (3.2Ghz)
SYSTEM MEMORY
512GB
PRIMARY STORAGE
1.6TB NVMe SSD
WEEKLY PRICE
$4,800/wk
MONTHLY PRICE
$16,000/mo
GPU ACCELERATORS
Four NVIDIA Tesla P100s with NVLink
Eight NVIDIA Tesla P100s with NVLink
Sixteen NVIDIA Tesla P100s
PROCESSORS
Dual Intel Xeon E5-2690 v4 (2.6Ghz)
Dual Intel Xeon E5-2667 v4 (3.2Ghz)
Dual Intel Xeon E5-2667 v4 (3.2Ghz)
SYSTEM MEMORY
512GB
512GB
512GB
PRIMARY STORAGE
1.6TB NVMe SSD
1.6TB NVMe SSD
1.6TB NVMe SSD
WEEKLY PRICE
$1,300/wk
$2,600/wk
$4,800/wk
MONTHLY PRICE
$4,200/mo
$8,400/mo
$16,000/mo

With these models you can:

  • Rent a dedicated, non-virtualized platform and load your application software

  • Rent space on a dedicated, non-virtualized platform with our full software library preloaded

  • Lease for an extended time period –week, month, or year– with discounts for more than 6 months

  • Rent to own. Pay monthly instead of making a large up-front purchase.

  • Purchase a new system and apply 25% of your first two months’ rent to the purchase price

  • Purchase the system you use and apply 50% of the first two months’ rent to the purchase price

Software Library

OPERATING SYSTEM: CentOS 7.3           NVIDIA DRIVERS: CUDA 8, CUDA 7.5

Preinstalled Machine Learning Frameworks

Preinstalled Machine Learning Libraries

MLPython • ML Dependencies (400MB Python) • cuDNN (5.0 & 5.1) • DIGITS

Caffe on Spark • CUDA & NVIDIA driver • CUB (CUDA building blocks) • NCCL

Each Customer Receives

  • A dedicated, sanitized system for the time they purchase

  • The ability to scale up or down as they require

  • Use of our vast available software library without additional costs or hidden fees

Security

Each HPCaaS client gets a secure, dedicated system with no virtual machines operating on the same system. A dedicated internet connection along with your own IP address aids in the isolation of your system.

HPC Cloud Services With the Leadership Features You Need

Dedicated Hardware

With root access to bare metal multi-GPU and clustered systems, you control the compute power to your entire platform. No virtualization and no sharing means all the GPUs drive your applications with world-class speed.

Security

SkyScale’s supercomputing platforms are located in world-class industry leading data centers with unmatched (and certified) physical and cyber security, plus redundancies resulting in not a single minute of downtime in seven years! Your access is always available and your data is always protected and secure. Download our Security and Reliability Document. 

Access to Experts

SkyScale automates equipment and processes, not communication. Working with SkyScale means always having access to experts who understand your needs and can resolve your needs quickly, leaving you to focus on the truly valuable things that drive your productivity and success.

Time Maximization

There’s nothing more important to SkyScale than your time. Our goal is to facilitate and enhance your productivity, leaving you more time to do what you want to do. Running your applications on SkyScale’s world-class platforms minimizes your time to results, value, and satisfaction.

Save on Capital Costs

Lease only what you need when you want it and scale up or down as needed, saving on security and capital costs for equipment and on-prem data center space and maintenance.

Highest Performance

With exclusive access to as many as 16 NVIDIA V100s or P100s in a single node, your applications are running on the world’s highest performing multi-GPU platform. Couple that with state-of-the-art high speed low latency NVMe storage and your project time to completion just got a lot shorter.

INDUSTRY SOLUTIONS

Deep Learning and A.I.

Financial

Oil and Gas

Life Science

Research

PARTNERS

Mellanox
Juniper
Mapd
Intel
One Stop Systems
NVidia
Snap Route
Edge Core
Scale Matrix
Bit Fusion
Bright
San Diego Super Computing System