X

General Availability of Virtual Machines with NVIDIA GPUs on Oracle Cloud Infrastructure

Karan Batta
Product Management

A few weeks ago, we announced preview availability for our virtual machine instances with NVIDIA Tesla Volta GPUs at the ISC conference in Germany. Customers have been using GPUs on Oracle Cloud Infrastructure for various use cases from engineering simulations and medical research to modern workloads such as machine learning training using frameworks such as TensorFlow.

This week we’re at NVIDIA’s GPU Technology Conference in Tokyo and I’m excited to announce General Availability for these virtual machines with NVIDIA Tesla Volta V100 GPUs in our London (UK) and Ashburn (US) regions this week. You’ll be able to login and launch these instances just the same way you normally launch instances on Oracle Cloud Infrastructure.

These virtual machines join the bare-metal compute instance we launched earlier in the year and provide you the entire server for very computationally intensive and accelerated workloads such as DNN training or run traditional high-performance computing (HPC) applications such as GROMACS or NAMD.

Finally, we’re also making our Pascal generation GPU instances available on Virtual Machines in our Ashburn (US) and Frankfurt (Germany) regions as a new cost-effective GPU option.

Data scientists, researchers, engineers and developers now have access to a portfolio of options ranging from a single P100 virtual machine to the cutting edge 8-way bare-metal instance with V100 Tesla GPUs. There’s something here for everyone!

 

Instance Shape

GPU Type

GPU(s)

Core(s)

Memory (GB)

Interconnect

Price (GPU/Hr)

BM.GPU3.8

Tesla V100 (SXM2)

8

52

768

P2P over NVLINK

$2.25

VM.GPU3.4

Tesla V100 (SXM2)

4

24

356

P2P over NVLINK

$2.25

VM.GPU3.2

Tesla V100 (SXM2)

2

12

178

P2P over NVLINK

$2.25

VM.GPU3.1

Tesla V100 (SXM2)

1

6

86

N/A

$2.25

BM.GPU2.2

Tesla P100

2

28

192

N/A

$1.275

VM.GPU2.1

Tesla P100

1

12

104

N/A

$1.275

You can additionally use NVIDIA GPU Cloud to launch HPC or AI application containers by simply deploying our pre-configured images along with NGC credentials. You can follow detailed step-by-step instructions here or visit our GPU product page for more information.

Finally, visit us this week in the exhibition hall at NVIDIA’s GTC Conference to talk to the engineering teams about Oracle Cloud Infrastructure or attend the breakout session on the 13th September at 12:10pm to learn more - https://www.nvidia.com/ja-jp/gtc/sessions/?sid=2018-1105.

We hope to see you there!

Karan

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.Captcha