At AWS re:Invent 2020, we’re celebrating 10 years of partnership and innovations — and paving the way for another decade of accelerated computing with Amazon Web Services’ newest virtual machine addition, the Amazon EC2 P4d instance, powered by the latest NVIDIA A100 Tensor Core GPU and now generally available.
AWS’s first GPU instance debuted 10 years ago, with the NVIDIA M2050. Since then, its portfolio of cloud NVIDIA GPU instances has grown to include the K80 (P2), K520 (G3), M60 (G3), V100 (P3/P3dn) and T4 (G4).
The P4d instance delivers AWS’s highest performance, most cost-effective GPU-based platform for machine learning training and high performance computing applications. The instances reduce the time to train machine learning models by up to 6x with TF32 compared to the default FP32 precision.
Demos and Sessions
As a Gold sponsor at this year’s re:Invent, we’ll be sharing some exciting demos, including NVIDIA CloudXR, our offering for streaming virtual reality, augmented reality and mixed reality content from any OpenVR XR application. Check out how Theia Interactive uses CloudXR to turn the design process into interactive Unreal Engine VR experiences with The Gettys Group in our virtual demo booth.
Need a car but not ready to head to a showroom? The NVIDIA RTX Ray-Tracing Car Configurator transports you to a virtual showroom where you can view cars in photorealistic configurations.
With the prevalence of remote work, many enterprises have moved to the cloud. So don’t miss the demo and session showing how AWS and NVIDIA are enabling engineers and designers to continue their work on GPU-accelerated virtual workstations. Also, be sure to check out the session titled “Virtual Workstations for Content Creation with G4 Instances” (CMP305) to hear how NVIDIA T4 Tensor Core GPUs with NVIDIA RTX technology enable creative studios to do their work anywhere from the cloud, presented by AWS.
To find the best machine learning model before serving it to the world, check out a tutorial that shows how to combine RAPIDS with Amazon SageMaker to accelerate hyperparameter optimization that shows a 12x speedup and 4.5x reduction in cost.
In our virtual booth, learn about conversational AI inference at data center scale using NVIDIA Riva, as well as how to easily deploy a BERT language model using NVIDIA Triton Inference Server on Amazon EKS.
Check all this out when our sponsor page goes live on Nov. 30.
In addition, we’ll be hosting 36 virtual theater sessions covering a range of topics:
- building next-generation healthcare AI on NVIDIA Clara
- accelerating natural language processing with GPUs for consumer internet companies
- driving fintech innovations
- advanced computer graphics for 3D design and animation
- powering autonomous driving with NVIDIA DRIVE Constellation and NVIDIA DRIVE Sim
If you’re interested in attending and engaging with our industry experts, check out the session schedule.
Free AI Training for Developers
During AWS re:Invent, the NVIDIA Deep Learning Institute is offering a free enrollment code for a self-paced course to the first 500 developers who sign up for our NVIDIA Developer Program. This program ensures developers have access to all the tools and training necessary to successfully build apps on all NVIDIA technology platforms.
DLI provides hands-on training in AI, accelerated computing and accelerated data science to help developers, data scientists and other professionals solve their most challenging problems. Select courses offer a certificate of competency to support career growth.
Learn how to set up an end-to-end project in just eight hours or how to apply a specific technology or development technique in two hours — anytime, anywhere — with just a computer and an internet connection.
Accelerating Startups
Through our partnership with AWS Activate, startups within the NVIDIA Inception acceleration program will be eligible for up to $100,000 in credits for the AWS cloud, which can be used to access the new NVIDIA A100 Tensor Core GPU.
In addition, NVIDIA Inception members can benefit from preferred pricing on on-premises GPUs, training through the DLI and go-to-market support.
Swing by NVIDIA’s virtual booth, starting Nov. 30, at AWS re:Invent 2020.