aws logo

AWS

Largest all around cloud platform by market share with GPU offerings

Fast Facts
Datacenter Tier: 4*Founded: 2006Founders: AmazonFunding: Publicly Trade CompanyGpu Variety: MediumHeadquarters: Seattle, USAMerits: SOC I, SOC II, SOC III, HIPAAPersistent Storage: Supported

$$$$
Price

Low

Availability
Available Regions

Overview

As the original public cloud, AWS is the leading all around cloud provider with the widest feature set and the largest footprint. In regards to GPU specific offerings, AWS provides GPU virtual machines under its EC2 offering. Compared to smaller GPU clouds, AWS offers a smaller set of Nvidia GPUs and often only VMs with a full set of 8 GPUs.

Highlights

Dots
Wide Geographic FootprintRegions across multiple continents with widespread availability
Dots
Reservation PricingReduce prices for long term commitments
Dots
SLA GauranteeProvider offers an SLA on uptime and performance

Instances

Features

01.

Shadeform Integrated

Integrated with Shadeform for console and API

02.

Fast Spin Up

Machines are deployed quickly when they are launched

03.

Persistent Storage

Disk storage that can be used across multiple instances

04.

Interconnected Clusters

Offers multiple GPU nodes hardware interconnected with a high speed connection

Pricing

AWS offers multiple GPU accelerated instance types such as H100s, A100s, A10g, and V100s. Unfortunately, AWS's cloud offerings are some of the highest price GPUs on the market owing to their extensive integrations with their existing platform. There are opportunities for discounts on their machines through long term reservations of multiple nodes.

Availability

Availabilty of AWS GPUs varies depending on the GPU type across its many regions and availability zones. For higher end GPUs like A100s and H100s, on-demand availability is relatively low as most machines are dedicated to reservations. However, for lower end GPUs like A10G and V100s, there is generally availabililty in multiple regions.

© Shadeform, Inc