AMD based Amazon EC2 G4ad instances accelerate game engine workflows in the cloud

george_watkins · ‎12-13-2021

Game developers looking to adopt the latest cloud-based workflows now have a powerful new option when shifting their production pipelines online: AMD based Amazon Elastic Compute Cloud (Amazon EC2) G4ad instances.

Powered by AMD processors, G4ad instances provide a flexible, cost-effective environment that can run game development tools like Unreal Engine, enabling studios to set up virtual workstations for their artists and programmers. Now available in five different configurations, the instances allow teams of all sizes to benefit from the power of the cloud. Teams can collaborate with freelancers and outsourcing studios around the world, while bringing new machines online during crunch periods at the click of a button.

What are G4ad instances?

Amazon EC2 G4ad instances allow studios to run software securely in the cloud on virtual workstations, with only the encrypted pixels of the desktop display being streamed to an artist's PC or laptop.

The instances are powered by AMD GPUs and CPUs, bringing the power of the AMD RDNA™ and “Zen 2” architectures to the cloud. The AMD Radeon™ PRO V520 GPUs accelerate viewport playback, while the second-generation AMD EPYC™ processors provide up to 64 vCPUs per instance, powering highly multi-threaded tasks like lightmap baking and shader compilation.

Available in five different configurations to suit all types of creative workflows, G4ad instances provide up to 45% better price performance over comparable Amazon GPU-based instances for graphics-intensive applications.1

Size	Great for	vCPU	Memory	GPU	Storage
G4ad.xlarge	For GPU-centric applications which require less CPU resource	4	16 GB	1x 8GB GPU	150
G4ad.2xlarge	For GPU-centric applications which require less CPU resource	8	32 GB	1x 8GB GPU	200
G4ad.4xlarge	Accelerating professional workstation-class workloads that require large amounts of CPU and GPU resources	16	64 GB	1x 8GB GPU	600
G4ad.8xlarge	Accelerating professional workstation-class workloads that require large amounts of CPU and GPU resources	32	128 GB	2x 8GB GPU	1,200
G4ad.16xlarge	Large multi-session deployments that requires ample amounts of CPU and GPU resource	64	256 GB	4x 8GB GPU	2.400

How do game development tools benefit from running on G4ad instances?

The combination of AMD Radeon PRO GPUs and AMD EPYC CPUs in Amazon EC2 G4ad instances accelerates art creation, level design and programming workflows inside game engines like Unreal Engine. G4ad instances' powerful Radeon PRO V520 GPUs accelerate viewport playback, both the Unreal Editor, and when testing is completed, games are then compiled.

The following three workload examples demonstrate how each instance can be selected for a particular task:

Using an instance with a smaller vCPU count

Using an instance with a larger vCPU count

Using an instance with a multi-GPU option

The smaller vCPU count of g4ad.xlarge, g4ad.2xlarge asset development tasks like modeling, compositing, or rendering that do not require extensive high vCPU count.

The high vCPU count of g4ad.4xlarge, g4ad.8xlarge and g4ad.16xlarge instances also benefit highly multi-threaded tasks, such as baking lightmaps, compiling shaders, packaging projects, and compiling game engines: an efficiency boost for engineers.

The multi-GPU configuration of g4ad.8xlarge and g4ad.16xlarge instances can greatly help efficiencies where multi-GPU support is enabled. For example, features like nDisplay now supports multi-GPU to render your Unreal Engine scenes on multiple synchronized display devices to deliver a powerful immersive experience.

How can game developers benefit from moving to the cloud?

Cloud-based workflows can provide a wide range of benefits for game development studios, including cutting hardware costs and deployment times, and streamlining online collaboration:

Cloud-based workflows accommodate artists working remotely

With more people than ever now working from home, moving from physical to virtual workstations helps studios shift to new, less centralized team structures. For individual artists, being able to work in a familiar environment boosts creativity, while studios are freed to attract talent from around the world without having to worry about sourcing hardware or software licenses locally, or providing technical support in person.

Cloud-based workflows scale easily as projects expand

Virtual workstations make it easier for studios to bring in freelancers for crunch periods, or to crew up when a project expands. Rather than waiting for new hardware to be delivered, managers can simply move artists to larger instances, or deploy new instances at the click of a button. Studios can even store virtual machine profiles for regular contractors or outsourcing agencies, enabling them to begin work again immediately.

Cloud-based workflows can reduce start-up costs

Renting virtual workstations on demand rather than having to buy hardware up front can reduce start-up costs for small teams. For larger studios, prioritizing Operating Expenditure (OpEx) over Capital Expenditure (CapEx) makes costs more predictable, and reduces the financial risks of new hires being unable to work while waiting for workstations to be delivered, or of costly machines standing idle during quieter periods.

Cloud-based workflows simplify collaboration and data security

For commercially sensitive projects, working in the cloud simplifies data security. Since all of the project files are stored on the virtual workstation, and only pixels are streamed to the artist's machine, studios have more control over who can access confidential data. And with both in-house artists and external freelancers able to access the same virtual machine, teams can collaborate more easily on art assets or project files.

Offering workstation-class performance with game development tools like Unreal Engine, including the upcoming Unreal Engine 5, AMD-powered Amazon EC2 G4ad instances provide studios with a simple, cost-effective way to shift their production pipelines online. Ideally suited for both art and programming tasks ranging from lighting and level design to gameplay and tools development, G4ad instances allow game developers of all sizes to benefit from the power of the cloud.

More resources:

George Watkins is a Product Marketing Manager for AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

1 for more information, visit: https://aws.amazon.com/about-aws/whats-new/2021/07/introducing-new-amazon-ec2-g4ad-instance-sizes/