By Edward Churchward, Founder & CTO, Arch Platform Technologies
Cloud workstations are revolutionizing how professionals manage complex workloads, offering the ability to access high-performance computing resources from virtually anywhere. By moving computationally intensive tasks like 3D graphics, simulation, machine learning, and video processing to the cloud as virtual desktop infrastructure (VDI), businesses improve teamwork and gain the flexibility to scale their resources on demand.
What is an Amazon EC2 ‘G’ workstation?
In AWS, G-type instances are a class of GPU-optimized EC2 instances designed for tasks that require powerful graphics processing units (GPUs). The "G" in G-type instances stands for GPU, as these instances are commonly used for applications like CGI, design and engineering, video editing, AI, simulation, and real-time spatial computing. They provide access to high-performance NVIDIA and AMD GPUs, which can handle parallel processing for both graphics and computation.
Translating the AWS naming convention!
When talking about workstation specifications, people normally refer to an i9 CPU with 128GB ram and a 4090 GPU. In the land of AWS we have to extrapolate these specifications from their naming conventions, for example a g5.8xlarge means it has 32 cpu cores, 128GB ram and an A10G GPU.
Here is a brief explanation of how these conventions work:
The G: (g6.4xl or gr6.4xl)
The G, as mentioned above, is the primary portion of the name and signifies this is a GPU based machine with the ability to do accelerated VDI (virtual desktop infrastructure) workflows. They sometimes modify the G to better describe the instance, like in the case of the Gr6 where they add an ‘r’ to show it has double the RAM of a standard G6.
Immediately after the G: (g6.4xl or g6e.4xl)
The characters after the G, and before the period, represent the generation and type of GPU that the workstation will have. A G6 has an NVIDIA L4 but a G6e has an NVIDIA L40S. In general, the bigger the number, the newer and faster the workstation is, but there is nuance to this whereby a G5 can be faster than a G6 for specific AI workloads due to vRAM bandwidth.
After the period: (g6.4xl)
The final part of AWS instance type naming that is separated by a period defines the specific configuration of the given generation and type of workstation computer. For example, a g6.2xlarge has twice the vCPUs, twice the RAM and more NVMe than that of a g6.xlarge. A general rule of thumb is that as you go up by the power of 2, (i.e. 2,4,8,16), you double the RAM and vCPUs. The Figure 1 table below illustrates configurations within a G5 generation of workstations related to this AWS’ naming convention.
Figure 1 - Amazon Web Services (AWS) G5 instances and related naming convention
From the EC2 G4dn and G4ad, which are purpose built for standard 2D/3D and video workloads, to the G5 and G6e instances designed for demanding 3D graphics, simulation and AI workflows, AWS provides robust solutions for media & entertainment, visual effects, AI, industrial design and product engineering, AEC, energy, healthcare, and scientific computing.
NVIDIA GRID virtual workstation drivers are included as part of the EC2 instance cost. In addition, a free high performance workstation VDI client called Amazon DCV (Windows, Linux or Mac) is available. AWS instances run either Windows or Linux.
Advantages of EC2 Workstations on AWS
As more industries transition to remote work and cloud-based infrastructures, using Amazon EC2 workstations on AWS has emerged as a powerful solution for professionals across various fields like software development, 3D graphics, video production, and machine learning. AWS and Arch Platform Technologies provide secure, high-performance workstations with the ability to scale according to project demands, all while minimizing upfront costs. Here’s an in-depth look at the advantages of AWS EC2 workstations and an overview of the newest instance types tailored to specific workflows.
1. Scalability and Flexibility
One of the greatest advantages of cloud workstations is their ability to scale according to a user's needs. Whether you’re working on a small project or need a powerhouse for rendering complex 3D graphics, AWS allows you to easily select the resources required. This flexibility ensures that you only pay for the computing power you use, minimizing waste and improving cost efficiency. With AWS and the Arch Platform, orchestration of resources is fast with tremendous insight, eliminating requirements for highly specialized cloud DevOps resources.
2. Global Accessibility
AWS EC2 workstations allow users to work from anywhere, provided they have an internet connection. With global data centers, Arch + AWS ensure multi-region orchestration of low-latency virtual workstations, providing seamless performance. This is particularly useful for geographically distributed teams working on the same project, enabling collaboration without the need for powerful local hardware.
3. Security, Compliance and Resilience
AWS and Arch Platform Technologies have a robust set of security measures in place, including encryption, identity and access management (IAM), and compliance certifications. This makes Amazon EC2 workstations a secure choice for industries that deal with sensitive data, such as healthcare, product development & manufacturing, and media. AWS’s cloud infrastructure is built to meet the strictest regulatory standards, allowing users to maintain compliance across various regions and industries. And, if a machine doesn’t work for some reason, you can always get a new box in seconds - so infrastructure is always working.
4. Cost Efficiency
With EC2 workstations on AWS, there is no need to invest in expensive local hardware and the frequent over-provisioning of that hardware. Users can start and stop instances as needed, ensuring that you’re only charged when the instance is in use. For businesses that require high-performance workstations on-demand, such as rendering studios or development teams working in cycles, this can lead to significant cost savings.
5. High-Performance Computing
AWS offers cutting-edge EC2 workstation instance types that are optimized for performance-heavy tasks, such as CGI, product design, simulation, AI, and video editing. These instances are equipped with high-end GPUs and CPUs, ensuring that even the most demanding tasks can be performed with ease.
Amazon EC2 Workstation Instance Types and Their Use-Cases
Many modeling, CAD and other 2D/3D software apps have been optimized for single-threaded CPUs. Some of these are in transition to take better advantage of GPU and high core-count CPU architectures, but this transition may take time and/or not make technical sense for every workstation-class compute application. Please refer to the specific software in your overall workflow pipeline to orient to which workstation architecture meets the right balance between performance and cost. We have pulled together the table in Figure 2 below to address the AWS cloud workstation instance types along with important specifications that influence their range of performance, which speaks more specifically to their respective use-case.
Bringing it all together
With cloud workstation infrastructure, the technology adapts to project and user workflow requirements. It’s more or less future-proof due to AWS’s constant investment and expansion across a large base of business users. For example, if a workflow changes from 2K video files to 4K, 6K or 8K, it’s quick and easy to adjust with cloud infrastructure based on AWS and Arch’s platform. With on-prem, you spec to user needs in your specific business in a rigid way so that not long after, everyone must adjust to the technology rather than the other way around. This is why on-prem frequently over-provisions due to related uncertainties. In this vein, cloud is great for augmenting your existing on-prem infrastructure as well as for providing migration opportunities for your digital transformation initiatives.
An important benefit for Arch + AWS cloud workstation end-users is that their user profiles follow them from machine to machine, so their preferences roam to whichever machine they sign onto.
Related to instance types and cost-performance, there are important nuances to consider. For example, when you compare running a g5.8xlarge for Unreal Engine, Maya or V-Ray workloads to a gr6.4xl that still has 128 GB of RAM and a faster CPU with slightly fewer cores and a powerful GPU, you will save roughly 35% per month with the Gr6. Our Arch Platform team is ready to help with the overlaps in specific AWS instance types and insights to any new ones as they are released.
At the end of the day, the most important and valuable thing is time. Time of end-users, time of the technology team, and project cycle time. This is the real power of the cloud augmenting or replacing on-prem infrastructure. We gain valuable time which equates to financial and competitive advantages.
#CloudComputing, #CloudWorkstations, #AWS, #AWSCloud, #CloudRendering, #3DRendering, #CloudInfrastructure, #3DGraphics, , #GPUComputing, #MachineLearning, #HighPerformanceComputing, #AWSWorkstations, #RemoteWorkstation, #PostProduction, #VFX, #AIandML, #DataScience, #CloudInnovation, #DigitalTransformation, #GraphicsDesign, #VisualEffects, #Simulation, #IndustrialDesign, #AEC, #ScientificComputing
Comments