News
AWS Adds Lower-Cost GPU Instance to AppStream 2.0
Users of AppStream 2.0, the application streaming solution from Amazon Web Services (AWS), now have a new instance option for graphics-heavy applications.
AWS this week announced the launch of a GPU-optimized instance for AppStream 2.0 called Graphics Design, just over a month after unveiling the offering's first two GPU instances, Graphics Desktop and Graphics Pro.
Graphics Design offers more compute and cost flexibility that the other two GPU instances, coming in four different sizes compared to three for Graphics Pro and just one for Graphics Desktop.
Its configurations range from stream.graphics-design.large, which contains two virtual CPUs, 7.5GiB of system memory and 1GiB of GPU memory, to stream.graphics-design.4xlarge, with 16 virtual CPUs, 61GiB of system memory and 8GiB of GPU memory.
Pricing for the Graphics Design instance ranges from $0.25 per hour to $2 per hour. The smallest size in the Graphics Design family costs half as much as a Graphics Desktop instance and about an eighth of the smallest Graphics Pro instance.
[Click on image for larger view.] Comparison of the three GPU instances for AppStream 2.0, including pricing for AWS' Northern Virginia region. (Source: AWS)
The three AppStream 2.0 GPU instances are meant to enable users to stream more graphics-heavy applications, including those geared toward media, high-performance computing, engineering and design.
As the newest, Graphics Design is "ideal for delivering applications that rely on hardware acceleration of DirectX, OpenGL, or OpenCL, such as Adobe Premiere Pro, Autodesk Revit, and Siemens NX," AWS said in its announcement.
Graphics Design instances run on AMD FirePro S7150x2 Server GPUs (the other two instances are based on NVIDIA GPUs) and use AMD's Multiuser GPU technology.