Conclusion - Conclusion - WP557

Decoding the Versal AI Edge Series Gen 2 and Prime Series Gen 2 GPU for Next-Gen Visualization (WP557)

Document_ID
WP557
Release_Date
2025-01-03
Revision
1.0 English

For more information on the Versal AI Edge Series Gen 2 and Versal Prime Series Gen 2 adaptive SoCs, see the Versal Architecture and Product Data Sheet: Overview (DS950). The following table summarizes the features supported by these devices.

Table 1. Features, APIs, and Product Matrix
Feature Value Description
Anti-aliasing 4x, 8x, 16x MSAA Multisampling, anti-aliasing.
API support (graphics and compute)

OpenGL ES 3.2

Vulkan 1.4

Vulkan SC 1.0

OpenCL 3.0

Open GL SC 2.0

Full support for next-generation and legacy compute and graphics applications, including safety critical applications.
Adaptive scalable texture compression (ASTC) Low dynamic range (LDR) and High dynamic range (HDR). Supports both 2D and 3D images. ASTC helps improve image quality, reduce memory bandwidth, and thus, energy use.
Arm frame buffer compression (AFBC) Version 1.3, 4 x 4 pixel block size AFBC: Loss-less image compression format, helps reduce memory bandwidth.
Hardware partitions Up to two partitions GPU can be configured as a single partition with four shader cores or two partitions with two shader cores.
Multiple virtual machines Up to eight VMs GPU can support a maximum of eight VMs.
FMAX (evaluation) Up to 1050 MHz (-2MHP speed grade) Base device -1HP speed grade, @ 1000 MHz.
Vertex rate 2100 Mverts/s @1050 MHz (-2MHP speed grade) Number of vertices (points in 3D space), processed by the GPU per second.
Pixel fill rate 8.4 GPix/s @1050 MHz (-2MHP speed grade) Number of pixels a GPU can render per second.
Texture rate 16.8 GTex/s @1050 MHz (-2MHP speed grade) Speed that the GPU can process texture mapping operations.
FP32 Ops 268.8 GFLOPs @1050 MHz (-2MHP speed grade) Number of calculations involving floating-point numbers in one second using 32-bit precision.

Advantages of the Mali-G78AE GPU

Graphics Rendering Efficiency
The Mali-G78AE GPU offers significant improvements in graphics rendering efficiency, crucial for delivering high-quality visuals and user experiences.
Flexible Partitioning Architecture
The flexible partitioning architecture of the GPU allows for the isolation of critical workloads from non-critical ones, enhancing system stability and security. GPU virtualization enables the operation of multiple operating systems on the same GPU hardware, increasing system flexibility and resource utilization.
Heterogeneous Applications/Workloads
The Mali-G78AE GPU efficiently handles diverse workloads, including graphics rendering, general-purpose compute tasks, and machine learning inference, showcasing its versatility and suitability for complex, multiapplication environments.