“Vega” GPU Architecture
The Radeon™ Pro Vega graphics card features AMD’s cutting edge GPU architecture, purpose-built to handle big data sets and a diverse range of computational workloads.
Tremendous compute power to speed up professional applications where high performance is paramount, from rendering video effects to visualizing complex simulations.
Revolutionary Memory Engine
Create and design with larger, more detailed models and assets in real time.
The Radeon™ Pro Vega graphics card employs cutting-edge 2048-bit HBM2 memory delivering blazingly fast memory bandwidth of over 400GB/s to enable seamless performance at extreme resolutions in 4K and beyond.
Next-gen Compute Units (nCUs)
For many years GPUs have been optimized to process 32-bit floating point and integer data types, since these were best suited for standard 3D graphics rendering tasks. However as GPU workloads and rendering techniques have become more diverse, this one-size-fits-all approach is no longer always the best one. The processing units at the core of the “Vega” GPU architecture have been updated to address this new reality.
Next-Gen Compute Units (NCUs) provide super-charged pathways for doubling processing throughput when using 16-bit data types.1 In cases where a full 32 bits of precision is not necessary to obtain the desired result, they can pack twice as much data into each register and use it to execute two parallel operations. This is ideal for a wide range of computationally intensive applications including image/video processing, ray tracing, artificial intelligence, and game rendering.
OpenCL™ and Metal 2 Support
The Radeon Pro Vega graphics card enables professionals to tap into the parallel computing power of modern GPUs and multi-core CPUs to accelerate compute-intensive tasks in leading Media & Entertainment applications that support OpenCL and Metal 2.
Get accelerated performance for compute-intense workloads such as rendering and machine intelligence applications.
Enhanced Geometry Engine
The “Vega” architecture is designed to tackle the ever-increasing demand for more complex and detailed 3D graphics with an efficient new geometry engine. It combines the brute force of dedicated parallel processing pipelines with intelligent load balancing and new fast paths to deliver a major step forward in polygon throughput.
The most challenging workloads for a GPU can present it with millions of geometry primitives per frame, all of which must be evaluated to determine their contribution to the final image. New primitive shader technology allows Radeon Pro Vega graphics to perform geometry culling at an accelerated rate, eliminating unnecessary work for the rest of the GPU. An advanced workload distribution mechanism then assigns processing tasks to the available pipelines in a way that maximizes their utilization and avoids idle time. The result is Radeon Pro Vega graphics is capable of rendering extremely complex 3D models and scenes smoothly in real time.
Optimized Pixel Engine
Pixel counts are rising thanks to improvements in display and image-capture technologies. Radeon Pro Vega graphics feature a pixel engine to meet this challenge with a collection of optimizations intended to improve the throughput and efficiency of pixel processing.
The new rasterizer converts polygons into pixels more productively by breaking work into batches. These batches are sized to fit into on-chip cache. Smarter use of caching allows the Radeon Pro Vega graphics to reduce memory transactions and increase throughput compared to past approaches. Meanwhile, shade-once technology evaluates pixels early in the graphics pipeline and determines which ones are situated behind other objects on the screen. Non-visible pixels are safely discarded, eliminating shader processing work that won’t contribute to the final image.
At the same time, the Radeon Pro Vega can write images to memory more efficiently thanks to a new direct connection between the pixel engine and the on-chip cache. This improved pathway enables more effective use of cache memory in popular game engines and in other real-time graphics applications.
State-of-the-Art Power Management
New intelligent power management technologies adjust to your workload, providing thermal headroom to optimize peak performance and system acoustics. Advanced power agility features enable optimized performance during bursty workloads that are common in professional applications.
1. Discrete AMD Radeon™ GPUs and AMD FirePro™ GPUs based on the Graphics Core Next architecture consist of multiple discrete execution engines known as a Compute Unit (“CU”). Each CU contains 64 shaders (“Stream Processors”) working in unison. GRT-5