AMD has new professional workstation graphics cards (GPUs). Details were released this month detailing the new AMD Radeon PRO W7000 series cards built on the high-performance and energy-efficient AMD RDNA 3 architecture.
The new cards consist at the moment of the new Radeon PRO W7900 and Radeon PRO W7800.
World First Design
The new Radeon PRO W7000 series is the first major update since Architosh last reviewed the last round of Radeon PRO graphics cards. (see: Architosh, “Product Review: AMD Radeon Pro W6400 GPU for Workstations,” 16 Feb 22). These new graphics cards employ a world-first in the use of advanced AMD chiplet design, a semiconductor manufacturing technique that combines separate discreet “chiplets” into a more complex integrated circuit (IC). We’ll touch base on the advantages of chiplet design at the end of this article, but for now, it is important to recognize this new chiplet design is integral to the AMD RDNA 3 Architecture.
With the new architecture, new RDNA 3 compute units share resources between rendering, AI, and raytracing to boost such things as 50 percent faster raytracing than the previous generation. Part of this speed-up in the rendering technique is due to a second-generation raytracing engine.
There is also a new dedicated AI acceleration with new AI instructions offering over 2x performance in this area than the previous AMD RDNA 2 architecture. Also new is support for DisplayPort 2.1 and the AMD Radiance Display Engine, which boosts higher refresh rates, higher resolutions, and support for over 68 billion colors.
Two New Cards
The new AMD Radeon PRO W7000 series cards come in two models as noted below:
Radeon PRO W7900
- AMD RDNA 3 Architecture
- Stream Processors = 6144
- Compute Units = 96
- Peak Half Precision (FP16) Performance = 122.64 TFLOPs
- 57.8B transistors
- TSMC 5nm GCD 6nm MCD
- 48GB Memory with GDDR6
- AMD Infinity Cache = 96MB
- Peak Memory Bandwidth = 864 GB/s
- More details here.
Radeon PRO W7800
- AMD RDNA 3 Architecture
- Stream Processors = 4480
- Compute Units = 70
- Peak Half Precision (FP16) Performance = 90.5 TFLOPs
- 57.8B transistors
- TSMC 5nm GCD 6nm MCD
- 32GB Memory with GDDR6
- AMD Infinity Cache = 64MB
- Peak Memory Bandwidth = 576 GB/s
- More details here.
AMD continues to work with leading professional software application vendors to ensure AMD Radeon PRO graphics cards are built for demanding 24/7 environments and tested to meet exceptional standards, delivering the ideal balance of performance and stability. The list of certified applications can be found on the AMD website.
To learn more about the new AMD Radeon PRO W7000 series graphics card, visit here.