AMD’s next-generation and award-winning AMD RDNA™ 2 architecture serves as the foundation for the company’s latest professional graphics technologies and products—introducing the new AMD Radeon PRO W6800, the fastest AMD RDNA workstation card ever.
Also introduced are the AMD Radeon PRO W6600 graphics card and the AMD Radeon PRO W6600M GPU designed to power professional mobile workstations.
RDNA 2 — Breakthroughs
“AMD RDNA 2 was designed from the ground up to deliver world-class performance for a wide range of applications and workloads,” says Scott Herkelman, corporate vice president and general manager, Graphics Business Unit at AMD. “Bringing this breakthrough architecture to our workstation graphics lineup allows users to tackle much larger datasets, dramatically reduce render times, and speed processing of highly complex models and simulations. The AMD Radeon PRO W6000 series gives professionals a powerful new tool in their arsenal to accelerate projects and bring creative visions to life.”
Key Takeaways: These new GPUs offer up to 214 percent faster performance in AEC applications like Autodesk Revit 2021.1 and faster performance in SOLIDWORKS Visualize 2021 in hardware raytracing than a comparable rival NVIDIA RTX 5000 card, all for less money.
The new graphics products were designed for demanding architectural design workloads, ultra-high resolution media projects, complex design and engineering simulations, and advanced image and video editing and compositing applications.
AMD has released its more powerful workstation-class graphics card ever in the new AMD Radeon PRO W6800.
Here are key features applicable to the series and after the jump, we have detailed specs, images, and more details about each solution.
- Award-winning AMD RDNA 2 Architecture — built on an advanced 7nm manufacturing process, this latest graphics architecture from AMD offers state-of-the-art GPU technologies
- Enhanced Compute Units with Realtime Hardware-Accelerated Raytracing — Ray Accelerators offer up to 46 percent faster rendering than Radeon PRO graphics cards based on the previous generation architectures in real app performance like SOLIDWORKS Visualize 2021. Support for Variable Rate Shading (VRS) delivers real-time photorealistic viewports and rendering
- AMD Infinity Cache — up to 128MB of last-level data cache integrated onto the GPU die helps reduce latency and power consumption
- Smart Access Memory — unlocks higher performance for key professional workloads by giving AMD Ryzen 5000 Series Desktop Processors or select AMD Ryzen 3000 series Desktop Processors access to the entire high-speed GDDR6 graphics memory
- AMD Radeon PRO Viewport Boost — designed for today’s professional workloads and compatible software to help viewport frames-per-second performance increase with project file sizes
- Certified for leading professional applications — A list of fully certified pro applications can be found here.
The AMD Radeon PRO W6800 graphics card is a double-slot solution while the Radeon PRO W6600 is a single-slot solution. We have images, details, specs, and pricing on the next page.
next page: Price, Performance Specs, Images, and Key Details on the New Radeon Pro cards
Pricing
The AMD Radeon PRO W6800 is priced at USD 2,249 recommended retail while the AMD Radeon PRO W6600 is priced at USD 649, striking a very affordable sweet spot for many AEC, MCAD, and visualization professionals. The mobile version, the W6600M is available in the HP Fury ZBook G8 mobile workstation in select countries starting in July of this year.
Performance Specs
The new Radeon PRO W6000 series GPUs dramatically improve over the previous generation pro solutions from AMD offering up to 17.83 teraflops performance (FP32) and up to 35.66 (FP16) for the W68000 card, and up to 10.4 teraflops for the W6600. The former card supports up to 32GB of GDDR6 memory while the latter card supports 8GB.
In terms of rival comparisons, the AMD Radeon PRO W6800 sits between NVIDIA’s RTX 6000 and RTX 5000 workstation solutions. As you can see in the image below the AMD PRO W6800 offers double the memory and far greater peak teraflop performance for both fewer costs and less power consumption.
And AMD to NVIDIA comparable for the new W6000 series Radeon PRO graphics. (click image to see larger)
The Radeon PRO W6800 can support up to 6x mini DisplayPort connections while the Radeon PRO W6600 can support up to 4x total full-size DisplayPort connections. As noted before the W6800 is a dual-slot card and has 3840 stream processors while the W6600 is a single-slot card and has up to 1792 stream processors.
A view of the Radeon PRO W6000 series chip and board. The 7nm chip is sizeable at a die size of 520 square mm.
The new RDNA 2 architecture chip is a 7nm process part at 520 square mm die size and features dedicated ray accelerators, enhanced compute units to decrease latency, AMD Infinity Cache—really a bandwidth amplifier—and support for PCIe 4.0 x 16 to remove common bottlenecks. The chip also has full support for the Vulkan graphics API and Microsoft DirectX12 Ultimate APIs. For those interested in transistors the W6800 has a total of 26.8 billion of them.
The mobile version of this chip, the Radeon PRO W6600M features 1792 stream processors on a chip physically about half the size of its larger sibling at 237 square mm. It features 11.06 billion transistors.
To learn more about these latest Radeon PRO W6000 series graphics processors you can visit AMD here online.
Architosh Analysis and Commentary
These latest cards from AMD show massive performance advantages when running applications that are compatible with its Radeon PRO Viewport Boost feature for 4K viewports. When the feature is off in Revit turning it on with these cards generates a 214 percent enhancement in FPS. Autodesk 3ds Max sees a 143 percent gain in FPS, while Epic’s Twinmotion sees a 139 percent gain.
AMD says that in terms of a generational improvement, AMD RDNA 2 Architecture is about 179 percent faster than AMD RDNA 1 Architecture. This is a significant improvement for AMD and closes the gap with its chief rival NVIDIA if not surpassing it in key performance metrics. One such metric often overlooked is noise. The Radeon PRO W6800 card is at 45.9 dBA compared to the NVIDIA RTX 5000 at the same internal temperature.
Editor’s Note:
(1) AMD noted an error they put out in their press release on the teraflop values for the Radeon PRO W6800, this has now been corrected. The chart we previously had in the article is now eliminated. — 7:05 AM, 9 June 2021.