AMD Radeon Pro W6900X

AMD Radeon Pro W6900X

AMD Radeon Pro W6900X: Power for Professionals and Enthusiasts

Review of the graphics card for demanding tasks (April 2025)

Introduction

The AMD Radeon Pro W6900X is a professional graphics card designed for use in studios, research centers, and engineering projects. While its primary audience is professionals, many enthusiasts also take notice due to its unique characteristics. In this article, we will explore what sets the W6900X apart in 2025, how it performs in gaming and complex tasks, and who should consider purchasing it.


Architecture and Key Features

RDNA 2 Architecture: The W6900X is built on an enhanced version of the RDNA 2 architecture, which debuted in 2020. Despite the introduction of RDNA 3 and RDNA 4, this model remains relevant due to driver optimizations and support for new technologies.

Process Technology and Compute Units:

- 7nm process from TSMC;

- 5120 stream processors;

- 32 compute units (CU).

Unique Features:

- FidelityFX Super Resolution (FSR) 3.0: An upscaling technology that increases FPS in games with minimal quality loss. As of 2025, it is supported by most AAA titles.

- Ray Accelerators: Hardware blocks for ray tracing. Although performance is lower than NVIDIA's RTX 40xx series, it is sufficient for rendering in Blender or Maya.

- Infinity Cache: 128 MB of cache reduces memory latency.

Professional Features:

- Support for ECC memory to protect against errors in calculations;

- Optimization for OpenCL and Vulkan APIs.


Memory: Speed and Capacity

Type and Capacity:

- 32 GB HBM2e — high-speed memory with a stacked structure;

- Bandwidth: 1.6 TB/s.

Why HBM?

HBM2e provides record-breaking bandwidth, which is critical for tasks such as:

- 8K video rendering;

- Working with neural network models;

- Simulations in MATLAB or ANSYS.

Impact on Gaming:

Despite 32 GB of memory, the advantage of HBM2e in gaming is less pronounced due to optimizations for GDDR6/GDDR6X. However, at 4K resolution and above, the card shows stability due to its large buffer size.


Gaming Performance: Not the Main Focus, but Impressive

The W6900X is not designed for gaming, but its power is sufficient for comfortable gaming experiences:

Average FPS (2025, Pro 23.Q4 Drivers):

- Cyberpunk 2077 (4K, Ultra, FSR 3.0 Quality): 48–55 FPS;

- Starfield (1440p, Ultra, without FSR): 65–70 FPS;

- Horizon Forbidden West (4K, Ultra): 60 FPS (with FSR 3.0).

Ray Tracing:

Enabling ray tracing drops FPS by 30–40%. For instance, in Control (4K, RT High), the performance falls to 28–35 FPS. For ray tracing games, it is better to consider NVIDIA GeForce RTX 4080/4090.

Summary: The card is suitable for casual gamers at 4K or enthusiasts who value stability. However, for the same price of $2500–3000, more gaming-oriented solutions can be found.


Professional Tasks: Where the W6900X Shines

3D Rendering:

- Blender (Cycles): Performs on par with NVIDIA RTX A6000 due to optimization for HIP API;

- Autodesk Maya: 8K textures are processed without lag.

Video Editing:

- DaVinci Resolve: Renders an 8K project in 12 minutes (compared to 15 minutes for RTX 3090);

- Adobe Premiere Pro: Effect acceleration via OpenCL.

Scientific Calculations:

- Support for OpenCL and ROCm 5.5;

- SPECviewperf 2025 test: 15% faster than RTX A5000 in CFD tasks.

Why Not CUDA?

AMD is betting on open standards (OpenCL, Vulkan), but NVIDIA's CUDA still dominates in niche scientific packages.


Power Consumption and Heat Dissipation

TDP: 300 watts — which is high but acceptable for a workstation.

Recommendations:

- Power Supply: At least 750 watts with an 80+ Gold certification;

- Cooling: The card is equipped with a turbine cooler. For building, choose a case with good ventilation (e.g., Fractal Design Meshify 2);

- Temperatures: Under load — up to 78°C, throttling starts after 85°C.


Comparison with Competitors

NVIDIA RTX A6000 (48 GB GDDR6):

- Pros: Better RT performance, more memory;

- Cons: Price ($5500), higher power consumption (320 watts).

AMD Radeon Pro W7800 (32 GB GDDR6):

- Pros: New RDNA 4 architecture, lower price ($2000);

- Cons: Lower memory bandwidth.

Conclusion: The W6900X remains a solid choice for those who prioritize HBM2e speed and the reliability of Pro drivers.


Practical Tips

1. Power Supply: Corsair RM850x (850 watts) or similar.

2. Compatibility:

- Requires PCIe 4.0 x16;

- macOS support only in Mac Pro 2023+.

3. Drivers:

- Use AMD Pro Edition — they are more stable for professional tasks;

- For hybrid use (gaming + work), you can install Adrenalin, but conflicts may arise.


Pros and Cons

Pros:

- 32 GB HBM2e — ideal for rendering;

- Support for ECC memory;

- Optimization for professional software.

Cons:

- Price ($2700–3000);

- Weak RT performance in games;

- Noisy cooler under load.


Final Verdict: Who Should Consider the W6900X?

This graphics card is designed for:

- Video Editors: Working with 8K and heavy effects;

- 3D Artists: Rendering complex scenes without overload;

- Engineers: Calculations in CAD and simulations;

- Enthusiasts: Who want a powerful card for mixed tasks.

If you are a gamer or have a limited budget — consider the Radeon RX 7900 XT or NVIDIA RTX 4080. But for professionals who value reliability and speed, the W6900X is a sound investment in 2025.

Basic

Label Name
AMD
Platform
Desktop
Launch Date
August 2021
Model Name
Radeon Pro W6900X
Generation
Radeon Pro Mac
Base Clock
1825MHz
Boost Clock
2150MHz
Bus Interface
PCIe 4.0 x16
Transistors
26,800 million
RT Cores
80
Compute Units
80
TMUs
?
Texture Mapping Units (TMUs) serve as components of the GPU, which are capable of rotating, scaling, and distorting binary images, and then placing them as textures onto any plane of a given 3D model. This process is called texture mapping.
320
Foundry
TSMC
Process Size
7 nm
Architecture
RDNA 2.0

Memory Specifications

Memory Size
32GB
Memory Type
GDDR6
Memory Bus
?
The memory bus width refers to the number of bits of data that the video memory can transfer within a single clock cycle. The larger the bus width, the greater the amount of data that can be transmitted instantaneously, making it one of the crucial parameters of video memory. The memory bandwidth is calculated as: Memory Bandwidth = Memory Frequency x Memory Bus Width / 8. Therefore, when the memory frequencies are similar, the memory bus width will determine the size of the memory bandwidth.
256bit
Memory Clock
2000MHz
Bandwidth
?
Memory bandwidth refers to the data transfer rate between the graphics chip and the video memory. It is measured in bytes per second, and the formula to calculate it is: memory bandwidth = working frequency × memory bus width / 8 bits.
512.0 GB/s

Theoretical Performance

Pixel Rate
?
Pixel fill rate refers to the number of pixels a graphics processing unit (GPU) can render per second, measured in MPixels/s (million pixels per second) or GPixels/s (billion pixels per second). It is the most commonly used metric to evaluate the pixel processing performance of a graphics card.
275.2 GPixel/s
Texture Rate
?
Texture fill rate refers to the number of texture map elements (texels) that a GPU can map to pixels in a single second.
688.0 GTexel/s
FP16 (half)
?
An important metric for measuring GPU performance is floating-point computing capability. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy.
44.03 TFLOPS
FP64 (double)
?
An important metric for measuring GPU performance is floating-point computing capability. Double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy, while single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
1376 GFLOPS
FP32 (float)
?
An important metric for measuring GPU performance is floating-point computing capability. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
21.58 TFLOPS

Miscellaneous

Shading Units
?
The most fundamental processing unit is the Streaming Processor (SP), where specific instructions and tasks are executed. GPUs perform parallel computing, which means multiple SPs work simultaneously to process tasks.
5120
L1 Cache
128 KB per Array
L2 Cache
4MB
TDP
300W
Vulkan Version
?
Vulkan is a cross-platform graphics and compute API by Khronos Group, offering high performance and low CPU overhead. It lets developers control the GPU directly, reduces rendering overhead, and supports multi-threading and multi-core processors.
1.3
OpenCL Version
2.1
OpenGL
4.6
DirectX
12 Ultimate (12_2)
Power Connectors
None
Shader Model
6.5
ROPs
?
The Raster Operations Pipeline (ROPs) is primarily responsible for handling lighting and reflection calculations in games, as well as managing effects like anti-aliasing (AA), high resolution, smoke, and fire. The more demanding the anti-aliasing and lighting effects in a game, the higher the performance requirements for the ROPs; otherwise, it may result in a sharp drop in frame rate.
128
Suggested PSU
700W

Benchmarks

FP32 (float)
Score
21.58 TFLOPS
Vulkan
Score
105424
OpenCL
Score
141178

Compared to Other GPU

FP32 (float) / TFLOPS
23.177 +7.4%
20.325 -5.8%
19.1 -11.5%
Vulkan
382809 +263.1%
140875 +33.6%
61331 -41.8%
34688 -67.1%
OpenCL
385013 +172.7%
167342 +18.5%
75816 -46.3%
57474 -59.3%