AMD Radeon RX 6900 XT

AMD Radeon RX 6900 XT

AMD Radeon RX 6900 XT in 2025: Complete Analysis of the Flagship RDNA 2

An updated look at the legendary graphics card: what's relevant and what has become outdated?


1. Architecture and Key Features: RDNA 2 in Detail

The RX 6900 XT graphics card is built on the RDNA 2 architecture, which remains relevant in 2025 due to optimizations from AMD. The chip is manufactured using the 7nm TSMC process and contains 26.8 billion transistors. Key innovations include:

- Ray Accelerators — hardware blocks for ray tracing. There are 80 of them (one for each Compute Unit). Compared to NVIDIA's RTX 30 series, ray tracing performance is lower, but drivers from 2024-2025 have significantly improved the situation.

- Infinity Cache — 128 MB of cache memory that reduces latency and enhances efficiency when working with GDDR6.

- Smart Access Memory (SAM) — a technology that gives the CPU access to the entire video memory. An AMD Ryzen 5000/7000 processor and a motherboard with PCIe 4.0 are required to activate it.

FidelityFX Super Resolution (FSR) 3.0 — AMD's main advantage in the battle for FPS. Unlike NVIDIA’s DLSS 3.0, FSR works on any GPU, including competitors, and employs upscaling + frame generation. In 2025, FSR 3.0 support is available in 90% of new AAA games.


2. Memory: GDDR6 and Its Potential

The RX 6900 XT is equipped with 16 GB GDDR6 with a 256-bit bus and a bandwidth of 512 GB/s (memory frequency — 16 Gbps). This is sufficient for:

- 4K gaming with high texture settings.

- Working with 8K video and demanding 3D scenes.

However, in 2025, competitors like the NVIDIA RTX 4080 offer GDDR6X or even HBM3, which provides an advantage in memory-intensive tasks (for example, neural network rendering).


3. Gaming Performance: Is 4K Still Comfortable?

In 2025, the RX 6900 XT remains a solid choice for 1440p and 4K, but ray tracing requires compromises:

- Cyberpunk 2077 (4K, Ultra, RT Ultra): 34–42 FPS without FSR, 55–60 FPS with FSR 3.0.

- Starfield (4K, Ultra): 68–75 FPS.

- Alan Wake 2 (1440p, Medium RT): 48–55 FPS with FSR 3.0.

For 1080p, the card is overkill—it hits CPU bottlenecks here. In 1440p, the average FPS in new games is 90–120 (without RT); in 4K, it’s 50–70. Ray tracing reduces performance by 30-50%, but FSR 3.0 partially mitigates the losses.


4. Professional Tasks: Not Just Gaming

- Video Editing: Support for AV1 (decoding) and HEVC (encoding) codecs in DaVinci Resolve and Premiere Pro. Rendering a 4K project takes 15-20% longer than on the RTX 3090.

- 3D Modeling: In Blender and Maya, the RX 6900 XT shows results close to the RTX 3080 Ti thanks to optimizations in the Cycles engine for RDNA 2.

- Scientific Calculations: Support for OpenCL and ROCm (an alternative to CUDA). For machine learning, the card is less preferred due to its limited ecosystem.


5. Power Consumption and Cooling: Is an Upgrade Needed?

- TDP: 300W (peak consumption — up to 350W).

- Recommendations:

- A power supply of at least 850W (for systems with Ryzen 9 7900X or Intel i7-14700K).

- A case with good ventilation (minimum of 3 fans).

- An optimal solution is hybrid cooling (for example, Sapphire Nitro+ Hybrid) or custom liquid cooling for overclocking.

The card heats up to 75-80°C under load, but modern models (from 2024-2025) have received updated cooling systems with solder instead of thermal paste.


6. Comparison with Competitors: Who Does It Compete Against?

- NVIDIA RTX 4080 (16 GB GDDR6X, $1000): 25-30% faster in ray tracing, supports DLSS 3.5, but is more expensive.

- AMD RX 7900 XT (20 GB GDDR6, $750): A younger sibling with a newer RDNA 3 architecture. Performance on par with 6900 XT, but better energy efficiency.

- Intel Arc A770 (16 GB GDDR6, $350): Cheaper but weaker in 4K and lacks an equivalent to FSR 3.0.

Conclusion: The RX 6900 XT in 2025 is a choice for those seeking a balance between price ($650-750) and performance without a fanatic emphasis on ray tracing.


7. Practical Tips: How to Avoid Problems

- Power Supply: 850W with an 80+ Gold certification. The best models include Corsair RM850x, Be Quiet! Straight Power 11.

- Compatibility:

- Motherboards with PCIe 4.0 and support for Resizable BAR (for SAM).

- AMD Ryzen 5000/7000 processors or Intel 12th-14th generation processors.

- Drivers: Adrenalin Edition 2025 is stable, but check the lists of known bugs before installing updates.


8. Pros and Cons

Pros:

- 16 GB of memory for future games and professional tasks.

- Excellent 4K performance without RT.

- Support for FSR 3.0 and open standard.

Cons:

- Ray tracing is weaker than NVIDIA's.

- No hardware support for frame generation like in DLSS 3.5.

- High power consumption.


9. Overall Conclusion: Who is the RX 6900 XT Suitable For?

This graphics card is an ideal choice for:

- Gamers playing in 4K without ultra RT settings.

- Enthusiasts upgrading their PC on a budget (price lower than RTX 4080 by 30%).

- Professionals using OpenCL and looking to save money.

However, if you are a fan of cyberpunk games with ultra RT settings or work within the CUDA ecosystem, consider NVIDIA.


Prices are current as of April 2025. Retail prices for new devices in the U.S. are listed.

Basic

Label Name
AMD
Platform
Desktop
Launch Date
October 2020
Model Name
Radeon RX 6900 XT
Generation
Navi II
Base Clock
1825MHz
Boost Clock
2250MHz
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
16GB
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.
288.0 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.
720.0 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.
46.08 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.
1440 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.
22.579 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
2x 8-pin
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

Shadow of the Tomb Raider 2160p
Score
77 fps
Shadow of the Tomb Raider 1440p
Score
145 fps
Shadow of the Tomb Raider 1080p
Score
190 fps
Cyberpunk 2077 2160p
Score
67 fps
Cyberpunk 2077 1440p
Score
75 fps
Cyberpunk 2077 1080p
Score
118 fps
Battlefield 5 2160p
Score
116 fps
Battlefield 5 1440p
Score
196 fps
Battlefield 5 1080p
Score
197 fps
GTA 5 2160p
Score
124 fps
GTA 5 1440p
Score
155 fps
GTA 5 1080p
Score
193 fps
FP32 (float)
Score
22.579 TFLOPS
3DMark Time Spy
Score
20661
Blender
Score
2669
Vulkan
Score
158828
OpenCL
Score
161327

Compared to Other GPU

Shadow of the Tomb Raider 2160p / fps
193 +150.6%
45 -41.6%
34 -55.8%
24 -68.8%
Shadow of the Tomb Raider 1440p / fps
292 +101.4%
67 -53.8%
49 -66.2%
Shadow of the Tomb Raider 1080p / fps
310 +63.2%
101 -46.8%
72 -62.1%
Cyberpunk 2077 2160p / fps
90 +34.3%
24 -64.2%
Cyberpunk 2077 1440p / fps
185 +146.7%
35 -53.3%
Cyberpunk 2077 1080p / fps
203 +72%
48 -59.3%
Battlefield 5 2160p / fps
194 +67.2%
56 -51.7%
Battlefield 5 1440p / fps
203 +3.6%
Battlefield 5 1080p / fps
213 +8.1%
169 -14.2%
139 -29.4%
122 -38.1%
GTA 5 2160p / fps
174 +40.3%
GTA 5 1440p / fps
191 +23.2%
73 -52.9%
GTA 5 1080p / fps
231 +19.7%
156 -19.2%
141 -26.9%
86 -55.4%
FP32 (float) / TFLOPS
27.215 +20.5%
20.441 -9.5%
3DMark Time Spy
36233 +75.4%
9097 -56%
Blender
15026.3 +463%
3514.46 +31.7%
1064 -60.1%
552 -79.3%
Vulkan
382809 +141%
91662 -42.3%
61331 -61.4%
34688 -78.2%
OpenCL
385013 +138.7%
167342 +3.7%
75816 -53%
57474 -64.4%