AMD Radeon HD 7990

AMD Radeon HD 7990

AMD Radeon HD 7990: A Legend of the Past in the Realities of 2025

Analysis of an outdated flagship for enthusiasts and nostalgia lovers


Introduction

Released in 2013, the AMD Radeon HD 7990 was the pinnacle of engineering achievement for its time. However, in 2025, this graphics card is more of a museum exhibit than a viable solution. Nevertheless, it can still be found in budget builds or with collectors. Let's explore whether it has a place in the modern world and who might still find it useful.


Architecture and Key Features

Architecture: The HD 7990 is built on two Tahiti XT chips (architecture GCN 1.0) using a 28 nm process. It is one of AMD's first "dual-chip" cards, combining two GPUs on a single board.

Unique Features:

- AMD Eyefinity — support for up to 6 monitors simultaneously.

- CrossFire X — technology for combining multiple GPUs (though unnecessary for the HD 7990, as it is already dual-chip).

- DirectX 11.1 and OpenGL 4.2 — standards of its time, but in 2025, this is insufficient for modern games with DX12 Ultimate or Vulkan Ray Tracing.

What’s Missing:

- Ray tracing (RT) and AI technologies like FSR or DLSS. The HD 7990 does not even support FidelityFX Super Resolution due to the lack of driver updates.


Memory: Advantages and Limitations

- Type and Capacity: 6 GB GDDR5 (3 GB per GPU) with a 384-bit bus per chip.

- Bandwidth: 288 GB/s for each GPU (theoretically totaling 576 GB/s, but the actual efficiency is lower due to CrossFire characteristics).

Impact on Performance:

In 2013, 6 GB was excessive, but for games in 2025, even 6 GB of GDDR5 is a serious limitation. For instance, in Cyberpunk 2077: Phantom Liberty (2024), at 1080p, the "High" texture packs require at least 4 GB of VRAM, and due to slow memory and a narrow bus, the HD 7990 suffers from FPS drops.


Gaming Performance: Nostalgia vs. Reality

Average FPS Examples (Medium Settings, 2023 Drivers — Latest Available):

- CS2 (1080p): ~90-120 FPS (but there may be lags in competitive scenes).

- Elden Ring (1080p): 25-35 FPS (lacking full DirectX 12 support).

- Hogwarts Legacy (1080p): 15-20 FPS (unplayable).

- Fortnite (1080p, without RT): 40-50 FPS.

Resolutions:

- 1080p: The only workable option for non-modernized projects (like Dota 2, Overwatch 2).

- 1440p and 4K: Not recommended — lack of memory and computational power.

Ray Tracing: Not supported at the hardware level.


Professional Tasks: It's Complicated

- Video Editing: In Premiere Pro 2025, rendering is possible via OpenCL, but the speed is 3-4 times lower than that of an RTX 4050.

- 3D Modeling: Blender Cycles with OpenCL support will show modest results — rendering a scene with 1 million polygons will take ~2 hours compared to 15 minutes on a modern RX 7600.

- Scientific Calculations: OpenCL support is available but low efficiency due to outdated instructions.

Conclusion: The HD 7990 is not suited for professional tasks in 2025, except for basic operations.


Power Consumption and Heat Dissipation

- TDP: 375W — even by 2025 standards, this is excessive.

- Cooling Recommendations:

- Minimum case with 6-8 fans for ventilation.

- Liquid cooling is justified if the card is used 24/7.

- Thermal Interface: Replacing the thermal paste is essential (it dries out after 10+ years of use).


Comparison with Competitors

2013 Analogues:

- NVIDIA GeForce GTX 690: Lower TDP (300W) but worse optimization for modern drivers. In 2025 games, the HD 7990 is slightly more stable.

Modern Budget Competitors (2025):

- AMD Radeon RX 7600 (Price: $250): 3 times the performance, support for FSR 3.0 and RT.

- NVIDIA GeForce RTX 4050 ($300): DLSS 3.5, and half the power consumption.


Practical Advice

1. Power Supply: At least 700W with an 80+ Bronze certification.

2. Compatibility:

- Requires a motherboard with PCIe 3.0 x16 (compatible with PCIe 4.0/5.0, but with no speed increase).

- Not suitable for processors consuming more than 150W (risk of overloading the PSU).

3. Drivers: Use the latest version of Adrenalin 2023 Edition — newer operating systems (like Windows 12) may not be supported.


Pros and Cons

Pros:

- Unique collectible item.

- Low price on the secondary market ($80-120).

- Support for multi-monitor setups.

Cons:

- No support for modern APIs and technologies.

- High power consumption.

- Risk of component failure (the card’s age — 12+ years).


Final Conclusion: Who Is the HD 7990 For?

1. Collectors and Enthusiasts: For retro builds or a showcase with GPU history.

2. Owners of Old PCs: If an upgrade is needed for games from the 2010s (e.g., Skyrim, GTA V on Medium).

3. Budget Users: Only if the card was obtained for free or at a symbolic price.

Important: The HD 7990 is not an investment for the future. Even a budget RX 6400 ($150) will outperform it on all fronts. But as a part of the history of computer technology — it is a legend worthy of respect.


If you find an HD 7990 in your attic or at a garage sale, try to revive it — perhaps it will become the heart of your retro project. But for modern tasks, it’s better to look at new solutions.

Basic

Label Name
AMD
Platform
Desktop
Launch Date
April 2013
Model Name
Radeon HD 7990
Generation
Southern Islands
Base Clock
950MHz
Boost Clock
1000MHz
Bus Interface
PCIe 3.0 x16
Transistors
4,313 million
Compute Units
32
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.
128
Foundry
TSMC
Process Size
28 nm
Architecture
GCN 1.0

Memory Specifications

Memory Size
3GB
Memory Type
GDDR5
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.
384bit
Memory Clock
1500MHz
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.
288.0 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.
32.00 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.
128.0 GTexel/s
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.
1024 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.
4.178 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.
2048
L1 Cache
16 KB (per CU)
L2 Cache
768KB
TDP
375W
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.
1.2
OpenCL Version
1.2
OpenGL
4.6
DirectX
12 (11_1)
Power Connectors
2x 8-pin
Shader Model
5.1
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.
32
Suggested PSU
750W

Benchmarks

FP32 (float)
Score
4.178 TFLOPS
Hashcat
Score
278176 H/s

Compared to Other GPU

FP32 (float) / TFLOPS
4.311 +3.2%
4.074 -2.5%
4.014 -3.9%
Hashcat / H/s
330579 +18.8%
304761 +9.6%
245484 -11.8%
210867 -24.2%