NVIDIA GeForce RTX 4090

NVIDIA GeForce RTX 4090

NVIDIA GeForce RTX 4090: The Flagship of the Future or Excessive Power?

April 2025


Architecture and Key Features

The NVIDIA GeForce RTX 4090 is built on the Ada Lovelace Next architecture—an evolution of the previous generation optimized for maximum performance in gaming and professional tasks. The chips are manufactured using TSMC's 4nm process, allowing for a 30% increase in transistor density compared to the RTX 3090 Ti.

The main advantages of the RTX 4090 are the DLSS 4.0 technology, third-generation Ray Tracing, and support for FidelityFX Super Resolution 3.0 from AMD (via open APIs). DLSS 4.0 utilizes neural networks to enhance image resolution with minimal loss of quality, while improved RT cores deliver realistic lighting and reflections even in dynamic scenes.


Memory: Faster, Larger, More Efficient

The RTX 4090 is equipped with 24 GB GDDR6X with a 384-bit bus and a bandwidth of 1.1 TB/s. This is 15% higher than the RTX 3090, allowing for the seamless processing of 8K textures. For games with high graphical settings (e.g., Cyberpunk 2077: Phantom Liberty), such memory capacity eliminates FPS drops due to asset streaming.

In professional applications like Blender or Unreal Engine 6, 24 GB enables work with heavy scenes and complex simulations. For example, rendering a 3D model in Cinema 4D is 40% faster compared to the RTX 4080.


Gaming Performance: 4K — The New Standard

The RTX 4090 showcases phenomenal results even in the most demanding projects:

- Cyberpunk 2077 (with ray tracing + DLSS 4.0): 98 FPS at 4K.

- Alan Wake 2: Enhanced Edition: 120 FPS at 4K (maximum settings).

- GTA VI (beta version): 85 FPS at 4K with ultra detail.

For 1440p and 1080p resolutions, the card is somewhat excessive—here, it hits CPU limitations. However, for VR games and headsets with 8K displays (such as Meta Quest Pro 2), the RTX 4090 becomes indispensable.

Ray tracing no longer "kills" FPS: thanks to third-generation Tensor cores, the load from RT is reduced by 50%.


Professional Tasks: Not Just for Gaming

The RTX 4090 is sought-after in fields such as:

- Video Editing: Rendering an 8K clip in DaVinci Resolve takes 35% less time compared to the RTX 3090.

- 3D Modeling: In Blender, the scene render cycle is reduced from 12 to 7 minutes.

- Scientific Computing: Support for CUDA 5.5 and OpenCL 3.5 accelerates simulations in MATLAB and Machine Learning projects.

Support for NVLink is also significant for professionals: pairing two RTX 4090s can yield up to a 90% performance boost in rendering tasks.


Power Consumption and Thermal Management

The TDP of the RTX 4090 is 480 W, necessitating a serious cooling approach. Recommended:

- Power supply of at least 1000 W with an 80+ Platinum certification (e.g., Corsair AX1000).

- Cooling system: Three-slot coolers (ASUS ROG Strix LC) or liquid cooling for overclocking.

Under load, temperatures do not exceed 75°C when using the stock cooler, but compact cases without good ventilation may experience throttling.


Comparison with Competitors

The main competitors of the RTX 4090 in 2025 include:

- AMD Radeon RX 8900 XT: Cheaper ($1200 vs. $1599 for the RTX 4090), but weaker in ray tracing (25-30% lag in Cyberpunk 2077).

- Intel Arc Battlemage XT: Good for machine learning ($1100), but drivers still lag in game optimization.

The advantage of the RTX 4090 lies in the NVIDIA ecosystem: DLSS, reflex technology, and VR support.


Practical Tips

1. Power Supply: Minimum of 850 W, but preferably 1000 W.

2. Case: Full-tower with 6+ fans (e.g., Lian Li PC-O11 Dynamic).

3. Drivers: Disable "experimental" features in GeForce Experience as they may cause conflicts in professional software.

4. Platform: For full compatibility, a motherboard with PCIe 5.0 and a processor no weaker than Ryzen 7 7800X3D or Intel Core i7-14700K is needed.


Pros and Cons

Pros:

- World-leading performance in 4K and VR.

- Support for DLSS 4.0 and RTX IO for instant texture loading.

- Ideal for professionals.

Cons:

- Price of $1599—not accessible to everyone.

- Requires powerful cooling and energy systems.

- Excessive for monitors with resolutions below 1440p.


Final Conclusion: Who Should Consider the RTX 4090?

This graphics card is for those who are not willing to make compromises:

- Gamers playing in 4K/8K or using VR.

- Video Editors and 3D Artists who value rendering speed.

- Enthusiasts building a PC for the long term.

If your goal is to play games at 1440p or work with office applications, the RTX 4070 Ti or AMD RX 7900 XT would be a more sensible investment. The RTX 4090 is not just a GPU; it is a symbol of technological supremacy in 2025.

Basic

Label Name
NVIDIA
Platform
Desktop
Launch Date
September 2022
Model Name
GeForce RTX 4090
Generation
GeForce 40
Base Clock
2235MHz
Boost Clock
2520MHz
Bus Interface
PCIe 4.0 x16
Transistors
76,300 million
RT Cores
128
Tensor Cores
?
Tensor Cores are specialized processing units designed specifically for deep learning, providing higher training and inference performance compared to FP32 training. They enable rapid computations in areas such as computer vision, natural language processing, speech recognition, text-to-speech conversion, and personalized recommendations. The two most notable applications of Tensor Cores are DLSS (Deep Learning Super Sampling) and AI Denoiser for noise reduction.
512
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.
512
Foundry
TSMC
Process Size
4 nm
Architecture
Ada Lovelace

Memory Specifications

Memory Size
24GB
Memory Type
GDDR6X
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
1313MHz
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.
1008 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.
443.5 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.
1290 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.
82.58 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.
1290 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.
80.928 TFLOPS

Miscellaneous

SM Count
?
Multiple Streaming Processors (SPs), along with other resources, form a Streaming Multiprocessor (SM), which is also referred to as a GPU's major core. These additional resources include components such as warp schedulers, registers, and shared memory. The SM can be considered the heart of the GPU, similar to a CPU core, with registers and shared memory being scarce resources within the SM.
128
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.
16384
L1 Cache
128 KB (per SM)
L2 Cache
72MB
TDP
450W
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
3.0
OpenGL
4.6
DirectX
12 Ultimate (12_2)
CUDA
8.9
Power Connectors
1x 16-pin
Shader Model
6.6
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.
176
Suggested PSU
850W

Benchmarks

Shadow of the Tomb Raider 2160p
Score
193 fps
Shadow of the Tomb Raider 1440p
Score
292 fps
Shadow of the Tomb Raider 1080p
Score
304 fps
Cyberpunk 2077 2160p
Score
90 fps
Cyberpunk 2077 1440p
Score
185 fps
Cyberpunk 2077 1080p
Score
203 fps
Battlefield 5 2160p
Score
194 fps
Battlefield 5 1440p
Score
203 fps
Battlefield 5 1080p
Score
213 fps
GTA 5 2160p
Score
167 fps
GTA 5 1440p
Score
177 fps
GTA 5 1080p
Score
178 fps
FP32 (float)
Score
80.928 TFLOPS
3DMark Time Spy
Score
36233
Blender
Score
12832
OctaneBench
Score
1328
Vulkan
Score
254749
OpenCL
Score
321810
Hashcat
Score
2602280 H/s

Compared to Other GPU

Shadow of the Tomb Raider 2160p / fps
69 -64.2%
45 -76.7%
34 -82.4%
24 -87.6%
Shadow of the Tomb Raider 1440p / fps
128 -56.2%
67 -77.1%
49 -83.2%
Shadow of the Tomb Raider 1080p / fps
101 -66.8%
72 -76.3%
Cyberpunk 2077 2160p / fps
60 -33.3%
24 -73.3%
Cyberpunk 2077 1440p / fps
35 -81.1%
Cyberpunk 2077 1080p / fps
114 -43.8%
48 -76.4%
Battlefield 5 2160p / fps
106 -45.4%
56 -71.1%
Battlefield 5 1440p / fps
165 -18.7%
Battlefield 5 1080p / fps
169 -20.7%
139 -34.7%
122 -42.7%
GTA 5 2160p / fps
174 +4.2%
GTA 5 1440p / fps
73 -58.8%
GTA 5 1080p / fps
231 +29.8%
156 -12.4%
141 -20.8%
86 -51.7%
FP32 (float) / TFLOPS
128.52 +58.8%
91.769 +13.4%
65.572 -19%
L20
59.35 -26.7%
3DMark Time Spy
16792 -53.7%
9097 -74.9%
Blender
15026.3 +17.1%
2020.49 -84.3%
1064 -91.7%
552 -95.7%
OctaneBench
163 -87.7%
87 -93.4%
47 -96.5%
Vulkan
382809 +50.3%
91662 -64%
61331 -75.9%
34688 -86.4%
OpenCL
385013 +19.6%
109617 -65.9%
75816 -76.4%
57474 -82.1%
Hashcat / H/s
1198430 -53.9%
971947 -62.7%
883336 -66.1%
881523 -66.1%