NVIDIA GeForce GTX 970

NVIDIA GeForce GTX 970

NVIDIA GeForce GTX 970 in 2025: Nostalgia or Relevance?

Introduction

Released in 2014, the NVIDIA GeForce GTX 970 has become a legend among gamers thanks to its price-to-performance ratio. However, after 11 years, its capabilities are being called into question. Let's explore how relevant this card is in 2025, how it handles modern tasks, and whether it should be considered for a PC build.


1. Architecture and Key Features

Maxwell Architecture: A Modest Foundation

The GTX 970 is built on the Maxwell architecture (GM204), developed using a 28nm process. This generation brought a 40% improvement in energy efficiency compared to the previous Kepler series. However, modern GPUs such as the RTX 40 series utilize a 4nm process and the Ada Lovelace architecture.

Lack of Modern Technologies

The GTX 970 does not support ray tracing (RTX), DLSS, or FidelityFX Super Resolution. These features have become standard in gaming since 2018, but their operation requires hardware blocks (like RT cores) that Maxwell lacks.

Key Features from 2014

- Dynamic Super Resolution (DSR): Enhances image clarity in games.

- MFAA: Anti-aliasing with lower resource usage.

- ShadowPlay: Gameplay recording with minimal impact on FPS.


2. Memory: A Controversial Aspect

GDDR5 and the "3.5 GB Problem"

The card is equipped with 4 GB of GDDR5 memory, but due to its non-standard configuration, access to 512 MB is slower. This has led to FPS drops in games that use more than 3.5 GB (for instance, Cyberpunk 2077 even on low settings in 2025 requires 6 GB).

Bandwidth

The 256-bit bus and 7 GHz frequency provide 224 GB/s. In comparison, the RTX 4060 with GDDR6 (192-bit, 17 GHz) offers 272 GB/s. In modern games with high-resolution textures, the GTX 970 lags due to its narrow bandwidth.


3. Gaming Performance

1080p: Acceptable for Older Titles

- CS2: 120–150 FPS (low settings).

- The Witcher 3: 45–55 FPS (medium settings).

- GTA V: 60–70 FPS (high settings).

1440p and 4K: Not Recommended

Even in less demanding games like Fortnite, average FPS drops to 30–40 at 1440p. The card is unfit for 4K gaming.

Ray Tracing: No Support

All games with RTX (such as Alan Wake II) on GTX 970 either run without ray tracing or require software emulation, reducing FPS to 10–15.


4. Professional Tasks

CUDA: Limited Assistance

The card features 1664 CUDA cores, allowing for accelerated rendering in Blender or Adobe Premiere Pro. However, the 4 GB of memory is a serious limitation. For example, rendering a 4K scene may fail due to insufficient VRAM.

Scientific Computing

For machine learning or simulation tasks, the GTX 970 falls behind even budget modern cards. For instance, the RTX 3050 (8 GB GDDR6) is 3–4 times faster thanks to its Ampere architecture.


5. Power Consumption and Heat Generation

TDP of 145W: Modest, but Needs Attention

The recommended power supply is 500W (for example, Corsair CX550). The card is compatible with most Mid-Tower cases but requires good ventilation.

Cooling

The reference version was equipped with a turbine, but partner models (ASUS Strix, MSI Gaming) offered 2–3 fans. In 2025, even used cards should be cleaned of dust and have their thermal paste replaced.


6. Comparison with Competitors

AMD Radeon R9 390 (2015)

- 8 GB GDDR5, GCN 1.1 architecture.

- Better performance in games that use a lot of VRAM.

- Higher power consumption (275W TDP).

NVIDIA GTX 1060 (2016)

- 6 GB GDDR5, Pascal architecture.

- 15–20% faster in DX12 games.

- Support for newer drivers.

In 2025, Competitors — Budget Newcomers

- Intel Arc A580 (8 GB GDDR6): 120–150 FPS at 1080p, supports XeSS.

- AMD RX 6600 (8 GB GDDR6): 100–130 FPS at 1080p, priced at $199.


7. Practical Tips

Power Supply

Minimum 500W with an 8-pin cable. Avoid cheap noname models — unstable voltage can damage the card.

Compatibility with Platforms

- PCIe 3.0 x16: works on PCIe 4.0/5.0, but without speed gain.

- OS Support: Current drivers are only available for Windows 10. Windows 11 and Linux require manual configuration.

Drivers

NVIDIA discontinued official support for the GTX 900 series in 2021. The last stable version is 472.12. New games may encounter issues.


8. Pros and Cons

Pros

- Low price on the secondary market ($50–80).

- Energy efficiency for its class.

- Reliability (if kept cool).

Cons

- Only 4 GB of memory (3.5 GB is fast).

- No support for ray tracing and DLSS.

- Outdated drivers.


9. Final Conclusion: Who Should Consider the GTX 970?

This card is suitable for:

1. Budget Builds: If you need a PC for office tasks and older games.

2. Retro Gamers: To run projects from the 2010s on medium settings.

3. Temporary Solution: While you save for an RTX 4060 or RX 7600.

Why Not to Choose GTX 970 in 2025?

Even budget newcomers (like the Intel Arc A380 at $120) offer support for modern technologies and warranty. The GTX 970 is an option for those willing to accept limitations in exchange for nostalgia or an ultra-low budget.


If you find a GTX 970 in perfect condition for $50 — give it a try. But remember: the future belongs to GPUs with AI accelerators and larger memory capacities.

Basic

Label Name
NVIDIA
Platform
Desktop
Launch Date
September 2014
Model Name
GeForce GTX 970
Generation
GeForce 900
Base Clock
1050MHz
Boost Clock
1178MHz
Bus Interface
PCIe 3.0 x16
Transistors
5,200 million
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.
104
Foundry
TSMC
Process Size
28 nm
Architecture
Maxwell 2.0

Memory Specifications

Memory Size
4GB
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.
256bit
Memory Clock
1753MHz
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.
224.4 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.
65.97 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.
122.5 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.
122.5 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.
3.842 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.
1664
L1 Cache
48 KB (per SMM)
L2 Cache
2MB
TDP
148W
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 (12_1)
CUDA
5.2
Power Connectors
2x 6-pin
Shader Model
6.4
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.
56
Suggested PSU
300W

Benchmarks

Shadow of the Tomb Raider 2160p
Score
15 fps
Shadow of the Tomb Raider 1440p
Score
29 fps
Shadow of the Tomb Raider 1080p
Score
41 fps
GTA 5 2160p
Score
43 fps
GTA 5 1440p
Score
45 fps
GTA 5 1080p
Score
96 fps
FP32 (float)
Score
3.842 TFLOPS
3DMark Time Spy
Score
3708
Blender
Score
318
OctaneBench
Score
76
Vulkan
Score
31919
OpenCL
Score
26896
Hashcat
Score
157087 H/s

Compared to Other GPU

Shadow of the Tomb Raider 2160p / fps
26 +73.3%
1 -93.3%
Shadow of the Tomb Raider 1440p / fps
95 +227.6%
75 +158.6%
54 +86.2%
Shadow of the Tomb Raider 1080p / fps
141 +243.9%
107 +161%
79 +92.7%
46 +12.2%
GTA 5 2160p / fps
146 +239.5%
68 +58.1%
55 +27.9%
GTA 5 1440p / fps
103 +128.9%
82 +82.2%
62 +37.8%
GTA 5 1080p / fps
213 +121.9%
136 +41.7%
FP32 (float) / TFLOPS
4.073 +6%
3.981 +3.6%
3.612 -6%
3.454 -10.1%
3DMark Time Spy
4844 +30.6%
2325 -37.3%
1506 -59.4%
Blender
1506.77 +373.8%
848 +166.7%
45.58 -85.7%
OctaneBench
345 +353.9%
150 +97.4%
22 -71.1%
Vulkan
98446 +208.4%
69708 +118.4%
40716 +27.6%
5522 -82.7%
OpenCL
66428 +147%
46137 +71.5%
13442 -50%
8849 -67.1%
Hashcat / H/s
160182 +2%
157126 +0%
154346 -1.7%
151963 -3.3%