NVIDIA GeForce GTX 660

NVIDIA GeForce GTX 660

NVIDIA GeForce GTX 660 in 2025: Nostalgia or Practicality?

An Analysis of an Obsolete Legend in a Modern Context


Introduction: Why is GTX 660 Still Being Discussed?

The NVIDIA GeForce GTX 660, released in 2012, became a symbol of affordable gaming in its time. However, in 2025, mentioning it raises questions: who can benefit from a 13-year-old graphics card? In this article, we will analyze its features through the lens of modern requirements and explore whether it has any chances in the era of ray tracing and AI rendering.


Architecture and Key Features

Kepler: The Foundation of a Bygone Era

The GTX 660 is built on the Kepler architecture (GK106 chip) using a 28 nm process. Unlike modern GPUs, there is no hint of hardware support for ray tracing (RTX) or DLSS here. The card focused on basic computations and DirectX 11.

Unique features (none):

- RT Cores and Tensor Cores — not available.

- DLSS, FSR, XeSS — not supported.

- Maximum DirectX version — 12 (with limited functionality).

Conclusion: The architecture is hopelessly outdated for tasks requiring parallel processing and AI acceleration.


Memory: The Minimum for Survival

- Type: GDDR5 (compared to GDDR6X/HBM in modern counterparts).

- Capacity: 2 GB (critically low for 2025 — even browsers consume more).

- Bus: 192-bit.

- Bandwidth: 144 GB/s (3-4 times lower than budget cards in 2025).

Impact on performance:

- Games from 2020+: High-resolution textures do not fit in the buffer, leading to FPS drops and stutters.

- Professional tasks: Rendering complex scenes is impossible due to memory shortages.


Gaming Performance: What Can GTX 660 Handle in 2025?

1080p Resolution (Low/Medium Settings):

- CS:2 (2023): ~40-50 FPS (without anti-aliasing).

- Fortnite (main mode): ~30-35 FPS (Performance Mode).

- GTA V (2015): ~55-60 FPS (High).

- Cyberpunk 2077 (2020): ~15-20 FPS (Low, no Ray Tracing).

1440p and 4K: Not recommended — not enough video memory and computational power.

Ray Tracing: No hardware support. Software emulation (e.g., via Proton) drops FPS to unplayable values.


Professional Tasks: CUDA at Its Limits

- Video Editing: In Premiere Pro or DaVinci Resolve, the card can only handle rendering at resolutions up to 1080p. Effects and color grading cause lag.

- 3D Modeling: Blender Cycles on CUDA works, but rendering a medium-level scene takes 5-7 times longer than on an RTX 3050.

- Scientific Calculations: Suitable only for educational projects (e.g., MATLAB for students).

Plus: Support for OpenCL 1.2 and CUDA 3.0 allows running older professional applications.


Power Consumption and Heat Generation

- TDP: 140 W (comparable to modern budget GPUs).

- Cooling Recommendations:

- A system with good ventilation is essential (2-3 case fans).

- Replace thermal paste every 2-3 years (due to drying out).

- Compatibility with Cases: The card is compact (~24 cm in length) and suitable for small form factors.


Comparison with Competitors

Analogues from 2012-2013:

- AMD Radeon HD 7870: Comparable in performance but less optimized for modern drivers.

- NVIDIA GTX 750 Ti: Less powerful but more energy-efficient.

Modern Budget Alternatives (2025):

- NVIDIA RTX 3050 (6 GB): $199 — supports DLSS 3.5, Ray Tracing, 8 GB GDDR6.

- AMD RX 7600: $229 — FSR 3.1, 8 GB GDDR6.

Conclusion: Even new $150-200 GPUs in 2025 outperform the GTX 660 by 4-5 times.


Practical Tips for Enthusiasts

1. Power Supply: At least 450 W (preferably with 80+ Bronze certification).

2. Motherboard: Requires PCIe 3.0 x16 (compatible with PCIe 4.0/5.0, but no speed increase).

3. Drivers: Official support has ended. Use modified drivers (e.g., from the NVCleanstall community) for compatibility with Windows 11.

4. Use Cases:

- Building a PC for older games (up to 2015).

- Backup card for testing components.

- Media center (supporting 4K video through decoding).


Pros and Cons of the GTX 660 in 2025

Pros:

- Extremely low price on the secondary market ($20-40).

- Low heat generation by modern standards.

- Support for legacy projects and retro gaming.

Cons:

- No support for modern APIs (DirectX 12 Ultimate, Vulkan 1.3).

- Insufficient memory even for web applications.

- Risk of failure (component wear).


Final Conclusion: Who is the GTX 660 Suitable For?

This graphics card is a choice for:

1. Retro gaming enthusiasts building PCs for 2000s games.

2. Owners of old systems where an upgrade is impossible (e.g., motherboards without UEFI).

3. Students learning the basics of working with CUDA using “hardware” from a bygone era.

Important: The GTX 660 shouldn't be considered a primary card for work or modern gaming. Its niche lies in budget niche tasks and nostalgic projects. In 2025, even a used RTX 2060 ($50-70) would be a more sensible investment.


Epilogue: Why Are We Still Writing About GTX 660?

This graphics card serves as a reminder of how quickly technology evolves. What was once a mid-range top performer 13 years ago now barely manages basic tasks. However, its longevity proves that even obsolete "hardware" can find its place in the hands of creative users.

Basic

Label Name
NVIDIA
Platform
Desktop
Launch Date
September 2012
Model Name
GeForce GTX 660
Generation
GeForce 600
Base Clock
980MHz
Boost Clock
1032MHz
Bus Interface
PCIe 3.0 x16
Transistors
2,540 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.
80
Foundry
TSMC
Process Size
28 nm
Architecture
Kepler

Memory Specifications

Memory Size
2GB
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.
192bit
Memory Clock
1502MHz
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.
144.2 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.
20.64 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.
82.56 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.
82.56 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.
2.021 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.
960
L1 Cache
16 KB (per SMX)
L2 Cache
384KB
TDP
140W
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.
1.1
OpenCL Version
3.0
OpenGL
4.6
DirectX
12 (11_0)
CUDA
3.0
Power Connectors
1x 6-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.
24
Suggested PSU
300W

Benchmarks

FP32 (float)
Score
2.021 TFLOPS
3DMark Time Spy
Score
1285
Blender
Score
126
OctaneBench
Score
28
Vulkan
Score
11719
OpenCL
Score
11135
Hashcat
Score
25551 H/s

Compared to Other GPU

FP32 (float) / TFLOPS
2.157 +6.7%
2.099 +3.9%
3DMark Time Spy
5182 +303.3%
2755 +114.4%
1769 +37.7%
Blender
1506.77 +1095.8%
848 +573%
194 +54%
OctaneBench
123 +339.3%
69 +146.4%
Vulkan
98446 +740.1%
69708 +494.8%
40716 +247.4%
18660 +59.2%
OpenCL
62821 +464.2%
38843 +248.8%
21442 +92.6%
11291 +1.4%
Hashcat / H/s
33607 +31.5%
31509 +23.3%
24493 -4.1%
23908 -6.4%