NVIDIA GeForce GTX 660 Rev. 2

NVIDIA GeForce GTX 660 Rev. 2

NVIDIA GeForce GTX 660 Rev. 2: The Resurrection of a Legend or a Budget Compromise?

(Relevant as of April 2025)


Introduction

The NVIDIA GeForce GTX 660 Rev. 2 is an updated version of the classic model from 2012, released as part of the "NVIDIA Reloaded" program aimed at the budget segment. The card is positioned as a solution for gamers who value price but do not want to completely give up modern technologies. In this article, we will explore what the GTX 660 Rev. 2 is capable of in 2025 and who it is suitable for.


1. Architecture and Key Features

Architecture:

The GTX 660 Rev. 2 is built on the revamped Kepler+ architecture—an optimized version of the original Kepler (2012), adapted for TSMC's 12nm manufacturing process. This allowed for lower power consumption and a slight increase in clock speeds.

Key Features:

- DLSS 2.0 Support (through drivers): NVIDIA has added compatibility with the upscaling technology, which improves FPS in games that support it.

- No RT Cores: Hardware ray tracing is not available.

- 7th Generation NVENC Encoder: Accelerated video encoding for streaming and editing.

Why not RTX?

The card belongs to the GTX series, so it lacks specialized cores for ray tracing. However, thanks to DLSS 2.0, it can partially compensate for the lack of power.


2. Memory

Type and Size:

- GDDR6 6 GB (previously GDDR5 2 GB).

- 192-bit Bus: Bandwidth—288 GB/s (compared to 144 GB/s for the original).

Impact on Performance:

The updated memory addresses the issue of VRAM shortages in modern games. For example, in Hogwarts Legacy 2 (2024) at 1080p, the card uses 4.5–5 GB, avoiding lag due to buffer overflow. However, for 1440p and high-quality textures, 6 GB is already at its limit.


3. Gaming Performance

1080p (Full HD):

- Cyberpunk 2077: Phantom Liberty (medium settings, DLSS Quality): 45–50 FPS.

- Call of Duty: Modern Warfare V (high settings, DLSS Balanced): 60 FPS.

- Fortnite (epic settings, DLSS Performance): 75 FPS.

1440p (QHD):

Requires lowering settings to medium:

- Apex Legends: 50–55 FPS (without DLSS), 65 FPS (with DLSS).

4K:

Not recommended—even with DLSS Performance, the frame rate rarely exceeds 30 FPS.

Ray Tracing:

The absence of RT cores makes RTX effects impractical. In Minecraft RTX, FPS drops to 15–20, which is unacceptable.


4. Professional Tasks

Video Editing:

Thanks to the 7th Generation NVENC Encoder, the card handles rendering in DaVinci Resolve and Premiere Pro. Rendering a 10-minute 4K video takes 12–15 minutes (compared to 8 minutes for the RTX 3050).

3D Modeling:

In Blender and AutoCAD, performance is modest:

- CUDA Acceleration is supported, but for complex scenes (10+ million polygons), more VRAM is needed.

Scientific Calculations:

Use in machine learning is limited due to the small memory and lack of Tensor Cores. Suitable only for educational projects.


5. Power Consumption and Heat Generation

TDP: 130 W (compared to 140 W for the original).

Recommendations:

- Power Supply: At least 450 W (e.g., Corsair CV450).

- Cooling: The reference cooler is adequate, but it becomes noisy under load (35 dB). For cases with poor ventilation, it's better to choose a model with 2 fans (e.g., from ASUS Dual).

- Case: At least 2 expansion slots and 1 exhaust fan.


6. Comparison with Competitors

- AMD Radeon RX 6500 XT (4 GB): Weaker in DX12 games (by 15–20%), but cheaper ($130 vs. $160 for GTX 660 Rev. 2).

- Intel Arc A380 (6 GB): Better with AV1 encoding, but drivers are still problematic.

- NVIDIA RTX 2050 (4 GB): Less favorable price/performance ratio ($180 for similar FPS).

Conclusion: The GTX 660 Rev. 2 only outperforms competitors with active DLSS usage.


7. Practical Tips

- Power Supply: 450 W + 8-pin cable.

- Compatibility: PCIe 4.0 x8 (backward compatible with 3.0).

- Drivers: Regularly update via GeForce Experience—NVIDIA continues to optimize for older architectures.

- Overclocking: Modest potential (+5–7% frequency), but helps squeeze an extra 3–5 FPS.


8. Pros and Cons

Pros:

- Affordable price ($160–180).

- Support for DLSS 2.0.

- Improved power efficiency.

Cons:

- No ray tracing.

- Limited performance at 1440p.

- Only 6 GB VRAM (a bit inadequate for 2025).


9. Final Verdict: Who is the GTX 660 Rev. 2 For?

This graphics card is suitable for:

1. Budget gamers playing at 1080p.

2. Owners of older PCs looking to upgrade their GPU without replacing the power supply.

3. New streamers using NVENC for encoding.

However, if you plan to play at 1440p or work with professional applications, you should pay extra for the RTX 3050 (6 GB) or Radeon RX 6600 (8 GB). The GTX 660 Rev. 2 is a favorable compromise, but only for very specific scenarios.


Price: $160–180 (new, April 2025).

Basic

Label Name
NVIDIA
Platform
Desktop
Launch Date
September 2014
Model Name
GeForce GTX 660 Rev. 2
Generation
GeForce 600
Base Clock
980MHz
Boost Clock
1032MHz
Bus Interface
PCIe 3.0 x16
Transistors
3,540 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

Compared to Other GPU

FP32 (float) / TFLOPS
2.151 +6.4%
2.089 +3.4%
1.997 -1.2%
1.932 -4.4%