Quick Answer
The GeForce GTX 1660 and GTX 1070 are graphics cards from different generations with comparable performance. The GTX 1070, being a previous-generation high-end model, typically offers slightly better performance, especially at higher resolutions. The GTX 1660, a newer mid-range card, provides a more power-efficient architecture and often a more accessible price point.
GeForce GTX 1660 vs GeForce GTX 1070: Full Comparison
Introduction
When looking for graphics performance in a certain tier, it’s common to compare cards from different generations. The NVIDIA GeForce GTX 1070, launched in 2016, was a high-performance option in its time. The GeForce GTX 1660, released in 2019, arrived as a mid-range contender based on a newer architecture. This comparison breaks down their specifications, performance, and features to help clarify where each card stands, which is useful for those considering options in the used or previous-generation market.
Architecture and Specifications
The core difference lies in their underlying technology. The GTX 1070 is based on the older Pascal architecture (GP104 GPU), while the GTX 1660 utilizes the newer Turing architecture (TU116 GPU), albeit without dedicated RT or Tensor cores for ray tracing and DLSS.
- GPU Cores: The GTX 1070 features 1920 CUDA cores. The GTX 1660 has 1408 CUDA cores, but its newer Turing cores are generally more efficient per core.
- Memory: This is a key differentiator. The GTX 1070 uses 8GB of GDDR5 memory on a 256-bit bus, offering high bandwidth. The GTX 1660 uses 6GB of GDDR5 memory on a 192-bit bus.
- Process Node: The GTX 1660 is built on a more advanced 12nm process compared to the GTX 1070’s 16nm, contributing to better power efficiency.
Performance Comparison
In real-world gaming, the performance between these two cards is often close, with outcomes varying by title and resolution.
- 1080p Gaming: Both cards are capable of delivering high frame rates in most titles at 1080p resolution. The GTX 1070 typically holds a small lead, often in the range of 5-10%, due to its higher memory bandwidth and core count.
- 1440p Gaming: At 1440p, the GTX 1070’s 8GB of VRAM and wider memory bus can provide a more noticeable advantage in games that use more texture data, leading to more consistent performance.
- Power Efficiency: The GTX 1660 has a clear advantage here, with a lower typical board power (120W vs 150W for the GTX 1070), which can result in less heat and lower power supply requirements.
Features and Technologies
While both support core NVIDIA technologies like Ansel and G-Sync, the architectural generation brings some differences.
- Turing NVENC: The GTX 1660 includes an improved hardware video encoder (NVENC) based on the Turing architecture, which can provide better quality and efficiency for video recording and streaming compared to the Pascal encoder in the GTX 1070.
- Advanced Shading: The Turing architecture in the GTX 1660 supports concurrent execution of floating-point and integer operations, which can improve performance in certain modern game engines.
- Ray Tracing & DLSS: Neither card has dedicated hardware for real-time ray tracing (RT cores) or AI-accelerated super sampling (Tensor cores). These features are reserved for the RTX series.
Comparison Table
| Feature | GeForce GTX 1660 | GeForce GTX 1070 |
|---|---|---|
| Launch Date | March 2019 | June 2016 |
| GPU Architecture | Turing (TU116) | Pascal (GP104) |
| Process Node | 12nm | 16nm |
| CUDA Cores | 1408 | 1920 |
| VRAM | 6GB GDDR5 | 8GB GDDR5 |
| Memory Bus | 192-bit | 256-bit |
| Memory Bandwidth | 192.1 GB/s | 256 GB/s |
| Base / Boost Clock | ~1530 MHz / ~1785 MHz | ~1506 MHz / ~1683 MHz |
| Typical Board Power (TBP) | 120W | 150W |
| Recommended PSU | 450W | 500W |
| Video Encoder (NVENC) | 7th Gen (Turing) | 6th Gen (Pascal) |
| DirectX Support | 12 Ultimate (Feature Level 12_1) | 12 (Feature Level 12_1) |
| Ports (Typical) | 1x DisplayPort, 1x HDMI, 1x DVI-D | 1x DisplayPort, 1x HDMI, 1x DVI-D |
Frequently Asked Questions (FAQ)
Which is faster, the GTX 1660 or GTX 1070?
In most gaming benchmarks, the GTX 1070 is generally slightly faster than the GTX 1660, particularly at resolutions like 1440p where its larger 8GB VRAM and higher memory bandwidth provide an advantage. The performance difference at 1080p is typically smaller.
Does the GTX 1660 have ray tracing?
No. The GeForce GTX 1660 does not include dedicated RT Cores for hardware-accelerated ray tracing. While some ray tracing effects can be run on its standard CUDA cores, the performance impact is usually very significant, making it impractical for gaming.
Is the 6GB VRAM on the GTX 1660 a limitation?
For 1080p gaming, 6GB of VRAM is generally sufficient for most titles. At 1440p or with heavily modded games, the 8GB on the GTX 1070 can be beneficial and may prevent texture-related stuttering in more demanding scenarios.
Which card is more power-efficient?
The GTX 1660 is more power-efficient, with a lower typical board power (120W vs 150W). This can lead to lower heat output and potentially quieter operation, and it may allow for use in systems with less robust power supplies.
Final Thoughts
The choice between the GeForce GTX 1660 and GTX 1070 often comes down to availability and specific needs. The GTX 1070 typically offers marginally higher raw performance, especially at higher resolutions, and comes with more VRAM. The GTX 1660, as a newer generation card, provides better power efficiency, an improved video encoder for content creators, and can often be found at a lower cost in many markets. For users focused on 1080p gaming, both are capable options, with the decision potentially leaning on factors like power consumption, heat, and local pricing rather than a decisive performance gap.