NVIDIA GeForce GTX 1080 vs NVIDIA GeForce GTX 1660 Comparison

The Verdict

The benchmark data presents a decisive overall picture: the NVIDIA GeForce GTX 1080 wins 9 of 10 head-to-head tests against the NVIDIA GeForce GTX 1660. The GTX 1080's average benchmark score of 11,960 sits 2.4% above the GTX 1660's 11,680, and its percentile ranking among all GPUs matches the GTX 1660 at the 51st percentile. For raw performance across most workloads, the GTX 1080 is the clear choice, with particularly large margins in DirectX 11 (57% ahead), DirectX 10 (52.5% ahead), and compute tasks (53.4% ahead). However, the GTX 1660 is not without its own territory: it wins the Vulkan benchmark decisively, scoring 50,137 versus the GTX 1080's 30,398, a 39.4% advantage. That single win indicates that for Vulkan-based applications, the GTX 1660 is dramatically superior, making it the pick for users whose workloads are heavily Vulkan-centric. Conversely, the GTX 1080 should be chosen by anyone prioritizing DirectX legacy titles, compute workloads, or overall 3D performance, as its PassMark G3D score of 15,586 versus 11,646 demonstrates a 33.8% lead.

Architecture Differences

The two cards come from different NVIDIA architectures and manufacturing nodes. The GTX 1080 uses the GP104 chip built on the Pascal architecture, fabricated on a 16 nm process at TSMC. The GTX 1660 uses the TU116 chip on the Turing architecture, also from TSMC but on a 12 nm node. Despite the older process, the GTX 1080 packs more transistors: 7,200 million versus 6,600 million on the GTX 1660. The die sizes differ as well, with the GTX 1080 at 314 mm² and the GTX 1660 at 284 mm², resulting in transistor densities of 22.9M per mm² and 23.2M per mm², respectively — the GTX 1660 is slightly denser.

Memory configurations diverge sharply. The GTX 1080 carries 8 GB of GDDR5X on a 256-bit bus, delivering 320.3 GB/s of bandwidth. The GTX 1660 has 6 GB of GDDR5 on a 192-bit bus, yielding 192.1 GB/s. Clock speeds also differ: the GTX 1080 has a base clock of 1607 MHz and a boost of 1733 MHz, while the GTX 1660 starts lower at 1530 MHz but boosts higher to 1785 MHz. Memory clocks show the GTX 1080 at 1251 MHz (10 Gbps effective) versus the GTX 1660's 2001 MHz (8 Gbps effective).

Core counts heavily favor the GTX 1080. It has 2,560 shading units, 160 texture mapping units, and 64 ROPs. The GTX 1660 has 1,408 shading units, 88 TMUs, and 48 ROPs. This translates to a substantial compute advantage for the GTX 1080: FP32 performance is 8.873 TFLOPS versus 5.027 TFLOPS. Interestingly, FP16 performance flips: the GTX 1080 manages only 138.6 GFLOPS (1:64 ratio), while the GTX 1660 achieves 10.05 TFLOPS (2:1 ratio). Both cards lack ray tracing and tensor cores, and both support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Power and physical specifications also differ. The GTX 1080 has a 180 W TDP with a suggested 450 W PSU, while the GTX 1660 draws 120 W and recommends a 300 W PSU. Both are dual-slot cards with a single 8-pin power connector. The GTX 1080 is longer at 267 mm versus 229 mm, and slightly wider at 40 mm versus 35 mm, though heights are nearly identical at 112 mm and 111 mm. Display outputs differ: the GTX 1080 offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a, while the GTX 1660 has 1x DVI, 1x HDMI 2.0, and only 1x DisplayPort 1.4a.

FAQ

Q: Which card has the higher average benchmark score?

A: The GTX 1080 has an average benchmark score of 11,960, which is 2.4% higher than the GTX 1660's 11,680.

Q: Does the GTX 1660 win any benchmark tests against the GTX 1080?

A: Yes, the GTX 1660 wins the Geekbench Vulkan test with a score of 50,137 versus the GTX 1080's 30,398, a 39.4% margin.

Q: How large is the GTX 1080's lead in DirectX 11 performance?

A: The GTX 1080 scores 124 in PassMark DirectX 11 versus the GTX 1660's 79, giving the GTX 1080 a 57% advantage.

Q: What is the memory bandwidth difference between the two cards?

A: The GTX 1080 delivers 320.3 GB/s over a 256-bit bus with GDDR5X, while the GTX 1660 provides 192.1 GB/s over a 192-bit bus with GDDR5.

Q: Are both cards the same physical width?

A: No, the GTX 1080 is 40 mm wide (1.6 inches) while the GTX 1660 is 35 mm wide (1.4 inches); both are dual-slot.

Q: Which card has a higher boost clock?

A: The GTX 1660 boosts to 1785 MHz, while the GTX 1080 boosts to 1733 MHz, despite the GTX 1080 having a higher base clock of 1607 MHz versus 1530 MHz.

Specification Differences

| Specification | GTX 1080 | GTX 1660 |

|---|---|---|

| Chip | GP104 | TU116 |

| Architecture | Pascal | Turing |

| Process Node | 16 nm | 12 nm |

| Transistors | 7,200 million | 6,600 million |

| Die Size | 314 mm² | 284 mm² |

| Transistor Density | 22.9M / mm² | 23.2M / mm² |

| Base Clock | 1607 MHz | 1530 MHz |

| Boost Clock | 1733 MHz | 1785 MHz |

| Memory Clock | 1251 MHz (10 Gbps effective) | 2001 MHz (8 Gbps effective) |

| Memory Size | 8 GB | 6 GB |

| Memory Type | GDDR5X | GDDR5 |

| Memory Bus | 256 bit | 192 bit |

| Memory Bandwidth | 320.3 GB/s | 192.1 GB/s |

| Shading Units | 2560 | 1408 |

| TMUs | 160 | 88 |

| ROPs | 64 | 48 |

| Pixel Rate | 110.9 GPixel/s | 85.68 GPixel/s |

| Texture Rate | 277.3 GTexel/s | 157.1 GTexel/s |

| FP32 | 8.873 TFLOPS | 5.027 TFLOPS |

| FP16 | 138.6 GFLOPS (1:64) | 10.05 TFLOPS (2:1) |

| TDP | 180 W | 120 W |

| Suggested PSU | 450 W | 300 W |

| Length | 267 mm (10.5 inches) | 229 mm (9 inches) |

| Width | 40 mm (1.6 inches) | 35 mm (1.4 inches) |

| Display Outputs | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a |

| Release Date | 2016-05-26 | 2019-03-13 |

| Launch MSRP | 599 USD | 219 USD |

Head-to-Head Benchmarks

The data reveals that the GTX 1080 dominates in nearly every tested workload, but the margins vary considerably. The largest win for the GTX 1080 comes in PassMark DirectX 11, where it scores 124 against the GTX 1660's 79 — a 57% difference. DirectX 10 shows a similar story: 93 versus 61, a 52.5% lead. Compute performance is also heavily one-sided, with the GTX 1080 posting 7,614 in PassMark GPU Compute versus 4,963, a 53.4% margin. In the 3DMark Steel Nomad DX12 test, the GTX 1080 scores 1,560 compared to 1,065, a 46.5% advantage. PassMark G3D shows a 33.8% lead (15,586 vs 11,646), and DirectX 9 shows a 19.2% edge (211 vs 177). The GTX 1080 also wins PassMark DirectX 12 by 12.2% (55 vs 49), Geekbench OpenCL by 7% (51,204 vs 47,850), and PassMark G2D by 14.4% (888 vs 776). The sole exception is Geekbench Vulkan, where the GTX 1660 wins decisively with 50,137 versus 30,398, a 39.4% margin — the only test where the Turing card's architecture provides a clear advantage. Notably, the GTX 1080's wins in DirectX 10 and 11 are particularly lopsided, suggesting legacy DirectX performance is a major strength of the Pascal architecture.

Where Each One Wins

NVIDIA GeForce GTX 1080 wins in: DirectX 9, 10, 11, and 12 workloads, with margins ranging from 12.2% to 57%. Its PassMark G3D score of 15,586 makes it the stronger choice for general 3D rendering and gaming across most APIs. Compute-heavy tasks clearly favor the GTX 1080, as its PassMark GPU Compute score of 7,614 is 53.4% ahead. The OpenCL benchmark also goes to the GTX 1080 (51,204 vs 47,850), reinforcing its compute dominance. The 3DMark Steel Nomad DX12 test shows a 46.5% lead, indicating that even modern DX12 titles will perform substantially better on the GTX 1080. Additionally, the GTX 1080's larger memory capacity (8 GB vs 6 GB) and higher bandwidth (320.3 GB/s vs 192.1 GB/s) make it the pick for texture-heavy or high-resolution workloads.

NVIDIA GeForce GTX 1660 wins in: Geekbench Vulkan, where its score of 50,137 is 39.4% higher than the GTX 1080's 30,398. This is a substantial and specific advantage for Vulkan-based applications, which may include modern game engines and compute APIs that leverage Vulkan. The GTX 1660 also has a higher boost clock (1785 MHz vs 1733 MHz) and superior FP16 performance (10.05 TFLOPS vs 138.6 GFLOPS), although the benchmark data does not include a head-to-head FP16 test. For users whose primary software stack is Vulkan-centric, the GTX 1660 is the data-backed recommendation, despite losing the overall average benchmark comparison.