NVIDIA GeForce GTX 1660 Ti vs NVIDIA GeForce GTX 980 Comparison

The GTX 980 and GTX 1660 Ti represent two distinct eras of NVIDIA’s mid-range and prior flagship designs. The data reveals a generational clash where the newer card wins the majority of tests, yet the older card holds a surprising and decisive advantage in one specific API workload. This analysis breaks down the benchmark results, architectural shifts, and practical implications for each card based solely on the provided facts.

Where Each One Wins

The benchmark results paint a clear picture of specialization. The GTX 1660 Ti is the dominant performer in the vast majority of workloads, winning 9 out of 10 head-to-head comparisons. Its victories span legacy and modern APIs, including a significant 55.3% lead in the 3DMark Steel Nomad DX12 test, a 32% lead in Geekbench OpenCL, and a 13.8% advantage in Passmark G3D. This indicates a comprehensively faster card for almost all standard rendering and compute tasks.

However, the GTX 980 claims a single, monumental victory in the Geekbench Vulkan test. Its score of 22,543 is 286.9% higher than the GTX 1660 Ti’s 5,827. This is not a marginal win; it is a categorical one. The data suggests that while the newer card excels in compute and DirectX-based scenarios, the older Maxwell architecture has a massive optimization or driver advantage in this specific Vulkan workload. For users running Vulkan-based applications that stress the API similarly, the GTX 980 is the clear choice, despite its age.

Architecture Differences

The two cards are built on fundamentally different architectures and manufacturing processes. The GTX 980 uses the GM204 chip based on the Maxwell 2.0 architecture, manufactured on a 28 nm process at TSMC. In contrast, the GTX 1660 Ti uses the TU116 chip based on the Turing architecture, built on a more advanced 12 nm process, also at TSMC. This node shrink allows the newer card to pack more transistors into a smaller space: 6,600 million transistors on a 284 mm² die versus 5,200 million on a 398 mm² die. Consequently, the GTX 1660 Ti’s transistor density is 23.2M / mm², nearly double the GTX 980’s 13.1M / mm².

The core configurations also differ significantly. The GTX 980 has more raw processing units, featuring 2048 shading units, 128 TMUs, and 64 ROPs. The GTX 1660 Ti, despite its newer architecture, has fewer physical units: 1536 shading units, 96 TMUs, and 48 ROPs. This is a critical trade-off. The newer card compensates with much higher clock speeds, running a base clock of 1500 MHz and a boost of 1770 MHz, compared to the GTX 980’s 1127 MHz base and 1216 MHz boost. The higher clocks allow the GTX 1660 Ti to achieve a higher FP32 throughput of 5.437 TFLOPS versus 4.981 TFLOPS for the GTX 980, despite having fewer cores.

Neither card features dedicated ray tracing or tensor cores, as both are from the GTX line. Memory architecture also diverges. The GTX 980 uses 4 GB of GDDR5 on a 256-bit bus, providing 224.4 GB/s of bandwidth. The GTX 1660 Ti uses 6 GB of GDDR6 on a narrower 192-bit bus, but achieves a higher bandwidth of 288.0 GB/s due to the faster memory clock. The GTX 1660 Ti also supports FP16 compute at a 2:1 rate, a feature not listed for the GTX 980.

FAQ

Q: Which card has more raw shader processing power?

A: The GTX 980 has more shading units (2048) and TMUs (128) than the GTX 1660 Ti (1536 and 96, respectively). However, the GTX 1660 Ti’s higher clock speeds give it a greater FP32 performance of 5.437 TFLOPS compared to 4.981 TFLOPS.

Q: Is the newer GTX 1660 Ti more power-efficient based on the specifications?

A: Yes. The GTX 1660 Ti has a lower TDP of 120 W, while the GTX 980 has a higher TDP of 165 W. The newer card also requires a lower suggested PSU of 300 W versus the 450 W suggested for the GTX 980.

Q: Why does the GTX 980 win the Vulkan benchmark so decisively?

A: In the Geekbench Vulkan test, the GTX 980 scored 22,543, which is 286.9% higher than the GTX 1660 Ti’s score of 5,827. The data does not specify why, but it indicates a massive performance advantage for the Maxwell architecture in this specific workload.

Q: Does the GTX 1660 Ti have more memory bandwidth?

A: Yes. Despite having a narrower 192-bit memory bus, the GTX 1660 Ti’s GDDR6 memory provides 288.0 GB/s of bandwidth, which is higher than the GTX 980’s 224.4 GB/s from its 256-bit GDDR5 setup.

Q: Which card is physically larger?

A: The GTX 980 is longer and wider, measuring 267 mm in length and 40 mm in width. The GTX 1660 Ti is shorter at 229 mm and slimmer at 35 mm. Both cards are the same height at 111 mm.

Q: Which card has a higher average benchmark score?

A: The GTX 980 has a higher average benchmark score of 8167, placing it in the 43rd percentile of all GPUs. The GTX 1660 Ti has an average score of 7723 and is in the 41st percentile.

Specification Differences

• Process Node: GTX 980 is 28 nm, GTX 1660 Ti is 12 nm.
• Transistors: GTX 980 has 5,200 million, GTX 1660 Ti has 6,600 million.
• Die Size: GTX 980 is 398 mm², GTX 1660 Ti is 284 mm².
• Base Clock: GTX 980 is 1127 MHz, GTX 1660 Ti is 1500 MHz.
• Boost Clock: GTX 980 is 1216 MHz, GTX 1660 Ti is 1770 MHz.
• Memory Size: GTX 980 has 4 GB, GTX 1660 Ti has 6 GB.
• Memory Type: GTX 980 uses GDDR5, GTX 1660 Ti uses GDDR6.
• Memory Bus Width: GTX 980 is 256-bit, GTX 1660 Ti is 192-bit.
• Memory Bandwidth: GTX 980 is 224.4 GB/s, GTX 1660 Ti is 288.0 GB/s.
• Shading Units: GTX 980 has 2048, GTX 1660 Ti has 1536.
• TMUs: GTX 980 has 128, GTX 1660 Ti has 96.
• ROPs: GTX 980 has 64, GTX 1660 Ti has 48.
• FP32 Performance: GTX 980 is 4.981 TFLOPS, GTX 1660 Ti is 5.437 TFLOPS.
• FP16 Performance: GTX 980 is not listed, GTX 1660 Ti is 10.87 TFLOPS (2:1).
• TDP: GTX 980 is 165 W, GTX 1660 Ti is 120 W.
• Power Connectors: GTX 980 uses 2x 6-pin, GTX 1660 Ti uses 1x 8-pin.
• Suggested PSU: GTX 980 is 450 W, GTX 1660 Ti is 300 W.
• Dimensions (LxW): GTX 980 is 267 mm x 40 mm, GTX 1660 Ti is 229 mm x 35 mm.

Head-to-Head Benchmarks

The most striking result is in the 3DMark Steel Nomad DX12 test. The GTX 1660 Ti scores 1061, which is 55.3% higher than the GTX 980’s 474. This is the largest single delta in the entire dataset and highlights the newer card’s superior DirectX 12 performance. The GTX 1660 Ti also shows a substantial 32% lead in Geekbench OpenCL, scoring 51,029 versus 34,676, indicating better general-purpose compute performance.

In the Passmark suite, the GTX 1660 Ti wins across the board. It leads by 19.7% in DirectX 10 (66 vs 53), 19.4% in DirectX 11 (103 vs 83), and 13.7% in DirectX 9 (190 vs 164). The margins narrow in more modern APIs, with a 9.8% lead in DirectX 12 (51 vs 46). In the Passmark G3D test, which is a broad gaming metric, the GTX 1660 Ti scores 12,871, a 13.8% improvement over the GTX 980’s 11,095. The GPU Compute test shows a smaller 9.1% advantage for the GTX 1660 Ti, with scores of 5,226 and 4,753 respectively. The 2D test is nearly identical, with the GTX 1660 Ti winning by a slim 1.5% margin (804 vs 792).

The GTX 980’s sole victory is in Geekbench Vulkan, where it scores 22,543 against the GTX 1660 Ti’s 5,827. This 286.9% delta is extraordinary and completely reverses the narrative of the other tests. It suggests that the GTX 980’s Vulkan driver implementation is exceptionally efficient, or that the benchmark itself is heavily optimized for the Maxwell architecture’s specific feature set. This single result is a powerful reminder that raw architectural age does not always correlate with performance in every API.

The Verdict

The data is unequivocal for most users: the GTX 1660 Ti is the superior card for the vast majority of workloads. It wins in 9 out of 10 benchmarks, including all DirectX variants, OpenCL, and the comprehensive 3DMark test. Its higher clock speeds, greater FP32 throughput, faster memory, and lower power draw make it a more capable and efficient choice for modern gaming and general compute. The 13.8% lead in Passmark G3D and the 55.3% lead in Steel Nomad firmly establish it as the faster gaming card.

The GTX 980, however, is not without its strengths. Its monumental 286.9% victory in the Geekbench Vulkan test is a specific and significant advantage. For a user whose primary workload is Vulkan-based and mirrors the characteristics of that benchmark, the GTX 980 would be the better performer. Additionally, the GTX 980 has a higher average benchmark score (8167 vs 7723) and a slightly higher percentile ranking (43rd vs 41st), which is driven by its exceptional Vulkan result. The GTX 980 also offers more shading units and ROPs, which could theoretically benefit certain rasterization patterns, though the benchmark data does not confirm this in any other test. Ultimately, the GTX 1660 Ti is the safe, well-rounded pick, while the GTX 980 is a specialist with a single, powerful weapon.