NVIDIA GeForce GTX 1080 Ti vs NVIDIA GeForce GTX 1660 SUPER Comparison
NVIDIA GeForce GTX 1080 Ti
GeForce GTX 1660 SUPER
PERFORMANCE BENCHMARKS
Analysis: NVIDIA GeForce GTX 1080 Ti vs NVIDIA GeForce GTX 1660 SUPER
Head-to-Head Benchmarks
The benchmark data presents a clear overall winner, but one unexpected reversal complicates the story. Across the ten head-to-head tests, the GTX 1080 Ti wins nine, while the GTX 1660 SUPER takes one. The margin of victory in those nine wins ranges from modest to overwhelming.
The largest gap appears in compute-focused workloads. In the Passmark GPU Compute test, the GTX 1080 Ti scores 9632 against 5076 for the GTX 1660 SUPER, a delta of 89.8%. That is nearly double the throughput. This result aligns with the raw specification difference in shading units, which we will examine later. The DX10 legacy test shows a similar pattern: 118 versus 65, a delta of 81.5%. These are not close calls; the older flagship simply outclasses the newer mid-range part in raw parallel execution.
The most mainstream gaming metric, Passmark G3D, shows the GTX 1080 Ti scoring 18600 against 12699, a delta of 46.5%. This is a substantial lead, but not as extreme as the compute gap. The DX11 test, still relevant for many older titles, shows a 45.2% delta (151 versus 104). DX12 narrows the gap somewhat: 66 versus 50, a delta of 32%. This suggests that the newer architecture of the GTX 1660 SUPER helps it close the distance under modern API overhead, but it does not eliminate the raw performance deficit.
The 3DMark Steel Nomad DX12 test shows a 74.6% delta (2231 versus 1278). This is a demanding, modern workload, and the GTX 1080 Ti wins decisively. OpenCL also favors the older card by 29.4% (67929 versus 52490).
The single GTX 1660 SUPER win is in Geekbench Vulkan. Here, the newer card scores 57102, while the GTX 1080 Ti manages 40511. That is a 29.1% swing in favor of the GTX 1660 SUPER. This is a notable outlier. Vulkan is a low-level API that can benefit from driver optimizations and architectural efficiencies. The Turing architecture's newer scheduler and cache hierarchy may be better suited to Vulkan's explicit control model. The data suggests that in Vulkan-specific titles, the GTX 1660 SUPER could be the better choice, despite losing everywhere else.
The remaining tests are closer. Passmark G2D shows 939 versus 805, a delta of 16.6%. DX9 shows 231 versus 189, a delta of 22.2%. These are wins for the GTX 1080 Ti, but the margins are not enormous. The overall pattern is clear: the GTX 1080 Ti is the dominant performer in the vast majority of workloads, with the GTX 1660 SUPER's Vulkan result being the sole exception.
Architecture Differences
The two cards come from different architectural generations and different process nodes. The GTX 1080 Ti uses the GP102 chip, built on the Pascal architecture, fabricated on a 16 nm process at TSMC. The GTX 1660 SUPER uses the TU116 chip, built on the Turing architecture, fabricated on a 12 nm process, also at TSMC. This node shrink is significant, but it does not translate into a performance advantage for the newer card.
The transistor counts tell a story of scale. The GTX 1080 Ti packs 11,800 million transistors on a 471 mm² die, yielding a transistor density of 25.1M per mm². The GTX 1660 SUPER has 6,600 million transistors on a 284 mm² die, with a density of 23.2M per mm². The GTX 1080 Ti is a much larger, more complex chip. The density difference is minor, but the absolute scale is roughly 1.8x in transistor count.
The compute resources differ dramatically. The GTX 1080 Ti has 3584 shading units, 224 texture mapping units (TMUs), and 88 raster operation units (ROPs). The GTX 1660 SUPER has 1408 shading units, 88 TMUs, and 48 ROPs. This is a 2.5x difference in shader count, a 2.5x difference in TMUs, and a 1.8x difference in ROPs. These numbers explain the massive compute and texture rate gaps.
The memory subsystems are also very different. The GTX 1080 Ti uses 11 GB of GDDR5X on a 352-bit bus, delivering 484.4 GB/s of bandwidth. The GTX 1660 SUPER uses 6 GB of GDDR6 on a 192-bit bus, delivering 336.0 GB/s. The GTX 1080 Ti has a 44% bandwidth advantage. The memory clock rates are specified as 1376 MHz (11 Gbps effective) for the GTX 1080 Ti and 1750 MHz (14 Gbps effective) for the GTX 1660 SUPER. The newer GDDR6 runs faster per pin, but the wider bus of the older card wins overall.
The FP32 compute rates reflect the shader count disparity: 11.34 TFLOPS for the GTX 1080 Ti versus 5.027 TFLOPS for the GTX 1660 SUPER. Interestingly, the FP16 rates are reversed in character. The GTX 1080 Ti has a low FP16 rate of 177.2 GFLOPS, which is 1:64 of its FP32 rate. The GTX 1660 SUPER has an FP16 rate of 10.05 TFLOPS, which is 2:1 of its FP32 rate. This means the GTX 1660 SUPER can handle FP16 workloads much faster relative to its FP32, but in absolute terms, the GTX 1080 Ti's FP32 is still more than double.
Power consumption is a major differentiator. The GTX 1080 Ti has a TDP of 250 W and requires a 600 W suggested PSU. The GTX 1660 SUPER has a TDP of 125 W and requires a 300 W suggested PSU. The newer card uses exactly half the power. The GTX 1080 Ti needs two power connectors (1x 6-pin + 1x 8-pin), while the GTX 1660 SUPER needs only one 8-pin. Physical dimensions also differ: the GTX 1080 Ti is 267 mm long, 112 mm high, and 40 mm wide. The GTX 1660 SUPER is 229 mm long, 111 mm high, and 35 mm wide. The newer card is shorter and slimmer.
Both cards support the same API levels: DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Neither has ray tracing cores or tensor cores. The display outputs differ: the GTX 1080 Ti has 1x HDMI 2.0 and 3x DisplayPort 1.4a, while the GTX 1660 SUPER has 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a.
FAQ
Q: Which card is faster in the majority of benchmark tests?
A: The GTX 1080 Ti wins 9 out of 10 head-to-head tests. It leads in 3DMark Steel Nomad, OpenCL, DirectX 9, 10, 11, 12, G2D, G3D, and GPU Compute.
Q: Is there any test where the GTX 1660 SUPER wins?
A: Yes, the GTX 1660 SUPER wins the Geekbench Vulkan test with a score of 57102 versus 40511, a delta of 29.1% in its favor.
Q: How large is the average performance gap between the two cards?
A: The average benchmark score for the GTX 1080 Ti is 15548, while the GTX 1660 SUPER averages 12986. This represents a roughly 19.7% lead for the GTX 1080 Ti in average score across all tests.
Q: What is the biggest single-test advantage for either card?
A: The largest delta is in Passmark GPU Compute, where the GTX 1080 Ti leads by 89.8%, scoring 9632 versus 5076. The largest win for the GTX 1660 SUPER is its 29.1% Vulkan advantage.
Q: How do the memory configurations compare?
A: The GTX 1080 Ti has 11 GB of GDDR5X on a 352-bit bus with 484.4 GB/s bandwidth. The GTX 1660 SUPER has 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth. The GTX 1080 Ti has a 44% bandwidth advantage.
Q: How do the power requirements differ?
A: The GTX 1080 Ti has a TDP of 250 W and a suggested PSU of 600 W. The GTX 1660 SUPER has a TDP of 125 W and a suggested PSU of 300 W. The GTX 1660 SUPER draws exactly half the power.
The Verdict
The data supports a straightforward conclusion for most use cases. The GTX 1080 Ti is the superior performer in raw compute, DirectX workloads, and overall 3D rendering. Its average benchmark score of 15548 places it in the 58th percentile of all GPUs, while the GTX 1660 SUPER sits at 12986 and the 53rd percentile. The GTX 1080 Ti's nearest rivals include the RTX 2060 (delta of 1.7%) and the GTX 580 (delta of 1.7%), while the GTX 1660 SUPER is closest to the RTX 3050 Ti Mobile (delta of 0.4%) and the RX 580 (delta of 0.4%).
For users who prioritize maximum frame rates in DirectX-based games, the GTX 1080 Ti is the clear choice. Its wins in DX9, DX10, DX11, and DX12 tests are consistent and often large. The 46.5% lead in G3D is decisive. The 32% lead in DX12 suggests that even in modern titles, the older card maintains a significant advantage.
However, the Vulkan result is a genuine exception. The GTX 1660 SUPER's 29.1% lead in Geekbench Vulkan indicates that Vulkan-native games or workloads could run noticeably better on the newer card. This is not a trivial outlier; it is a 57000+ score versus 40000+, which is a substantial difference. Anyone whose primary workload is Vulkan-based should consider the GTX 1660 SUPER.
Power efficiency is another consideration. The GTX 1660 SUPER uses 125 W versus 250 W. For a system with a 300 W PSU or a small form factor case, the GTX 1660 SUPER is the only viable option. The GTX 1080 Ti requires a 600 W PSU and a longer chassis (267 mm versus 229 mm).
The GTX 1080 Ti is end-of-life, as is the GTX 1660 SUPER. Both have successors in the GeForce 20 series. The launch MSRP for the GTX 1080 Ti was 699 USD. The launch MSRP for the GTX 1660 SUPER was 229 USD.
Specification Differences
| Specification | GTX 1080 Ti | GTX 1660 SUPER |
|---|---|---|
| Architecture | Pascal | Turing |
| Process Node | 16 nm | 12 nm |
| Transistors | 11,800 million | 6,600 million |
| Die Size | 471 mm² | 284 mm² |
| Transistor Density | 25.1M / mm² | 23.2M / mm² |
| Base Clock | 1481 MHz | 1530 MHz |
| Boost Clock | 1582 MHz | 1785 MHz |
| Memory Clock | 1376 MHz (11 Gbps effective) | 1750 MHz (14 Gbps effective) |
| Memory Size | 11 GB | 6 GB |
| Memory Type | GDDR5X | GDDR6 |
| Memory Bus Width | 352 bit | 192 bit |
| Memory Bandwidth | 484.4 GB/s | 336.0 GB/s |
| Shading Units | 3584 | 1408 |
| TMUs | 224 | 88 |
| ROPs | 88 | 48 |
| Pixel Rate | 139.2 GPixel/s | 85.68 GPixel/s |
| Texture Rate | 354.4 GTexel/s | 157.1 GTexel/s |
| FP32 | 11.34 TFLOPS | 5.027 TFLOPS |
| FP16 | 177.2 GFLOPS (1:64) | 10.05 TFLOPS (2:1) |
| TDP | 250 W | 125 W |
| Power Connectors | 1x 6-pin + 1x 8-pin | 1x 8-pin |
| Suggested PSU | 600 W | 300 W |
| Display Outputs | 1x HDMI 2.0, 3x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a |
| Length | 267 mm (10.5 inches) | 229 mm (9 inches) |
| Height | 112 mm (4.4 inches) | 111 mm (4.4 inches) |
| Width | 40 mm (1.6 inches) | 35 mm (1.4 inches) |
| Release Date | 2017-03-09 | 2019-10-28 |
Where Each One Wins
The GTX 1080 Ti wins in nearly every measured category. Its victories span legacy APIs (DX9 by 22.2%, DX10 by 81.5%), modern APIs (DX11 by 45.2%, DX12 by 32%), compute (GPU Compute by 89.8%, OpenCL by 29.4%), and overall 3D rendering (G3D by 46.5%, Steel Nomad by 74.6%). It also wins in 2D performance (G2D by 16.6%). This card is the dominant choice for any DirectX-based gaming, content creation using OpenCL, or general GPGPU compute. Its 3584 shading units and 484.4 GB/s of bandwidth are simply too much for the GTX 1660 SUPER to overcome in most scenarios.
The GTX 1660 SUPER wins only in Geekbench Vulkan. The score of 57102 versus 40511 is a 29.1% advantage. This is the sole bright spot. Users who run Vulkan-native games or Vulkan-based compute workloads may see better performance on the GTX 1660 SUPER. This could be due to the Turing architecture's more modern scheduling or driver optimizations for the newer chip. The GTX 1660 SUPER also has a massive advantage in power efficiency, using exactly half the power (125 W versus 250 W). This makes it the right choice for low-power builds, small form factor cases, or systems with a 300 W PSU.
The GTX 1660 SUPER also has a higher boost clock (1785 MHz versus 1582 MHz) and a higher base clock (1530 MHz versus 1481 MHz), but the GTX 1080 Ti's sheer scale of execution units overwhelms this clock speed advantage. The GTX 1660 SUPER's FP16 performance (10.05 TFLOPS) is far superior to the GTX 1080 Ti's FP16 (177.2 GFLOPS), but this is only relevant for workloads that specifically use FP16, and the absolute FP32 rate of the GTX 1080 Ti is still more than double.
In summary, the GTX 1080 Ti is the performance champion, winning 9 of 10 tests with an average score that is 19.7% higher. The GTX 1660 SUPER is the efficiency champion and the Vulkan specialist. The choice depends on whether the user needs raw compute power or power efficiency and Vulkan compatibility.