NVIDIA GeForce GTX 1660 SUPER vs NVIDIA P106-090 Comparison

The NVIDIA P106-090 and the NVIDIA GeForce GTX 1660 SUPER represent two very different approaches to the same silicon vendor, separated by over two years of architectural evolution. The P106-090 is a Pascal-based mining card stripped of display outputs, while the GTX 1660 SUPER is a Turing-based consumer graphics card. Benchmark data shows a complete sweep for the GTX 1660 SUPER across every common test, but the margin of victory and the underlying specifications tell a more nuanced story about what each card is designed to do.

Head-to-Head Benchmarks

The head-to-head benchmark data is unambiguous: the NVIDIA GeForce GTX 1660 SUPER wins all three recorded comparisons, with the NVIDIA P106-090 trailing by significant double-digit percentages. The largest gap appears in the Geekbench Vulkan test, where the GTX 1660 SUPER scores 57,102 against the P106-090’s 18,596, a delta of -67.4% for the mining card. This is a massive gulf, indicating that the Turing architecture’s Vulkan performance is in a different league entirely.

In the 3DMark Steel Nomad DX12 test, the GTX 1660 SUPER scores 1,278 while the P106-090 manages only 509, a -60.2% difference. This test stresses modern DirectX 12 workloads, and the result aligns with the raw compute disparity between the two cards. The GTX 1660 SUPER delivers 5.027 TFLOPS of FP32 performance versus the P106-090’s 2.352 TFLOPS, which explains the roughly 2.5x advantage in this synthetic workload.

The Geekbench OpenCL result is slightly closer but still firmly in the GTX 1660 SUPER’s favor: 52,490 versus 21,304, a -59.4% delta. This test is more compute-oriented and less dependent on fixed-function hardware, yet the P106-090 still cannot close the gap. The pattern is consistent across all three benchmarks, with the GTX 1660 SUPER winning by roughly 60-67% in every case. The P106-090’s aggregate benchmark score of 13,470 is actually slightly higher than the GTX 1660 SUPER’s 12,986, but this is misleading because the P106-090 has only three benchmark entries while the GTX 1660 SUPER has ten, including several Passmark tests that drag its average down.

FAQ

Q: Which card has a higher average benchmark score?

A: The NVIDIA P106-090 has an average benchmark score of 13,470, which is 3.7% higher than the GTX 1660 SUPER’s 12,986. However, this is based on only three benchmarks for the P106-090 versus ten for the GTX 1660 SUPER, so the comparison is not apples-to-apples.

Q: How much faster is the GTX 1660 SUPER in Vulkan workloads?

A: In the Geekbench Vulkan test, the GTX 1660 SUPER scores 57,102 versus the P106-090’s 18,596. This represents a 67.4% advantage for the GTX 1660 SUPER, the largest margin of any head-to-head benchmark.

Q: What is the memory configuration difference?

A: The P106-090 uses 3 GB of GDDR5 memory with a 192-bit bus and 192.2 GB/s bandwidth, while the GTX 1660 SUPER uses 6 GB of GDDR6 memory on the same 192-bit bus but with 336.0 GB/s bandwidth. The GTX 1660 SUPER has both double the capacity and 74.8% more bandwidth.

Q: Does either card support ray tracing or DLSS?

A: Neither card has RT cores or tensor cores listed in the specifications. Both cards support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Q: What is the power consumption difference?

A: The P106-090 has a 75 W TDP with a 1x 6-pin power connector and a suggested 250 W PSU, while the GTX 1660 SUPER has a 125 W TDP with a 1x 8-pin connector and a suggested 300 W PSU. The GTX 1660 SUPER draws 66.7% more power.

Q: Which card has a higher pixel fill rate?

A: The GTX 1660 SUPER has a pixel rate of 85.68 GPixel/s, which is 16.6% higher than the P106-090’s 73.49 GPixel/s. Both cards have 48 ROPs, so the difference comes from the higher clock speeds of the Turing card.

Architecture Differences

The two cards are built on fundamentally different architectures. The P106-090 uses the GP106 chip based on Pascal, manufactured on a 16 nm process at TSMC. The GTX 1660 SUPER uses the TU116 chip based on Turing, manufactured on a 12 nm process, also at TSMC. This process shrink allows the GTX 1660 SUPER to pack 6,600 million transistors into a 284 mm² die, compared to the P106-090’s 4,400 million transistors in a 200 mm² die. The transistor density is similar (23.2M / mm² versus 22.0M / mm²), but the Turing chip has 50% more transistors overall.

The compute resources differ substantially. The GTX 1660 SUPER has 1,408 shading units and 88 texture mapping units, versus the P106-090’s 768 shading units and 48 TMUs. Both cards share the same 48 ROPs, but the GTX 1660 SUPER’s higher clock speeds (1,530 MHz base and 1,785 MHz boost versus 1,354 MHz base and 1,531 MHz boost) amplify the advantage. The texture rate tells the story: 157.1 GTexel/s for the GTX 1660 SUPER versus 73.49 GTexel/s for the P106-090, a 113.8% difference. The FP32 compute is similarly lopsided at 5.027 TFLOPS versus 2.352 TFLOPS.

Memory architecture also diverges. The P106-090 uses GDDR5 at 8 Gbps effective, while the GTX 1660 SUPER uses GDDR6 at 14 Gbps effective. Both use a 192-bit bus, but the faster GDDR6 memory yields 336.0 GB/s versus 192.2 GB/s. The FP16 performance is particularly telling: the GTX 1660 SUPER achieves 10.05 TFLOPS with a 2:1 ratio to FP32, while the P106-090 manages only 36.74 GFLOPS with a 1:64 ratio, highlighting the Turing architecture’s focus on mixed-precision workloads.

Specification Differences

The most obvious difference is the display outputs. The P106-090 has no outputs at all, while the GTX 1660 SUPER provides 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. This directly reflects their intended use cases: mining versus gaming. The bus interface also differs, with the P106-090 limited to PCIe 1.0 x1 while the GTX 1660 SUPER uses PCIe 3.0 x16. This is a critical constraint for the P106-090, as the narrow bus severely limits data transfer to the host system.

Physical dimensions differ as well. The P106-090 is 250 mm (9.8 inches) long, while the GTX 1660 SUPER is 229 mm (9 inches) long. The GTX 1660 SUPER also has listed height and width of 111 mm (4.4 inches) and 35 mm (1.4 inches), whereas the P106-090 has no height or width specified. Both are dual-slot cards, but the power connectors differ: the P106-090 uses a single 6-pin, and the GTX 1660 SUPER uses a single 8-pin. The TDP is 75 W for the P106-090 and 125 W for the GTX 1660 SUPER, with suggested PSU ratings of 250 W and 300 W respectively.

The release dates are separated by over two years: the P106-090 launched on 2017-07-30, and the GTX 1660 SUPER launched on 2019-10-28. The GTX 1660 SUPER has a launch MSRP of 229 USD, while the P106-090 has no listed launch MSRP. The series and generation fields also differ: the P106-090 belongs to the Mining GPUs generation with no series, while the GTX 1660 SUPER belongs to the GeForce 16 generation within the GeForce 10-series. The GTX 1660 SUPER lists a predecessor of GeForce 10 and a successor of GeForce 20, while the P106-090 has neither.

The Verdict

The data is unequivocal: the NVIDIA GeForce GTX 1660 SUPER outperforms the NVIDIA P106-090 in every single benchmark they share, with margins ranging from 59.4% to 67.4%. The GTX 1660 SUPER offers more than double the FP32 compute, more than double the memory capacity, and 74.8% more memory bandwidth. It also has proper display outputs, a full PCIe 3.0 x16 interface, and a modern Turing architecture with vastly superior FP16 performance.

However, the P106-090 is not without its own profile. It has a significantly lower TDP of 75 W versus 125 W, making it more power-efficient in absolute terms. Its average benchmark score of 13,470 is actually higher than the GTX 1660 SUPER’s 12,986, though this is an artifact of the test count difference. The P106-090 also occupies the 54th percentile of all GPUs, one point higher than the GTX 1660 SUPER’s 53rd percentile, which reinforces the idea that the P106-090’s limited benchmark suite flatters its standing.

The verdict depends entirely on use case. For any task involving graphics output, gaming, or modern API workloads, the GTX 1660 SUPER is the only choice. For compute-only deployments where power draw is the primary constraint and display output is irrelevant, the P106-090’s lower TDP may be attractive, but the performance gap is so large that it would take multiple P106-090 cards to match a single GTX 1660 SUPER.

Where Each One Wins

The GTX 1660 SUPER wins decisively in every head-to-head benchmark category. In 3DMark Steel Nomad DX12, it is 60.2% faster. In Geekbench OpenCL, it is 59.4% faster. In Geekbench Vulkan, it is 67.4% faster. The GTX 1660 SUPER also wins on memory bandwidth (336.0 GB/s versus 192.2 GB/s), texture rate (157.1 GTexel/s versus 73.49 GTexel/s), and FP32 compute (5.027 TFLOPS versus 2.352 TFLOPS). It has double the VRAM, a faster memory type, and a wider bus interface. For any gaming, rendering, or general-purpose GPU workload, the GTX 1660 SUPER is the clear winner.

The P106-090 wins only on power efficiency and physical footprint considerations. Its 75 W TDP is 40% lower than the GTX 1660 SUPER’s 125 W, and it requires a 250 W PSU versus 300 W. It is longer at 250 mm versus 229 mm, but it lacks the height and width dimensions listed for the GTX 1660 SUPER. The P106-090 also has a slightly higher average benchmark score (13,470 versus 12,986) and a one-point higher percentile ranking (54th versus 53rd), though these metrics are based on incomplete data. For a compute-only environment where power draw is the absolute priority and performance per watt matters more than absolute throughput, the P106-090 has a narrow appeal. The data shows no other scenario where the P106-090 is preferable.