NVIDIA GeForce GTX 1660 vs NVIDIA P106-090 Comparison

Head-to-Head Benchmarks

The recorded data presents a decisive picture in every shared test: the NVIDIA GeForce GTX 1660 wins all three head-to-head comparisons. In 3DMark Steel Nomad DX12, the GTX 1660 scores 1065 against the P106-090's 509, a delta of -52.2% from the P106-090's perspective. That means the GTX 1660 more than doubles the P106-090's performance in this DirectX 12 workload. In Geekbench OpenCL, the GTX 1660 reaches 47850 while the P106-090 manages 21304, a -55.5% delta. The gap widens further in Geekbench Vulkan: the GTX 1660 posts 50137 versus 18596, a -62.9% delta. The P106-090 does not secure a single win in the head-to-head set.

The margin is not marginal; it is systematic. In Vulkan, the GTX 1660 runs at roughly 2.7 times the P106-090's score. In OpenCL, the ratio is about 2.25 to 1. Even in the DX12 test, the GTX 1660 is ahead by a factor of 2.09. The P106-090's best relative showing is in 3DMark Steel Nomad, where it trails by just over half, but that is still a dominant win for the GTX 1660. The average benchmark score tells a slightly different story, however. The P106-090's average is 13470, while the GTX 1660's average is 11680. That puts the P106-090 ahead by roughly 15% in aggregate across all recorded benchmarks, despite losing every shared test. This is a curious inversion, and it suggests that the P106-090's recorded scores may be drawn from a different mix of workloads, or that the GTX 1660's extra benchmark entries (including several Passmark tests) drag its average down.

The percentile data adds context. The P106-090 sits at the 54th percentile among all GPUs, while the GTX 1660 sits at the 51st percentile. The P106-090's nearest rivals include the NVIDIA GeForce GTX 570 at -0.3%, the AMD Radeon Pro 555 at +0.5%, the AMD Radeon HD 7770M at -0.5%, and the AMD Radeon RX 9070 XT at -0.5%. The GTX 1660's nearest rivals are the AMD Radeon RX 7800 XT at +0.5%, the AMD Radeon Pro 5500M at +1.3%, the AMD Radeon RX 6500 XT at -1.4%, and the NVIDIA Tesla K20c at +1.8%. So in the broader database, the P106-090 is clustered with older and lower-tier cards, while the GTX 1660 is clustered with a mix of mid-range and professional parts. The GTX 1660's head-to-head dominance is not reflected in its percentile rank, which is lower than the P106-090's; the P106-090 benefits from a higher average score despite losing every direct comparison.

Architecture Differences

The two cards come from different architectural generations. The P106-090 uses the GP106 chip built on the Pascal architecture, fabricated on a 16 nm process at TSMC. The GTX 1660 uses the TU116 chip built on the Turing architecture, also at TSMC but on a 12 nm process. The transistor counts differ substantially: the P106-090 packs 4,400 million transistors on a 200 mm² die, while the GTX 1660 packs 6,600 million transistors on a 284 mm² die. Transistor density is similar, 22.0M per mm² for the P106-090 and 23.2M per mm² for the GTX 1660, so the GTX 1660's larger chip is mostly a matter of scale, not density.

The compute resources are heavily skewed toward the GTX 1660. The GTX 1660 has 1408 shading units, 88 texture mapping units, and 48 ROPs. The P106-090 has 768 shading units, 48 TMUs, and 48 ROPs. The ROP count is identical, but the GTX 1660 has nearly double the shading units and nearly double the TMUs. This explains the large gap in texture rate: the GTX 1660 delivers 157.1 GTexel/s versus 73.49 GTexel/s for the P106-090. Pixel rate also favors the GTX 1660, at 85.68 GPixel/s versus 73.49 GPixel/s, though the gap is smaller. FP32 throughput is 5.027 TFLOPS for the GTX 1660 and 2.352 TFLOPS for the P106-090, a difference of roughly 2.1 times.

Memory is another major differentiator. The P106-090 has 3 GB of GDDR5 on a 192-bit bus, yielding 192.2 GB/s of bandwidth. The GTX 1660 has 6 GB of GDDR5 on the same 192-bit bus, yielding 192.1 GB/s. The bandwidth is effectively identical, but the GTX 1660 has twice the capacity. Clock speeds also favor the GTX 1660: base clock 1530 MHz versus 1354 MHz, boost clock 1785 MHz versus 1531 MHz. Memory clock is nearly the same, 2001 MHz versus 2002 MHz, both at 8 Gbps effective.

The GTX 1660 also differs in features that matter for real-world use. It has display outputs: 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. The P106-090 has no display outputs at all, a clear sign of its mining-oriented design. The bus interface is also different: the GTX 1660 uses PCIe 3.0 x16, while the P106-090 uses PCIe 1.0 x1, an extremely narrow link that would bottleneck any data transfer. The P106-090's power draw is lower at 75 W versus 120 W, and its power connector is a single 6-pin versus the GTX 1660's single 8-pin. The suggested PSU is 250 W for the P106-090 and 300 W for the GTX 1660.

Where Each One Wins

The GTX 1660 wins every shared benchmark, so its territory is clear: any workload represented in the head-to-head set. The 3DMark Steel Nomad DX12 test, the Geekbench OpenCL test, and the Geekbench Vulkan test all go to the GTX 1660, and the margins are large. If the question is which card handles compute or graphics workloads better, the data points squarely at the GTX 1660. Its higher shading unit count, texture rate, and FP32 throughput give it a structural advantage that shows up in every recorded test.

The P106-090 has one virtue in the recorded data: its average benchmark score of 13470 is higher than the GTX 1660's 11680. That average is computed across all benchmarks in the database, not just the shared ones, and it suggests that the P106-090 may have been tested in workloads that favor its profile, or that its benchmark pool includes scores from different driver versions or test conditions. The P106-090 also has a lower TDP of 75 W, making it a lighter load on a power supply, and its dual-slot cooler with a single 6-pin connector may fit into systems with more limited power delivery. But in terms of recorded performance, the P106-090 has no wins.

The GTX 1660's Passmark results, which the P106-090 lacks, show a mixed profile: Passmark G3D score of 11646, Passmark GPU Compute score of 4963, Passmark G2D score of 776, and DirectX-specific scores of 177 (DX9), 79 (DX11), 61 (DX10), and 49 (DX12). These numbers are not directly comparable to the P106-090 because the P106-090 has no Passmark entries, but they illustrate the GTX 1660's range.

The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 1660 is the faster card. It wins all three head-to-head benchmarks by margins of -52.2%, -55.5%, and -62.9% (from the P106-090's perspective). It has more shading units, more TMUs, double the memory capacity, higher clock speeds, and a newer architecture on a smaller process node. The GTX 1660 also has display outputs, which the P106-090 lacks entirely, making the GTX 1660 usable as a standard graphics card while the P106-090 cannot drive a monitor. The GTX 1660 carries a launch MSRP of 219 USD, though pricing is not a factor in this analysis.

The P106-090's only recorded advantage is its higher average benchmark score of 13470 versus 11680, and its lower TDP of 75 W versus 120 W. For a system where power draw is critical and display output is unnecessary, the P106-090 might have a niche. For any application that requires actual graphics output or maximum compute throughput, the GTX 1660 is the clear choice. The P106-090's PCIe 1.0 x1 interface is a severe limitation, and its lack of display outputs confirms its mining-specific design. The GTX 1660 is the better card for nearly any use case that involves rendering, compute, or gaming.

FAQ

Q: Which GPU wins the 3DMark Steel Nomad DX12 benchmark?

A: The NVIDIA GeForce GTX 1660 wins with a score of 1065, while the NVIDIA P106-090 scores 509, a delta of -52.2% from the P106-090's perspective.

Q: How much memory does each card have?

A: The P106-090 has 3 GB of GDDR5, while the GTX 1660 has 6 GB of GDDR5. Both use a 192-bit bus, and bandwidth is nearly identical at 192.2 GB/s for the P106-090 and 192.1 GB/s for the GTX 1660.

Q: Does the P106-090 support display output?

A: No, the P106-090 has no display outputs. The GTX 1660 has 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a.

Q: What are the architecture and process node differences?

A: The P106-090 uses the Pascal architecture on TSMC's 16 nm process, while the GTX 1660 uses the Turing architecture on TSMC's 12 nm process. The GTX 1660 has 6,600 million transistors on a 284 mm² die, versus 4,400 million on 200 mm² for the P106-090.

Q: Which card has a higher average benchmark score?

A: The P106-090 has a higher average benchmark score of 13470, compared to 11680 for the GTX 1660, despite losing every head-to-head test.

Q: What is the power draw of each card?

A: The P106-090 has a TDP of 75 W and uses a 1x 6-pin power connector with a suggested PSU of 250 W. The GTX 1660 has a TDP of 120 W, uses a 1x 8-pin connector, and suggests a 300 W PSU.

Specification Differences

| Specification | NVIDIA P106-090 | NVIDIA GeForce GTX 1660 |

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

| Architecture | Pascal | Turing |

| Process Node | 16 nm | 12 nm |

| Transistors | 4,400 million | 6,600 million |

| Die Size | 200 mm² | 284 mm² |

| Transistor Density | 22.0M / mm² | 23.2M / mm² |

| Base Clock | 1354 MHz | 1530 MHz |

| Boost Clock | 1531 MHz | 1785 MHz |

| Memory Size | 3 GB | 6 GB |

| Memory Type | GDDR5 | GDDR5 |

| Memory Bus Width | 192 bit | 192 bit |

| Memory Clock | 2002 MHz (8 Gbps effective) | 2001 MHz (8 Gbps effective) |

| Memory Bandwidth | 192.2 GB/s | 192.1 GB/s |

| Shading Units | 768 | 1408 |

| TMUs | 48 | 88 |

| ROPs | 48 | 48 |

| Pixel Rate | 73.49 GPixel/s | 85.68 GPixel/s |

| Texture Rate | 73.49 GTexel/s | 157.1 GTexel/s |

| FP32 | 2.352 TFLOPS | 5.027 TFLOPS |

| FP16 | 36.74 GFLOPS (1:64) | 10.05 TFLOPS (2:1) |

| TDP | 75 W | 120 W |

| Power Connectors | 1x 6-pin | 1x 8-pin |

| Suggested PSU | 250 W | 300 W |

| Bus Interface | PCIe 1.0 x1 | PCIe 3.0 x16 |

| Display Outputs | No outputs | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a |

| Slot Width | Dual-slot | Dual-slot |

| Length | 250 mm (9.8 inches) | 229 mm (9 inches) |

| Height | Not specified | 111 mm (4.4 inches) |

| Width | Not specified | 35 mm (1.4 inches) |

| DirectX Support | 12 (12_1) | 12 (12_1) |

| OpenGL Support | 4.6 | 4.6 |

| Vulkan Support | 1.4 | 1.4 |

| Release Date | 2017-07-30 | 2019-03-13 |

| Production Status | End-of-life | End-of-life |

| Launch MSRP | Not specified | 219 USD |

| Predecessor | Not specified | GeForce 10 |

| Successor | Not specified | GeForce 20 |