NVIDIA GeForce GTX 1660

The NVIDIA GeForce GTX 1660 is a GeForce 16-generation GPU built on the Turing architecture using TSMC’s 12 nm process. The TU116 chip contains 6,600 million transistors on a 284 mm² die, for a transistor density of 23.2 million transistors per mm². Its average benchmark score is 13,265, which places the card at the 53rd percentile among all GPUs. The card was released on 2019-03-13, is now listed as end-of-life, and its successor is the GeForce 20 series. Its launch MSRP was 219 USD.

Memory Subsystem

The GTX 1660 uses 6 GB of GDDR5 memory on a 192-bit bus, with a memory clock of 2001 MHz and an effective data rate of 8 Gbps. The resulting memory bandwidth is 192.1 GB/s. This is a moderate memory configuration, and the 192-bit bus width is directly tied to that bandwidth figure. The processing context around this memory subsystem includes 1,408 shading units, 88 texture mapping units, and 48 ROPs. The pixel rate is 85.68 GPixel/s, and the texture rate is 157.1 GTexel/s.

For high resolutions, the 6 GB frame buffer is the first constraint. A 6 GB GDDR5 configuration is not a large capacity, and high-resolution rendering with sizeable textures can pressure it. The 192.1 GB/s bandwidth is likewise not in the high-bandwidth class; it supports the 85.68 GPixel/s pixel rate but can be stressed by high-resolution fill workloads. Benchmark results show a Geekbench OpenCL score of 57,988 and a Vulkan score of 55,845, indicating general compute and API throughput. The PassMark G2D score is 776, while the PassMark G3D score is 11,646, meaning the card is oriented far more toward 3D rendering than 2D desktop acceleration. In short, the memory subsystem is capable for lower-resolution rendering, but high resolutions will expose both the 6 GB capacity limit and the 192.1 GB/s bandwidth ceiling.

Ray Tracing and Feature Set

The TU116 chip has no RT cores and no tensor cores: both fields are null in the hardware specification. This means there is no dedicated ray tracing or tensor processing hardware on the GTX 1660. The rendering engine is built around standard shader hardware, with 1,408 shading units, 88 TMUs, and 48 ROPs. FP32 compute is listed at 5.027 TFLOPS, while FP16 compute is 10.05 TFLOPS with a 2:1 ratio. The card supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, so modern graphics APIs are available even though ray tracing acceleration is not.

The feature set also includes PCIe 3.0 x16 as the bus interface. Display outputs are one DVI, one HDMI 2.0, and one DisplayPort 1.4a. Power delivery is via a single 8-pin connector, with a TDP of 120 W and a suggested PSU of 300 W. Physically, the card is dual-slot, measuring 229 mm by 111 mm by 35 mm. In the 3DMark Steel Nomad DX12 test, the GTX 1660 scores 1,065, a modest result that is consistent with the absence of dedicated RT/Tensor hardware and with the lower DirectX 12 score observed in PassMark testing.

Who Should Consider It

The GTX 1660’s benchmark profile is strongly era-dependent. PassMark results show a DirectX 9 score of 177, a DirectX 10 score of 61, a DirectX 11 score of 79, and a DirectX 12 score of 49. The highest relative score is in DirectX 9, and the DirectX 11 score is notably higher than the DirectX 12 score. As a result, users running DirectX 9 or DirectX 11 workloads will see a more favorable performance profile than users targeting DirectX 12-based applications. The Vulkan score of 55,845 and the OpenCL score of 57,988 in Geekbench also indicate that Vulkan and compute workloads are better served than the DirectX 12 path.

The 53rd percentile overall placement means the card sits just above the middle of the GPU field. Its average benchmark score is 13,265, and its nearest rivals are all within a narrow band: the GTX 480 sits at -0.3%, the Radeon Pro 555 at -1%, the Radeon Pro 455 at -1.1%, and the MX350 at +1.3%. Users whose software library leans on older DirectX APIs or Vulkan should consider this card as a viable option. Users expecting strong DirectX 12 performance should weigh the PassMark DirectX 12 score of 49 and the 3DMark Steel Nomad DX12 score of 1,065. The 6 GB GDDR5 frame buffer and 192.1 GB/s bandwidth further point toward lower-resolution or less VRAM-intensive scenarios. Production status is end-of-life, so this is an existing-harware consideration rather than a current-generation purchase.

How It Compares

Against the NVIDIA GeForce GTX 480, the GTX 1660 records an average score of 13,265 versus 13,300. The delta is -0.3%, which means the GTX 1660 trails the GTX 480 by only 0.3 percent. In practice, this makes the two effectively equivalent in average benchmark terms.

Against the AMD Radeon Pro 555, the GTX 1660 averages 13,265 versus 13,395. The delta is -1%, so the GTX 1660 sits one percent behind the Radeon Pro 555. This is a small gap, but it is the same direction as the Pro 555 comparison.

Against the AMD Radeon Pro 455, the GTX 1660 averages 13,265 versus 13,416. The delta is -1.1%, the largest deficit among the four nearest rivals listed. Even so, the 1.1 percent margin is narrow and leaves the two cards in the same performance class.

Against the NVIDIA GeForce MX350, the GTX 1660 averages 13,265 versus 13,095. The delta is +1.3%, making the MX350 the only rival in the nearest set that the GTX 1660 leads. The margin is still small, but it is a positive delta and puts the GTX 1660 ahead of the MX350 in average score.

FAQ

Q: What is the memory configuration of the GTX 1660?

A: The GTX 1660 has 6 GB of GDDR5 memory on a 192-bit bus, with a memory clock of 2001 MHz, an 8 Gbps effective data rate, and 192.1 GB/s of bandwidth.

Q: Does the GTX 1660 include dedicated ray tracing or tensor cores?

A: No. The specification lists rtCores as null and tensorCores as null, so there is no dedicated ray tracing or tensor processing hardware on the chip.

Q: Which graphics APIs does the GTX 1660 support?

A: It supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Q: How does the GTX 1660 perform in the PassMark DirectX tests?

A: It scores 177 in DirectX 9, 61 in DirectX 10, 79 in DirectX 11, and 49 in DirectX 12. The DirectX 9 score is the highest of the four, and the DirectX 12 score is the lowest.

Q: What is the GTX 1660’s average benchmark score and percentile?

A: Its average benchmark score is 13,265, which places it at the 53rd percentile among all GPUs.

Q: What are the power requirements of the GTX 1660?

A: The card has a TDP of 120 W, requires one 8-pin power connector, and lists a suggested PSU of 300 W. It is a dual-slot card measuring 229 mm by 111 mm by 35 mm.