NVIDIA GeForce GTX 1660 Ti Mobile

The NVIDIA GeForce GTX 1660 Ti Mobile is an end-of-life mobile graphics solution built on the Turing architecture, specifically the TU116 chip manufactured on TSMC's 12 nm process. The benchmark data places this GPU at the 50th percentile of all GPUs in the database, indicating a mid-pack standing in overall performance. With 6 GB of GDDR6 memory on a 192-bit bus, the card delivers 288.0 GB/s of bandwidth, and its 1536 shading units produce a peak FP32 throughput of 4.884 TFLOPS. The release date is 2019-04-22, and the production status is end-of-life, with the predecessor being GeForce 10 Mobile and the successor being GeForce 20 Mobile. This analysis examines the measured specifications and their implications for gaming and compute workloads.

Benchmark Performance

The FACT PACK provides no raw benchmark scores, but the percentile field places the GTX 1660 Ti Mobile at exactly the 50th percentile among all GPUs. This is a median position, meaning half of all GPUs in the database perform better and half perform worse. The FP32 compute rate is 4.884 TFLOPS, which is the primary scalar throughput figure. The FP16 rate is 9.769 TFLOPS with a 2:1 ratio, indicating that half-precision operations run at double the rate of single-precision. The pixel rate is 76.32 GPixel/s and the texture rate is 152.6 GTexel/s, derived from the 48 ROPs and 96 TMUs respectively. The base clock is 1455 MHz with a boost of 1590 MHz.

These figures suggest a balanced mid-range performer. Without rival data, the 50th percentile is the only comparative anchor. The data shows a GPU that is neither a high-end contender nor a low-end part. It sits squarely in the middle of the distribution, which historically aligns with a mainstream mobile offering. The 2:1 FP16 ratio is notable because it allows certain compute tasks to accelerate, but the lack of tensor cores (null in the pack) means no dedicated AI acceleration hardware. The boost clock of 1590 MHz is the maximum sustained frequency under load, and the base clock of 1455 MHz represents the guaranteed minimum. In terms of fill rates, 76.32 GPixel/s means the GPU can write that many pixels per second across its 48 ROPs, which is adequate for full HD gaming at moderate settings. The texture rate of 152.6 GTexel/s, driven by 96 TMUs, supports complex texture mapping. Overall, the benchmark performance is defined by these raw throughput numbers, and the 50th percentile confirms a mid-pack position. The FP32 throughput of 4.884 TFLOPS is the headline compute figure, and it directly influences how the GPU handles vertex processing and pixel shading. The memory bandwidth of 288.0 GB/s complements this compute rate, ensuring that the shading units are not starved for data.

Ray Tracing and Feature Set

The FACT PACK explicitly lists `rtCores` as null and `tensorCores` as null. This means the GTX 1660 Ti Mobile has no dedicated ray tracing cores and no tensor cores. Therefore, hardware-accelerated ray tracing is not available on this GPU. The feature set is instead defined by its API support: DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The DirectX 12_1 feature level indicates support for conservative rasterization and other DX12 features, but it does not include the DirectX Raytracing (DXR) tier that requires dedicated RT cores. Vulkan 1.4 support means it can run modern Vulkan titles, but again, ray tracing via Vulkan extensions would rely on compute shaders, not dedicated hardware.

The lack of tensor cores means no DLSS (Deep Learning Super Sampling) or other AI-based upscaling features. The FP16 2:1 ratio (9.769 TFLOPS) is the only acceleration beyond standard FP32, and it can be used for some compute workloads, but it does not substitute for tensor cores. The display outputs are listed as "Portable Device Dependent," meaning the GPU's video outputs are determined by the laptop or mobile device it is integrated into. The bus interface is PCIe 3.0 x16, which provides adequate bandwidth for the 288.0 GB/s memory subsystem. The power connectors are "None," indicating that the mobile GPU draws power from the motherboard or MXM connector rather than external PCIe power cables. The slot width is "MXM Module," confirming it is a mobile form factor. The API support is solid for its era, but the absence of RT and tensor cores clearly delineates this as a rasterization-focused GPU. The DirectX 12_1 support ensures compatibility with modern game engines, while the OpenGL 4.6 and Vulkan 1.4 support provide cross-platform flexibility. The lack of ray tracing is a defining characteristic, positioning this GPU firmly in the traditional rasterization camp.

Memory Subsystem

The GTX 1660 Ti Mobile is equipped with 6 GB of GDDR6 memory. The bus width is 192 bits, and the memory clock is 1500 MHz, which translates to 12 Gbps effective data rate. The total bandwidth is 288.0 GB/s. This is a substantial memory bandwidth figure for a mid-range mobile GPU. For high-resolution gaming, 6 GB of VRAM is a consideration. At quad HD, many modern titles can exceed 6 GB with high texture settings, but at full HD, 6 GB is generally sufficient. The 288.0 GB/s bandwidth ensures that the 1536 shading units have ample data to process.

The 192-bit bus width is a balance between cost and performance. The effective 12 Gbps memory speed is standard for GDDR6 of that generation. The memory subsystem is a strong point of this GPU, as the bandwidth is high relative to its compute throughput (4.884 TFLOPS). This ratio suggests that the GPU is not memory-bandwidth-bound in most workloads. The 6 GB capacity is a hard limit, however. For ultra HD gaming, 6 GB is often insufficient for ultra textures, leading to stuttering or texture pop-in. For quad HD, it is borderline. The data shows a memory subsystem that is well-matched to the GPU's compute capabilities, but capacity may be a limiting factor in the long term. The 288.0 GB/s bandwidth is exactly specified, and the 192-bit bus width is the conduit. The memory type GDDR6 is faster than the older GDDR5, contributing to the overall performance. The 1500 MHz memory clock, combined with the 192-bit bus, yields a bandwidth figure that is competitive within its class, though the pack provides no direct rival comparisons.

How It Compares

The FACT PACK's `nearestRivals` array is empty, so no direct rival scores or delta percentages are provided. However, the percentile field provides a comparative anchor. The GPU sits at the 50th percentile of all GPUs in the database. This means it outperforms roughly half of all GPUs and underperforms the other half. Its predecessor is the GeForce 10 Mobile series, and its successor is the GeForce 20 Mobile series. The GeForce 20 Mobile series introduced RT and tensor cores, which this GPU lacks.

The GTX 1660 Ti Mobile is a Turing-architecture part, but it is a cut-down Turing without the dedicated AI and ray tracing hardware. Compared to its predecessor, the GeForce 10 Mobile series, the 1660 Ti Mobile offers a newer architecture (Turing) which typically improves shader efficiency, though the pack does not specify the 10-series architecture. The successor, GeForce 20 Mobile, adds RT cores and tensor cores, which are null here. Therefore, the 1660 Ti Mobile occupies a specific niche: it has the Turing architecture's improved shader efficiency but forgoes the headline features of the 20-series. The 50th percentile ranking places it in the mainstream tier. Without rival names, the analysis must rely on this percentile and the generation context. The data indicates that this is a mid-range mobile GPU that is end-of-life, meaning it has been discontinued. The release date is 2019-04-22, which is a specific point in time. Compared to the 20-series, it likely trades ray tracing for higher rasterization performance per watt, but the pack does not provide wattage. The empty rival list means no exact deltas can be cited, so the percentile and generational context serve as the only comparative metrics.

Who Should Consider It

Based on the 50th percentile ranking and the 6 GB memory capacity, this GPU is suitable for full HD gaming. The 4.884 TFLOPS FP32 performance is adequate for full HD at high settings in most titles of its era. The 288.0 GB/s bandwidth supports high texture quality at full HD. For quad HD, the GPU can handle medium to high settings, but the 6 GB VRAM may become a constraint with modern textures. The 76.32 GPixel/s pixel rate is enough for full HD at high refresh rates, but not for ultra HD. The lack of RT and tensor cores means users should not expect ray tracing or DLSS. Those features are absent.

Therefore, this GPU is best for gamers who prioritize raw rasterization performance at full HD and do not need AI upscaling or ray tracing. The 50th percentile suggests it is a mainstream part, not a high-end one. Users targeting ultra HD should look elsewhere, as 6 GB VRAM and 4.884 TFLOPS are insufficient for ultra HD ultra settings. For quad HD, it is a viable option for esports titles and older games. The FP16 2:1 ratio (9.769 TFLOPS) can benefit compute workloads that use half precision, but this is a niche advantage. The end-of-life status means it is no longer in production, so it is only available on the used market. The MXM Module slot width indicates it is a replaceable mobile module, but the power connectors are "None," meaning it draws from the system. The 50th percentile ranking also suggests that users upgrading from a lower-tier GPU will notice a significant improvement, while those coming from a higher-tier part will see a step down.

Architecture and Design

The GTX 1660 Ti Mobile is built on the Turing architecture, specifically the TU116 chip. The manufacturing process is TSMC's 12 nm node. The die size is 284 mm², and the transistor count is 6,600 million. The transistor density is 23.2 million per square millimeter. This density is a key indicator of the chip's complexity. The TU116 chip is a Turing variant that omits RT and tensor cores, as evidenced by the null values in the FACT PACK.

The core configuration includes 1536 shading units, 96 TMUs, and 48 ROPs. The shading units are the primary processing elements for FP32 and FP16 operations. The TMUs handle texture filtering, and the ROPs handle pixel output. The base clock is 1455 MHz, and the boost clock is 1590 MHz. The memory is 6 GB of GDDR6 on a 192-bit bus, with a bandwidth of 288.0 GB/s. The chip is manufactured by TSMC, and the foundry is specified. The 12 nm process is a mature node. The transistor count of 6,600 million is substantial, and the die size of 284 mm² is relatively large for a mid-range chip.

The power delivery is through the MXM module, with no external power connectors. The bus interface is PCIe 3.0 x16, which is a standard interface. The production status is end-of-life, and the release date is 2019-04-22. The predecessor is GeForce 10 Mobile, and the successor is GeForce 20 Mobile. The architecture is Turing, but it is a cut-down implementation without the specialized cores. The FP32 rate of 4.884 TFLOPS and FP16 rate of 9.769 TFLOPS (2:1) are derived from the 1536 shading units and their clock speeds. The 284 mm² die size and 6,600 million transistors indicate a relatively dense chip, which is consistent with the 12 nm process. The transistor density of 23.2 million per square millimeter is a direct measure of the manufacturing efficiency.

FAQ

Q: What is the memory bandwidth of the GTX 1660 Ti Mobile?

A: The memory bandwidth is 288.0 GB/s, achieved with 6 GB of GDDR6 on a 192-bit bus at 12 Gbps effective.

Q: Does this GPU support hardware ray tracing?

A: No. The FACT PACK lists `rtCores` as null and `tensorCores` as null, meaning there are no dedicated ray tracing or tensor cores.

Q: What is the FP32 compute performance?

A: The FP32 performance is 4.884 TFLOPS, with an FP16 rate of 9.769 TFLOPS (2:1 ratio).

Q: What is the process node and die size?

A: The process node is 12 nm from TSMC, with a die size of 284 mm² and 6,600 million transistors.

Q: What is the release date and production status?

A: The release date is 2019-04-22, and the production status is end-of-life.

Q: What API versions are supported?

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

Q: What is the percentile ranking of this GPU?

A: It is at the 50th percentile of all GPUs in the database, indicating a mid-pack performance position.