NVIDIA GeForce RTX 2060 SUPER

The NVIDIA GeForce RTX 2060 SUPER, built on the TU106 chip with Turing architecture and a 12 nm process, occupies a narrow performance band where a handful of older and newer cards converge. Its average benchmark score of 19774 places it in the 63rd percentile of all GPUs, indicating a solidly mid-to-upper-tier position. The data shows a tightly clustered set of rivals, with the RTX 2060 SUPER leading the nearest competitor by a margin of just 0.3%, which makes its placement highly sensitive to specific workload characteristics.

How It Compares

The closest rival is the NVIDIA GeForce GTX TITAN, which holds an average score of 19706. The RTX 2060 SUPER is only 0.3% ahead, a margin that is effectively negligible in real-world terms. This proximity is notable because the GTX TITAN is a much older, high-end card from a previous generation, while the RTX 2060 SUPER benefits from newer architecture features like dedicated ray tracing cores. The data suggests that in raw compute throughput, the two are nearly identical, but the feature set diverges significantly.

Next is the NVIDIA GeForce RTX 2070, with an average score of 19680. The RTX 2060 SUPER leads by 0.5%. This is a counterintuitive result, as the RTX 2070 is positioned higher in the product stack. The benchmark data indicates that the performance gap between these two is within the margin of error, meaning the RTX 2060 SUPER offers nearly the same level of performance as its more expensive sibling in the aggregate metrics tracked here.

The NVIDIA Quadro K5200 trails with an average score of 19623, placing it 0.8% behind the RTX 2060 SUPER. The Quadro K5200 is a professional workstation card, and its presence in this comparison highlights that the RTX 2060 SUPER’s compute capabilities are competitive with specialized prosumer hardware, at least in the general-purpose benchmarks represented in this database.

Finally, the NVIDIA Tesla K40m, a datacenter-oriented accelerator, scores 19519, which is 1.3% lower than the RTX 2060 SUPER. This is the largest delta among the listed rivals. The Tesla K40m’s design prioritizes compute density and reliability over consumer gaming features, yet the RTX 2060 SUPER still manages to edge it out in aggregate benchmark scores, underscoring the efficiency gains of the newer Turing architecture.

Benchmark Performance

The performance profile of the RTX 2060 SUPER is best understood through its individual benchmark results. In 3DMark Steel Nomad (DX12), the card scores 2011, which is a demanding test for modern GPUs. While no direct rival score is provided for this specific test, the aggregate data suggests this is a respectable result for a card of this class. The Geekbench OpenCL score of 87670 indicates strong general-purpose compute performance, while the Vulkan score of 83503 shows slightly lower but still robust API-specific throughput.

In the PassMark suite, the results are more granular. The DirectX 11 score of 130 is the highest among the DX tests, suggesting that the card performs optimally in legacy DirectX 11 titles. The DirectX 10 score of 111 is lower, and the DirectX 12 score of 61 drops significantly. This pattern is unusual, as modern cards typically excel in newer APIs. The data implies that the RTX 2060 SUPER’s driver optimization may be more mature for older APIs, or that the PassMark DX12 test does not fully utilize the card’s Turing-specific features. The DirectX 9 score of 218 is the highest overall, indicating excellent compatibility with very old games.

The PassMark G3D score of 16462 is a solid marker for overall 3D rendering capability, while the PassMark G2D score of 854 reflects 2D desktop performance. The GPU Compute score of 6721 is notably lower than the G3D score, suggesting that the card’s compute throughput, while competitive in aggregate, is not its primary strength relative to its rasterization performance. Comparing the RTX 2060 SUPER to the RTX 2070, the 0.5% lead in average score means that in most gaming scenarios, the difference is imperceptible. Against the GTX TITAN, the 0.3% lead is similarly marginal, but the RTX 2060 SUPER brings hardware-accelerated ray tracing, which the TITAN lacks entirely.

Ray Tracing and Feature Set

The RTX 2060 SUPER is equipped with 34 dedicated RT cores, which are specifically designed to accelerate ray-traced workloads. This is a fundamental architectural advantage over its nearest rivals, none of which feature RT cores. The presence of 272 tensor cores further enhances the card’s capability for AI-driven tasks, such as deep learning super sampling. The data shows that the card supports DirectX 12 Ultimate (12_2), which includes ray tracing and variable rate shading as core features. Additionally, it supports OpenGL 4.6 and Vulkan 1.4, making it compatible with a wide range of modern and legacy applications.

The memory subsystem consists of 8 GB of GDDR6 on a 256-bit bus, yielding a bandwidth of 448.0 GB/s. This is a substantial amount of memory bandwidth, which is critical for ray tracing operations that require large amounts of data to be fetched quickly. The card’s pixel rate is 105.6 GPixel/s and its texture rate is 224.4 GTexel/s, both of which are adequate for its performance class. The FP32 compute throughput is 7.181 TFLOPS, with FP16 performance doubling to 14.36 TFLOPS via a 2:1 ratio. This FP16 capability is particularly useful for AI inference and certain compute workloads that can leverage reduced precision.

The display outputs include 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C, providing flexibility for multi-monitor setups and VR headsets. The card is a dual-slot design, measures 229 mm in length, 113 mm in height, and 35 mm in width, and requires a single 8-pin power connector. The total board power is rated at 175 W, with a suggested power supply of 450 W. The bus interface is PCIe 3.0 x16, which is fully compatible with modern motherboards.

FAQ

Q: How does the RTX 2060 SUPER compare to the RTX 2070 in terms of performance?

A: The RTX 2060 SUPER has an average benchmark score of 19774, which is 0.5% higher than the RTX 2070’s average score of 19680. This indicates nearly identical performance in the aggregate benchmarks.

Q: What is the advantage of the RTX 2060 SUPER over the GTX TITAN?

A: While the RTX 2060 SUPER is only 0.3% faster in average score (19774 vs 19706), it includes 34 RT cores and 272 tensor cores, enabling hardware-accelerated ray tracing and AI features that the GTX TITAN does not support.

Q: Does the RTX 2060 SUPER support modern graphics APIs?

A: Yes, it supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. DirectX 12 Ultimate includes features like ray tracing and variable rate shading.

Q: What is the memory configuration of the RTX 2060 SUPER?

A: The card features 8 GB of GDDR6 memory on a 256-bit bus, providing a bandwidth of 448.0 GB/s. The memory clock is 1750 MHz, which translates to 14 Gbps effective.

Q: How does the card perform in legacy DirectX benchmarks?

A: In PassMark tests, the DirectX 9 score is 218, DirectX 10 is 111, DirectX 11 is 130, and DirectX 12 is 61. The card performs best in DirectX 9 and 11, while its DirectX 12 score is lower.

Q: What is the power requirement for this card?

A: The total board power is 175 W, and a 450 W power supply is suggested. It requires a single 8-pin power connector.

Who Should Consider It

The RTX 2060 SUPER is suited for gamers targeting 1440p resolution with high settings, given its 8 GB memory capacity and competitive rasterization scores. The data shows it can handle modern DX12 titles, though its PassMark DX12 score of 61 is the lowest of the DX tests, suggesting that performance in the latest APIs may be less stellar than in older ones. For 1080p gaming, the card is overqualified, delivering high frame rates in most titles. The 0.5% lead over the RTX 2070 means that users who find the RTX 2060 SUPER at a lower price point lose nothing in raw performance, while gaining access to RT cores for ray tracing in supported games.

Users interested in ray tracing should note that the 34 RT cores provide entry-level hardware acceleration, but the card’s overall compute throughput of 7.181 TFLOPS means that enabling ray tracing will require lower resolutions or settings to maintain playable frame rates. The tensor cores also make this card viable for light AI workloads, such as image processing or machine learning inference, where FP16 performance of 14.36 TFLOPS can be leveraged. However, for professional compute tasks, the Quadro K5200 and Tesla K40m, despite their lower average scores, are built for sustained workloads and may be more reliable in those environments.

The card’s 175 W power draw and dual-slot design make it a reasonable fit for most mid-tower cases, though the 229 mm length requires checking case clearance. With a 63rd percentile ranking, it is a solid mid-range option for users who want a balance of performance and modern features without stepping up to higher-tier cards. Given its end-of-life production status, it is a legacy purchase, but the benchmark data shows it remains competitive against its immediate successor’s predecessor lineup. For users on a GTX 10-series card, the RTX 2060 SUPER offers a significant feature uplift with ray tracing support and higher memory bandwidth, making it a worthwhile upgrade for those seeking to experience Turing’s capabilities.