NVIDIA GeForce RTX 2060 SUPER vs NVIDIA GeForce RTX 2070 Comparison

The NVIDIA GeForce RTX 2070 and RTX 2060 SUPER are close siblings, but the benchmark data clearly separates them. The RTX 2060 SUPER wins 7 out of 10 head-to-head tests, including a decisive 22% victory in the demanding 3DMark Steel Nomad DX12 test, while the RTX 2070 counters with strong wins in compute-oriented workloads like Geekbench OpenCL and Vulkan. With an average benchmark score of 18,093 for the 2060 SUPER versus 18,789 for the 2070, the performance gap is narrower than the wins count suggests, making the choice dependent on workload rather than a clear overall champion.

Head-to-Head Benchmarks

The largest single performance swing favors the RTX 2060 SUPER. In the 3DMark Steel Nomad DX12 test, the 2060 SUPER scores 2,011 against the 2070’s 1,569 — a 22% difference. This is the most significant margin in any test and indicates that in modern, DX12-based gaming workloads, the 2060 SUPER has a substantial advantage.

The RTX 2070, however, dominates in compute benchmarks. It leads by 3.9% in Geekbench OpenCL (79,966 vs 76,957) and by 6.6% in Geekbench Vulkan (82,521 vs 77,402). These are the two largest wins for the 2070, showing that its higher shading unit count translates directly into better raw compute throughput.

The Passmark suite tells a more nuanced story. The 2070 wins Passmark DirectX 10 with a tie at 111 points, but loses the remaining Passmark tests. The 2060 SUPER edges ahead in DirectX 11 (130 vs 129, a 0.8% margin), DirectX 12 (61 vs 60, a 1.6% margin), and DirectX 9 (218 vs 211, a 3.2% margin). The 2060 SUPER also wins Passmark G2D (854 vs 819, a 4.1% margin) and Passmark G3D (16,462 vs 16,094, a 2.2% margin).

The compute-heavy Passmark GPU Compute test also favors the 2060 SUPER, which scores 6,721 versus the 2070’s 6,411 — a 4.6% lead. This is notable because the 2070 has more shaders, yet the 2060 SUPER still outperforms it in this specific compute test, suggesting clock speed plays a larger role than core count in some workloads.

In terms of overall benchmark averages, the 2070 holds a 3.8% lead (18,789 vs 18,093). The 2070’s percentile rank among all GPUs is 63, while the 2060 SUPER sits at 62. The nearest rivals for the 2070 include the NVIDIA Tesla K80 (18,866, a 0.4% edge for the K80) and the AMD Radeon Pro 5700 XT (18,685, a 0.6% edge for the 2070). For the 2060 SUPER, the nearest rival is the NVIDIA GeForce RTX 3060 Mobile (18,159, a 0.4% edge for the 2060 SUPER).

The Verdict

The data points to the RTX 2060 SUPER as the better choice for modern gaming. Its 22% lead in 3DMark Steel Nomad DX12 is a decisive indicator of DX12 gaming performance, and it also wins Passmark DirectX 9, 11, and 12 tests. If the primary use case is playing current or upcoming games that leverage DX12, the 2060 SUPER is clearly superior.

The RTX 2070 is the pick for compute-focused tasks. Its wins in Geekbench OpenCL and Vulkan, combined with a higher overall average score, make it the better option for GPU-accelerated compute, rendering, or any workload that leverages OpenCL or Vulkan APIs. The 2070’s higher shading unit count (2,304 vs 2,176) and higher FP32 throughput (7.465 TFLOPS vs 7.181 TFLOPS) support this compute advantage.

For users who value a balanced mix of gaming and compute, the 2060 SUPER offers the better gaming performance at a lower launch MSRP of 399 USD compared to the 2070’s 499 USD. However, the 2070’s higher average benchmark score and better compute results mean it is not obsolete — it simply targets a different use case. The 2060 SUPER is the more versatile card for gaming-first systems, while the 2070 serves users who need strong compute throughput alongside decent gaming.

FAQ

Q: Which card performs better in DirectX 12 gaming?

A: The RTX 2060 SUPER wins decisively. In 3DMark Steel Nomad DX12, it scores 2,011 versus the RTX 2070’s 1,569, a 22% advantage. The 2060 SUPER also wins Passmark DirectX 12 (61 vs 60).

Q: Is the RTX 2070 better for compute workloads?

A: Yes. The RTX 2070 wins Geekbench OpenCL by 3.9% (79,966 vs 76,957) and Geekbench Vulkan by 6.6% (82,521 vs 77,402). However, the 2060 SUPER wins Passmark GPU Compute by 4.6% (6,721 vs 6,411), so the compute advantage is workload-specific.

Q: How do their average benchmark scores compare?

A: The RTX 2070 has a higher average benchmark score of 18,789 versus the RTX 2060 SUPER’s 18,093, a 3.8% difference. The 2070 also ranks at the 63rd percentile of all GPUs, while the 2060 SUPER ranks at the 62nd.

Q: Which card has more shading units?

A: The RTX 2070 has 2,304 shading units, while the RTX 2060 SUPER has 2,176. The 2070 also has 144 texture mapping units versus the 2060 SUPER’s 136, and 36 ray tracing cores versus 34.

Q: Are the memory specifications the same?

A: Yes. Both cards feature 8 GB of GDDR6 memory on a 256-bit bus with 448.0 GB/s of bandwidth and the same 14 Gbps effective memory clock.

Q: What is the launch MSRP difference?

A: The RTX 2060 SUPER launched at 399 USD, while the RTX 2070 launched at 499 USD. This is the only price information available in the data.

Specification Differences

The RTX 2070 and RTX 2060 SUPER share identical memory configurations, but their core specifications differ. The RTX 2070 has 2,304 shading units, 144 TMUs, and 64 ROPs, while the RTX 2060 SUPER has 2,176 shading units, 136 TMUs, and the same 64 ROPs. The 2070 also features 36 ray tracing cores and 288 tensor cores, compared to 34 and 272 on the 2060 SUPER.

Clock speeds favor the 2060 SUPER. Its base clock is 1,470 MHz and boost clock is 1,650 MHz, versus the 2070’s 1,410 MHz base and 1,620 MHz boost. This clock advantage helps the 2060 SUPER achieve a higher pixel rate of 105.6 GPixel/s and texture rate of 224.4 GTexel/s, compared to the 2070’s 103.7 GPixel/s and 233.3 GTexel/s. The 2070 still has higher FP32 throughput at 7.465 TFLOPS versus 7.181 TFLOPS.

Both cards have a TDP of 175 W, require a 450 W power supply, and use a single 8-pin power connector. They share the same dimensions (229 mm length, 113 mm height, 35 mm width) and dual-slot form factor. The display outputs are identical: 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C. The release dates differ, with the 2070 launching on 2018-10-16 and the 2060 SUPER on 2019-07-08.

Architecture Differences

Both cards are built on the same TU106 chip using NVIDIA’s Turing architecture on a 12 nm process at TSMC. They share the same transistor count of 10,800 million, die size of 445 mm², and transistor density of 24.3M per mm². The architecture is identical, meaning the differences in performance come from the configuration of the chip rather than fundamental design changes.

The core configuration is where they diverge. The RTX 2070 has 128 more shading units, 8 more TMUs, 2 more ray tracing cores, and 16 more tensor cores than the RTX 2060 SUPER. However, the 2060 SUPER compensates with higher clock speeds, boosting 30 MHz higher than the 2070. This clock advantage allows the 2060 SUPER to win several rasterization-based tests despite having fewer cores.

Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, with identical API support. They also share the same PCIe 3.0 x16 bus interface. The production status for both is end-of-life, and both have GeForce 10 as their predecessor and GeForce 30 as their successor in the product line.

Where Each One Wins

The RTX 2060 SUPER is the winner for gaming-focused workloads. Its 22% lead in 3DMark Steel Nomad DX12 is the strongest indicator, but it also wins across the Passmark DirectX tests (9, 11, 12) by margins ranging from 0.8% to 3.2%. The 2060 SUPER also wins Passmark G3D (16,462 vs 16,094) and Passmark G2D (854 vs 819), making it the better choice for general graphics performance and legacy DirectX titles.

The RTX 2070 wins in compute and modern API performance. Its 6.6% Vulkan lead and 3.9% OpenCL lead make it the superior choice for GPU compute tasks such as rendering, scientific simulation, or any workload using those APIs. The 2070 also wins Passmark DirectX 10 (tied at 111) and has a higher average benchmark score overall (18,789 vs 18,093), suggesting it has slightly better all-around performance when averaged across all tests.

The Passmark GPU Compute result complicates the picture: the 2060 SUPER wins this test by 4.6% despite the 2070’s higher core count. This indicates that for some compute workloads, the 2060 SUPER’s higher clock speeds (1,470 MHz base vs 1,410 MHz) outweigh the 2070’s additional cores. Users should consider their specific software: if it relies on OpenCL or Vulkan, the 2070 is better; if it uses a compute library that favors clock speed, the 2060 SUPER may surprise.

For users who want the highest average performance across a broad suite of benchmarks, the RTX 2070 edges ahead. For users who prioritize modern gaming performance, especially DX12 titles, the RTX 2060 SUPER is the clear winner. Both cards are end-of-life products, so availability and driver support are equal considerations.