NVIDIA GeForce RTX 2070 vs NVIDIA GeForce RTX 2070 SUPER Comparison

The NVIDIA GeForce RTX 2070 SUPER and the NVIDIA GeForce RTX 2070 are two Turing-architecture cards from the GeForce 20-series, released in 2019 and 2018 respectively. The data shows a clear hierarchy: the SUPER variant wins all ten head-to-head benchmark comparisons, with an average benchmark score of 20282 against the RTX 2070’s 18789, placing the former at the 65th percentile of all GPUs and the latter at the 63rd. This analysis breaks down where each card excels, the magnitude of the performance gap, and the architectural differences that drive it.

Where Each One Wins

The RTX 2070 SUPER is the outright winner in every single benchmark category recorded, making the use-case split a matter of degree rather than direction. The most decisive victories come in legacy DirectX API tests, where the SUPER leads by double-digit percentages. In Passmark’s DirectX 10 test, the SUPER scores 132 against the RTX 2070’s 111, a 18.9% advantage. Similarly, the DirectX 11 test shows a 17.1% lead (151 vs 129), and the DirectX 12 test shows an 11.7% lead (67 vs 60). These results indicate that the SUPER’s higher raw compute throughput translates directly into faster frame rates across both older and modern graphics APIs.

The compute-oriented workloads also strongly favor the SUPER. The Passmark GPU Compute test yields a 17.9% delta (7557 vs 6411), which aligns with the SUPER’s substantially higher FP32 throughput of 9.062 TFLOPS compared to the RTX 2070’s 7.465 TFLOPS. For users running GPGPU tasks, physics simulations, or rendering workloads that rely on shader compute, the SUPER is the clear pick. The Geekbench OpenCL score reinforces this, with the SUPER scoring 83358 versus 79966, a 4.2% edge.

In pure rasterization, the Passmark G3D test shows a 12.9% lead (18169 vs 16094). This is the broadest indicator of overall gaming performance, and the gap is substantial. The SUPER’s advantage in Vulkan is even larger; Geekbench Vulkan scores show a 9.8% delta (90637 vs 82521). The newer 3DMark Steel Nomad DX12 test, which is more representative of modern AAA game loads, shows a smaller but still meaningful 5.2% lead (1651 vs 1569). The smallest win is in DirectX 9 (5.7% delta, 223 vs 211), which suggests that for very old titles, the difference narrows, likely due to CPU or driver overhead becoming the bottleneck rather than GPU shader throughput.

The RTX 2070, having zero wins, has no category where it outperforms the SUPER. Its only relative advantage is power consumption; with a TDP of 175 W versus the SUPER’s 215 W, it draws less power and requires only a single 8-pin connector instead of the SUPER’s 1x 6-pin + 1x 8-pin setup. The RTX 2070 also has a shorter physical footprint at 229 mm (9 inches) compared to the SUPER’s 267 mm (10.5 inches), making it easier to fit in compact cases. However, in terms of raw benchmark scores, the RTX 2070 does not win a single test.

The Verdict

The data dictates a simple conclusion: choose the RTX 2070 SUPER if performance is the sole metric. It wins every benchmark in the comparison, with an average score that is 7.9% higher (20282 vs 18789). The SUPER’s percentile ranking of 65 versus the RTX 2070’s 63 confirms that it sits in a higher performance tier among all GPUs. For gamers, the Passmark G3D and DirectX 11/12 results show a clear 11.7% to 17.1% improvement in the APIs most commonly used by modern titles. For compute users, the 17.9% lead in GPU Compute and the higher FP32 rate make it the only rational choice.

The RTX 2070 is only appealing in a narrow set of circumstances derived from the data. Its lower 175 W TDP and single 8-pin power connector mean it can be dropped into systems with a 450 W suggested PSU, whereas the SUPER requires a 550 W PSU. Its 229 mm length is also more accommodating for smaller chassis. If a user has a strict power budget or a physically constrained case, the RTX 2070 is the fallback option, but they will sacrifice 12.9% in G3D performance and 17.9% in compute performance. The RTX 2070 also shares the same 8 GB GDDR6 memory and 448.0 GB/s bandwidth as the SUPER, so memory capacity is not a differentiator. There is no scenario in the benchmark data where the RTX 2070 is the faster card; the only reasons to pick it are physical or power-related.

The launch MSRP for both cards is 499 USD, meaning the SUPER delivers superior performance at the same launch price point, making the RTX 2070 a difficult recommendation unless the power or size constraints are absolute.

Head-to-Head Benchmarks

The biggest win for the RTX 2070 SUPER is in the Passmark DirectX 10 test, where it scores 132 versus the RTX 2070’s 111, a 18.9% delta. This is followed closely by the Passmark GPU Compute test, where the SUPER’s 7557 score beats the RTX 2070’s 6411 by 17.9%. The DirectX 11 test shows a 17.1% gap (151 vs 129), and the Passmark G3D test shows a 12.9% gap (18169 vs 16094). These four tests represent the largest performance deltas, all exceeding 12%.

The Geekbench Vulkan test is another notable win for the SUPER, with a 9.8% delta (90637 vs 82521). This is significant because Vulkan is a low-overhead API used in many modern game engines, and a near-10% advantage indicates that the SUPER’s extra shading units and higher clocks provide tangible benefits in draw-call-heavy scenarios. The Passmark DirectX 12 test shows an 11.7% delta (67 vs 60), which, while a smaller absolute score difference, represents a substantial relative improvement for DX12 titles.

The smallest deltas are in 3DMark Steel Nomad DX12 (5.2%, 1651 vs 1569) and Passmark DirectX 9 (5.7%, 223 vs 211). The Steel Nomad result is interesting because it is a modern, demanding test; the smaller gap suggests that the RTX 2070’s 256-bit memory bus and 448.0 GB/s bandwidth are sufficient to keep pace in certain memory-bound scenarios, reducing the impact of the SUPER’s higher compute throughput. The DirectX 9 result likely reflects a bottleneck elsewhere, as both cards are far beyond what legacy DX9 games require.

The remaining tests, Geekbench OpenCL (4.2% delta, 83358 vs 79966) and Passmark G2D (7.2% delta, 878 vs 819), show moderate wins for the SUPER. The G2D test is a 2D graphics performance metric, and the SUPER’s higher pixel rate of 113.3 GPixel/s versus the RTX 2070’s 103.7 GPixel/s explains its advantage in this area.

FAQ

Q: Is the RTX 2070 SUPER faster than the RTX 2070 in every benchmark?

A: Yes. The RTX 2070 SUPER wins all ten head-to-head tests in the data, with deltas ranging from 4.2% in Geekbench OpenCL to 18.9% in Passmark DirectX 10. The RTX 2070 records zero wins.

Q: What is the largest performance gap between the two cards?

A: The largest gap is in the Passmark DirectX 10 test, where the RTX 2070 SUPER scores 132 against the RTX 2070’s 111, a 18.9% difference. The Passmark GPU Compute test is close behind at 17.9%.

Q: How do the two cards compare in average benchmark score?

A: The RTX 2070 SUPER has an average benchmark score of 20282, while the RTX 2070 scores 18789. This represents a 7.9% advantage for the SUPER, which also ranks at the 65th percentile of all GPUs versus the RTX 2070’s 63rd percentile.

Q: Are the memory specifications identical between the two cards?

A: Yes. Both cards feature 8 GB of GDDR6 memory on a 256-bit bus, with 448.0 GB/s bandwidth and 14 Gbps effective memory speed. Memory capacity and bandwidth are not differentiating factors.

Q: What are the physical and power differences?

A: The RTX 2070 has a TDP of 175 W and requires a single 8-pin power connector, while the RTX 2070 SUPER has a TDP of 215 W and requires a 1x 6-pin + 1x 8-pin configuration. The suggested PSU is 450 W for the RTX 2070 and 550 W for the SUPER. The RTX 2070 is also shorter at 229 mm (9 inches) versus the SUPER’s 267 mm (10.5 inches).

Q: Which card has more shader units and RT cores?

A: The RTX 2070 SUPER has 2560 shading units and 40 RT cores, while the RTX 2070 has 2304 shading units and 36 RT cores. The SUPER also has 160 TMUs versus 144, and 320 tensor cores versus 288.

Architecture Differences

The RTX 2070 SUPER is built on the TU104 chip, while the RTX 2070 uses the TU106 chip. Both are manufactured on TSMC’s 12 nm process, but the TU104 is a physically larger die at 545 mm² compared to the TU106’s 445 mm², containing 13,600 million transistors versus 10,800 million. This larger die allows the SUPER to pack more compute resources: 2560 shading units, 160 TMUs, and 40 RT cores, compared to the RTX 2070’s 2304 shading units, 144 TMUs, and 36 RT cores. Both have 64 ROPs.

The clock speeds differ significantly. The RTX 2070 SUPER has a base clock of 1605 MHz and a boost clock of 1770 MHz, whereas the RTX 2070 operates at 1410 MHz base and 1620 MHz boost. This higher clock speed, combined with the additional shading units, yields a substantial increase in compute throughput: the SUPER achieves 9.062 TFLOPS FP32 and 18.12 TFLOPS FP16 (2:1), while the RTX 2070 manages 7.465 TFLOPS FP32 and 14.93 TFLOPS FP16 (2:1).

The pixel and texture rates also scale accordingly. The SUPER has a pixel rate of 113.3 GPixel/s and a texture rate of 283.2 GTexel/s, versus the RTX 2070’s 103.7 GPixel/s and 233.3 GTexel/s. This explains the SUPER’s advantage in 2D and texture-heavy workloads, as seen in the Passmark G2D score of 878 versus 819.

Memory configurations are identical: 8 GB of GDDR6 on a 256-bit bus with 448.0 GB/s bandwidth. The display outputs differ slightly; the RTX 2070 SUPER offers 1x HDMI 2.0 and 3x DisplayPort 1.4a, while the RTX 2070 offers 1x DVI, 1x HDMI 2.0, and 2x DisplayPort 1.4a. Both include 1x USB Type-C. The SUPER is slightly wider (116 mm vs 113 mm) and longer (267 mm vs 229 mm), but both are dual-slot cards with a 35 mm width. The RTX 2070’s lower TDP of 175 W and single 8-pin power requirement make it a more power-efficient option, but the SUPER’s 215 W TDP is the price for its 7.9% higher average benchmark score.