AMD Radeon RX 5700 vs NVIDIA GeForce RTX 2070 SUPER Comparison

The AMD Radeon RX 5700 and NVIDIA GeForce RTX 2070 SUPER are two end-of-life graphics cards from the same 2019 generation, yet benchmark results reveal a starkly different performance profile for each. The RX 5700, built on TSMC’s 7 nm process with the Navi 10 chip, delivers a decisive victory in one modern DirectX 12 test, while the RTX 2070 SUPER, using the larger 12 nm TU104 die, dominates across the vast majority of legacy and compute workloads. With an average benchmark score of 22170, the RX 5700 sits at the 67th percentile of all GPUs, while the RTX 2070 SUPER’s average of 20282 places it at the 65th percentile; despite this, the NVIDIA card wins 9 of 10 head-to-head comparisons, indicating that aggregate scores mask significant per-application advantages. The data suggests that the RX 5700 is a specialist for a specific next-generation API, whereas the RTX 2070 SUPER is the more versatile and consistently faster option for general gaming and compute tasks.

The Verdict

Based strictly on the benchmark data, the NVIDIA GeForce RTX 2070 SUPER is the superior choice for users prioritizing broad software compatibility and consistent high performance. It wins 9 out of 10 head-to-head tests, with particularly large margins in DirectX 10 (132 vs 87, a 34.1% advantage) and DirectX 11 (151 vs 101, a 33.1% advantage). Its PassMark G3D score of 18169 is 21.2% higher than the RX 5700’s 14314, and its Geekbench Vulkan score of 90637 surpasses the AMD card’s 64166 by 29.2%. The RTX 2070 SUPER also leads in compute, with a PassMark GPU Compute score of 7557 versus 6517, a 13.8% difference, and in OpenCL (83358 vs 69603, a 16.5% lead).

The AMD Radeon RX 5700, however, is not without a clear reason for consideration. It wins the 3DMark Steel Nomad DX12 test decisively, scoring 1862 against the RTX 2070 SUPER’s 1651, a 12.8% margin. This single victory is significant because it represents a modern, demanding DirectX 12 workload. For users whose primary applications are built on the latest DirectX 12 features, the RX 5700 may offer better performance. Yet, the RX 5700’s overall position is weaker; its average benchmark score of 22170 is only slightly above the RTX 2070 SUPER’s 20282, yet it loses the majority of tests. The data indicates that the RTX 2070 SUPER is the safer, more well-rounded purchase for a typical gamer, while the RX 5700 is a niche pick for those specifically targeting modern DX12 titles where its architectural strengths are apparent.

Where Each One Wins

The performance split is clear-cut. The NVIDIA GeForce RTX 2070 SUPER wins in every category except one: the 3DMark Steel Nomad DX12 test. This suggests that NVIDIA’s Turing architecture provides superior performance in legacy DirectX 9, 10, and 11 APIs, as well as in general compute and Vulkan. The RTX 2070 SUPER’s wins range from a razor-thin 0.1% in PassMark G2D (878 vs 877) to a substantial 34.1% in DirectX 10. It also shows a 19.4% lead in DirectX 12 PassMark (67 vs 54) and a 21.2% lead in the aggregate PassMark G3D test. In compute, the NVIDIA card is consistently ahead, with a 13.8% lead in GPU Compute and a 16.5% lead in Geekbench OpenCL.

The AMD Radeon RX 5700’s sole win is a significant one, however. In the 3DMark Steel Nomad DX12 test, it outperforms the RTX 2070 SUPER by 12.8% (1862 vs 1651). This is a modern, API-specific benchmark, and the RX 5700’s victory here suggests that its RDNA 1.0 architecture is more efficient at handling the specific instruction sets and workloads of the newest DirectX 12. This makes the RX 5700 a compelling option for users who are exclusively playing or developing for the latest DX12 titles. For everything else, from older games to OpenCL-based applications, the RTX 2070 SUPER is the winner. The data strongly implies that the RTX 2070 SUPER is the more versatile card, while the RX 5700 is a specialist with a narrow but important area of dominance.

Architecture Differences

The two cards are built on fundamentally different architectures and manufacturing processes. The AMD Radeon RX 5700 uses the Navi 10 chip, based on the RDNA 1.0 architecture, and is fabricated on a 7 nm TSMC process. This results in a die size of 251 mm² containing 10,300 million transistors, yielding a transistor density of 41.0M per mm². In contrast, the NVIDIA GeForce RTX 2070 SUPER uses the TU104 chip, based on the Turing architecture, on a larger 12 nm TSMC process. This results in a much larger die size of 545 mm², containing 13,600 million transistors, but with a lower transistor density of 25.0M per mm². The RX 5700’s smaller, denser process gives it a potential efficiency advantage, but the RTX 2070 SUPER has more raw hardware.

The compute resources differ significantly. The RTX 2070 SUPER has 2560 shading units, 160 TMUs, and 64 ROPs, while the RX 5700 has 2304 shading units, 144 TMUs, and 64 ROPs. This gives the NVIDIA card higher theoretical peak rates: 9.062 TFLOPS FP32 and 283.2 GTexel/s texture rate, compared to the RX 5700’s 7.949 TFLOPS and 248.4 GTexel/s. The RTX 2070 SUPER also features dedicated hardware that the RX 5700 lacks entirely: 40 RT cores for ray tracing and 320 tensor cores for AI acceleration. The RX 5700 has no equivalent units. Both cards feature 8 GB of GDDR6 memory on a 256-bit bus with 448.0 GB/s bandwidth, but the RX 5700 uses a PCIe 4.0 x16 interface, while the RTX 2070 SUPER uses PCIe 3.0 x16.

Clock speeds and power draw also differ. The RX 5700 has a base clock of 1465 MHz and a boost of 1725 MHz, with a game clock of 1625 MHz. The RTX 2070 SUPER has a higher base clock of 1605 MHz and a boost of 1770 MHz. Despite the higher clocks, the RTX 2070 SUPER has a higher TDP of 215 W, versus 180 W for the RX 5700, and requires a 550 W suggested PSU versus 450 W. In terms of API support, the RTX 2070 SUPER supports DirectX 12 Ultimate (12_2), while the RX 5700 only supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. The RTX 2070 SUPER also has an additional display output: a USB Type-C port, alongside 1x HDMI 2.0 and 3x DisplayPort 1.4a, while the RX 5700 has 1x HDMI 2.0b and 3x DisplayPort 1.4a.

FAQ

Q: Which card has a higher average benchmark score?

A: The AMD Radeon RX 5700 has a higher average benchmark score of 22170, compared to the NVIDIA GeForce RTX 2070 SUPER’s 20282. However, this aggregate score does not reflect the per-test results, where the RTX 2070 SUPER wins 9 out of 10 comparisons.

Q: Does the RTX 2070 SUPER support ray tracing?

A: Yes, the NVIDIA GeForce RTX 2070 SUPER is equipped with 40 RT cores for ray tracing. The AMD Radeon RX 5700 has no RT cores listed in its specifications.

Q: Which card is better for DirectX 11 games?

A: The NVIDIA GeForce RTX 2070 SUPER is significantly better, scoring 151 in the PassMark DirectX 11 test, which is 33.1% higher than the AMD Radeon RX 5700’s score of 101.

Q: What is the difference in their manufacturing process nodes?

A: The AMD Radeon RX 5700 is manufactured on a 7 nm TSMC process, while the NVIDIA GeForce RTX 2070 SUPER uses a 12 nm TSMC process. This leads to a significant difference in die size and transistor density.

Q: In which benchmark does the RX 5700 beat the RTX 2070 SUPER?

A: The AMD Radeon RX 5700 wins the 3DMark Steel Nomad DX12 test, scoring 1862 versus the RTX 2070 SUPER’s 1651, a 12.8% advantage.

Q: Which card has more shading units?

A: The NVIDIA GeForce RTX 2070 SUPER has 2560 shading units, which is more than the 2304 shading units found on the AMD Radeon RX 5700.

Head-to-Head Benchmarks

The data presents a clear picture of a dominant NVIDIA card with a single, notable AMD exception. The most significant win for the RTX 2070 SUPER comes in the PassMark DirectX 10 test, where it scores 132 against the RX 5700’s 87, a 34.1% difference. This is followed closely by a 33.1% lead in DirectX 11, with scores of 151 and 101, respectively. In the aggregate PassMark G3D test, the RTX 2070 SUPER’s 18169 score is 21.2% higher than the RX 5700’s 14314. The Vulkan performance gap is also substantial, with the NVIDIA card scoring 90637 versus 64166, a 29.2% advantage. These are the largest margins, indicating that the Turing architecture is far more efficient in these APIs.

The RTX 2070 SUPER also demonstrates consistent, if smaller, leads in other areas. Its PassMark DirectX 12 score of 67 is 19.4% higher than the RX 5700’s 54. In compute, the NVIDIA card leads by 13.8% in PassMark GPU Compute (7557 vs 6517) and by 16.5% in Geekbench OpenCL (83358 vs 69603). The DirectX 9 test shows a narrower 6.3% lead for NVIDIA (223 vs 209), and the PassMark G2D test is essentially a tie, with the RTX 2070 SUPER edging out a 0.1% victory (878 vs 877). These results show that the RTX 2070 SUPER is not just faster in one or two areas, but is consistently superior across a broad spectrum of workloads.

The single counterpoint is the 3DMark Steel Nomad DX12 test. Here, the AMD Radeon RX 5700 scores 1862, which is 12.8% higher than the RTX 2070 SUPER’s 1651. This is a significant margin in the most modern benchmark on the list. This result is anomalous when compared to the other tests, suggesting that the RX 5700’s RDNA 1.0 architecture has a specific strength in handling the particular features of this DX12 workload. The win is enough to give the RX 5700 a 1-9 head-to-head record, but it does not offset the NVIDIA card’s overwhelming superiority in the other nine tests. For a user prioritizing a single, modern API, the RX 5700 is the winner; for all other purposes, the RTX 2070 SUPER is the clear benchmark champion.