AMD Radeon RX 6800S vs NVIDIA GeForce RTX 2070 SUPER Comparison

Head-to-Head Benchmarks

The head-to-head comparison between the AMD Radeon RX 6800S and the NVIDIA GeForce RTX 2070 SUPER is lopsided in raw win count, with NVIDIA taking 9 of 10 recorded tests. However, the single AMD victory is a notable one in a modern DirectX 12 workload. In the 3DMark Steel Nomad DX12 test, the RX 6800S scores 1689 against the RTX 2070 SUPER's 1651, a 2.3% advantage. This result indicates that AMD's RDNA 2.0 architecture can edge out the older Turing design in a current-generation rasterization benchmark, even with the NVIDIA card's higher absolute compute figures.

The remaining nine tests all favor the RTX 2070 SUPER, and the margins vary widely. The most decisive NVIDIA win comes in PassMark DirectX 10, where the RTX 2070 SUPER scores 132 versus 91 for the RX 6800S, a 31.1% gap. This is the single largest delta in the entire benchmark set. Legacy API performance is clearly a strength for the NVIDIA card, and the pattern continues in PassMark DirectX 9, where it leads 223 to 200, a 10.3% margin. The DirectX 11 test shows a smaller but still clear advantage: 151 versus 139, a 7.9% difference. The DirectX 12 PassMark test is the closest contest of the entire set, with the RTX 2070 SUPER scoring 67 against 65, a slim 3% edge.

Compute workloads also favor NVIDIA substantially. In Geekbench OpenCL, the RTX 2070 SUPER posts 83358 against 73202, a 12.2% lead. The Vulkan test shows a 90637 to 81272 result, a 10.3% advantage for NVIDIA. PassMark GPU Compute tells a similar story: 7557 versus 6144, an 18.7% margin. The 2D performance gap is also pronounced, with the RTX 2070 SUPER scoring 878 in PassMark G2D versus 732 for the RX 6800S, a 16.6% difference. The overall PassMark G3D score reinforces the trend: 18169 for NVIDIA versus 15908 for AMD, a 12.4% lead.

The aggregate numbers in the database tell a more nuanced story. The RX 6800S has an average benchmark score of 24063 across all recorded tests, while the RTX 2070 SUPER averages 20282. That puts the AMD part at the 69th percentile of all GPUs and the NVIDIA part at the 65th percentile. The nearest rivals for the RX 6800S include the NVIDIA GeForce RTX 2080 SUPER, which averages 24170, a 0.4% difference, and the AMD Radeon RX 9070 at 23877, a 0.8% difference in favor of the 6800S. For the RTX 2070 SUPER, its closest comparables are the NVIDIA Quadro M4000M at 20480, a 1% difference, and the RTX 3070 Mobile at 20534, a 1.2% difference. These percentile and rival figures suggest that despite losing most head-to-head tests, the RX 6800S occupies a slightly higher overall performance tier in the database's full suite, likely because the head-to-head list skews toward legacy and compute workloads where NVIDIA excels.

The Verdict

The data points to a clear split based on workload priorities. For users running modern DirectX 12 titles, the AMD Radeon RX 6800S holds a measurable but modest 2.3% lead in Steel Nomad, and its average benchmark score of 24063 places it above the RTX 2070 SUPER's 20282 average. The RX 6800S also sits at a higher percentile rank, 69 versus 65, and its nearest rival, the RTX 2080 SUPER, is only 0.4% ahead, which means the mobile AMD part is effectively trading blows with a desktop-class flagship from the previous generation.

For everything else, the NVIDIA GeForce RTX 2070 SUPER is the stronger pick. It wins across DirectX 9, 10, 11, and 12 PassMark tests, all Geekbench compute workloads tested, and both PassMark GPU Compute and G2D. The largest margin, 31.1% in DirectX 10, shows that legacy API support is not merely a small edge but a dominant one. The Vulkan and OpenCL margins of roughly 10% to 12% indicate that general compute and cross-platform graphics tasks favor the Turing card. The RTX 2070 SUPER also has a higher raw FP32 throughput, 9.062 TFLOPS versus 8.602 TFLOPS, and more shading units, 2560 versus 2048, which aligns with its compute-heavy benchmark wins.

The verdict depends on the use case. Gamers focused on current DirectX 12 titles and who value the RX 6800S's higher average score and percentile may prefer the AMD card, especially given its 100 W TDP against the RTX 2070 SUPER's 215 W, a power efficiency advantage that matters in constrained systems. Users running legacy titles, compute workloads, or Vulkan-based applications should choose the RTX 2070 SUPER, as the data shows consistent double-digit leads in those areas. The RTX 2070 SUPER also offers a launch MSRP of 499 USD, though no equivalent launch pricing exists for the RX 6800S in the database.

FAQ

Q: Which GPU wins the most head-to-head benchmark tests?

A: The NVIDIA GeForce RTX 2070 SUPER wins 9 of 10 recorded tests, with the only AMD victory coming in the 3DMark Steel Nomad DX12 test.

Q: How much faster is the RX 6800S in the 3DMark Steel Nomad DX12 test?

A: The RX 6800S scores 1689 versus 1651 for the RTX 2070 SUPER, a 2.3% difference.

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

A: The largest gap is in PassMark DirectX 10, where the RTX 2070 SUPER leads by 31.1%, scoring 132 against 91.

Q: Which card has a higher average benchmark score in the database?

A: The AMD Radeon RX 6800S has an average benchmark score of 24063, while the NVIDIA GeForce RTX 2070 SUPER averages 20282.

Q: How do the two cards compare in Geekbench Vulkan performance?

A: The RTX 2070 SUPER scores 90637 in Geekbench Vulkan, which is 10.3% ahead of the RX 6800S's 81272.

Q: What are the closest rivals to each card in the database?

A: The RX 6800S's nearest rival is the NVIDIA GeForce RTX 2080 SUPER, with a 0.4% difference in average score. The RTX 2070 SUPER's nearest rival is the NVIDIA Quadro M4000M, with a 1% difference.

Specification Differences

The two cards differ in nearly every major specification category. The AMD Radeon RX 6800S uses the Navi 23 chip on a 7 nm process, while the NVIDIA GeForce RTX 2070 SUPER uses the TU104 chip on a 12 nm process. Transistor counts differ substantially: the RX 6800S has 11,060 million transistors on a 237 mm² die, while the RTX 2070 SUPER has 13,600 million on a 545 mm² die. Transistor density is 46.7 million per mm² for AMD versus 25.0 million per mm² for NVIDIA, reflecting the process node advantage.

Clock speeds also differ. The RX 6800S has a base clock of 1800 MHz and a boost clock of 2100 MHz, with a game clock of 1975 MHz. The RTX 2070 SUPER has a base clock of 1605 MHz and a boost clock of 1770 MHz, with no game clock listed. Memory clocks are 2000 MHz with 16 Gbps effective for AMD versus 1750 MHz with 14 Gbps effective for NVIDIA. Both cards have 8 GB of GDDR6 memory, but the bus widths differ: 128 bit for the RX 6800S and 256 bit for the RTX 2070 SUPER. Memory bandwidth is 256.0 GB/s for AMD and 448.0 GB/s for NVIDIA.

Compute unit counts and related metrics vary. The RX 6800S has 2048 shading units, 128 TMUs, 64 ROPs, and 32 ray tracing cores. The RTX 2070 SUPER has 2560 shading units, 160 TMUs, 64 ROPs, 40 ray tracing cores, and 320 tensor cores. Pixel rates are 134.4 GPixel/s for AMD and 113.3 GPixel/s for NVIDIA. Texture rates are 268.8 GTexel/s and 283.2 GTexel/s respectively. FP32 performance is 8.602 TFLOPS for AMD and 9.062 TFLOPS for NVIDIA, with FP16 at 17.20 and 18.12 TFLOPS respectively.

Power and physical specifications diverge sharply. The RX 6800S has a TDP of 100 W, is listed as an integrated graphics package (IGP) with no power connectors, and uses a PCIe 4.0 x8 interface. The RTX 2070 SUPER has a TDP of 215 W, is a dual-slot card requiring one 6-pin and one 8-pin power connector, suggests a 550 W PSU, and uses PCIe 3.0 x16. The NVIDIA card has defined dimensions of 267 mm length, 116 mm height, and 35 mm width, while the AMD card lists no dimensions. Display outputs are portable-device dependent for the RX 6800S, while the RTX 2070 SUPER offers 1x HDMI 2.0, 3x DisplayPort 1.4a, and 1x USB Type-C. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Architecture Differences

The architectural divide is between AMD's RDNA 2.0 and NVIDIA's Turing. The RX 6800S belongs to the Radeon RX 6000 series and the Navi Mobile generation, built on a 7 nm TSMC process. The RTX 2070 SUPER belongs to the GeForce 20-series, built on a 12 nm TSMC process. The manufacturing process difference is central: the 7 nm node enables a 237 mm² die with 46.7 million transistors per mm², while the 12 nm node requires a 545 mm² die with 25.0 million transistors per mm². The RX 6800S packs 11,060 million transistors into a much smaller area, while the RTX 2070 SUPER uses 13,600 million across a die more than twice the size.

Ray tracing hardware exists on both cards but in different configurations. The RX 6800S has 32 ray tracing cores, while the RTX 2070 SUPER has 40. The NVIDIA card also includes 320 tensor cores, which the AMD card lacks entirely. This means the RTX 2070 SUPER supports dedicated tensor core workloads, while the RX 6800S has no equivalent hardware. The AMD card compensates in other areas: its game clock of 1975 MHz and boost clock of 2100 MHz are substantially higher than the RTX 2070 SUPER's 1605 MHz base and 1770 MHz boost, which partially offsets the NVIDIA card's higher shading unit count of 2560 versus 2048.

Memory architecture also differs. The RX 6800S uses a 128 bit bus with 256.0 GB/s bandwidth, while the RTX 2070 SUPER uses a 256 bit bus with 448.0 GB/s bandwidth. The NVIDIA card has nearly double the memory bandwidth, which contributes to its stronger performance in compute and legacy API tests. However, the AMD card's higher pixel rate, 134.4 GPixel/s versus 113.3 GPixel/s, suggests a fill-rate advantage in certain rendering scenarios. The RTX 2070 SUPER has a higher texture rate, 283.2 GTexel/s versus 268.8 GTexel/s, and higher FP32 and FP16 throughput. The RX 6800S is an end-of-life product released in January 2022, with its predecessor being Polaris Mobile. The RTX 2070 SUPER, also end-of-life, was released in July 2019, with its predecessor being GeForce 10 and its successor being GeForce 30. Both cards support the same API feature set, including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, so the architectural differences manifest primarily in raw throughput, memory bandwidth, and power efficiency rather than API compatibility.