AMD Radeon RX 5700 XT vs NVIDIA GeForce GTX 1070 Ti Comparison

FAQ

Q: Which GPU has a higher average benchmark score, the AMD Radeon RX 5700 XT or the NVIDIA GeForce GTX 1070 Ti?

A: The AMD Radeon RX 5700 XT records an average benchmark score of 16361, compared to 14277 for the NVIDIA GeForce GTX 1070 Ti. The AMD card sits at the 59th percentile of all GPUs, while the NVIDIA card sits at the 56th percentile.

Q: How large is the performance gap in the 3DMark Steel Nomad DX12 test?

A: The RX 5700 XT scores 2134, while the GTX 1070 Ti scores 1471. That is a 45.1% advantage for the AMD card, its biggest head-to-head win outside of the Geekbench Metal test.

Q: Which GPU wins the Geekbench OpenCL benchmark, and by how much?

A: The NVIDIA GeForce GTX 1070 Ti wins Geekbench OpenCL with a score of 49219 against 9524 for the RX 5700 XT. That represents an 80.6% deficit for the AMD card, the only head-to-head test the NVIDIA card wins.

Q: What are the memory specifications of each card?

A: Both cards have 8 GB of memory and a 256-bit bus width. The RX 5700 XT uses GDDR6 with 448.0 GB/s of bandwidth, while the GTX 1070 Ti uses GDDR5 with 256.3 GB/s of bandwidth.

Q: What process node does each GPU use?

A: The AMD Radeon RX 5700 XT is built on a 7 nm process at TSMC, while the NVIDIA GeForce GTX 1070 Ti is built on a 16 nm process at TSMC. The RX 5700 XT also has a higher transistor density of 41.0M per mm² versus 22.9M per mm².

Q: What is the launch MSRP for both cards?

A: The launch MSRP for both the AMD Radeon RX 5700 XT and the NVIDIA GeForce GTX 1070 Ti is 399 USD. Both cards are end-of-life products.

The Verdict

The data is unambiguous: the AMD Radeon RX 5700 XT is the stronger card in nearly every measured scenario. It wins 10 of the 11 head-to-head benchmarks and posts an average score of 16361, which is 14.6% above the GTX 1070 Ti's 14277. The only test the NVIDIA card takes is Geekbench OpenCL, where it leads by an enormous margin. That single win is an outlier in an otherwise consistent sweep.

Buyers who prioritize modern DirectX 12 performance, Vulkan workloads, or general 3D rendering should select the RX 5700 XT. Its 45.1% lead in 3DMark Steel Nomad DX12 and 12% lead in Geekbench Vulkan demonstrate a clear generational advantage in low-level APIs. The GTX 5700 XT also wins every PassMark API test, including a 25.5% lead in DirectX 12 and a 10.9% lead in PassMark G3D.

The GTX 1070 Ti remains relevant only for workloads that heavily favor its OpenCL implementation. Its 49219 OpenCL score dwarfs the RX 5700 XT's 9524, a 416.8% advantage. For compute tasks using that API, the older Pascal card is the correct choice. Outside of that niche, the RX 5700 XT is the recommended pick across gaming, rasterization, and general GPU compute.

Head-to-Head Benchmarks

The head-to-head results show a dominant performance profile for the RX 5700 XT. In 3DMark Steel Nomad DX12, the AMD card scores 2134 against 1471, a 45.1% delta. That is the largest margin in any DirectX-related test. Geekbench Metal is even more lopsided: 74647 for AMD against 22180 for NVIDIA, a 236.6% delta, though Metal is a macOS-specific API and not a typical Windows gaming metric. Geekbench Vulkan lands at 68310 for AMD and 61006 for NVIDIA, a 12% advantage.

PassMark results reinforce the trend across different API generations. DirectX 10 scores 97 for AMD and 90 for NVIDIA, a 7.8% delta. DirectX 11 improves to 116 against 107, an 8.4% delta. DirectX 12 shows 64 against 51, a 25.5% delta. DirectX 9 yields 226 against 207, a 9.2% delta. PassMark G2D is close at 922 versus 876, a 5.3% delta. PassMark G3D sees 16277 versus 14674, a 10.9% delta. PassMark GPU Compute scores 7650 against 7165, a 6.8% delta.

The NVIDIA card's only win in Geekbench OpenCL deserves scrutiny. NVIDIA scores 49219 while AMD scores 9524, a 80.6% delta. The magnitude of this win indicates that OpenCL compute on the GTX 1070 Ti scales far better than its rasterization and other compute metrics suggest. Every other compute-oriented benchmark, including PassMark GPU Compute, favors AMD. Overall, the database records 10 wins for AMD and 1 win for NVIDIA.

Specification Differences

The two cards differ across nearly every hardware specification category. The process node is 7 nm for AMD and 7 nm for NVIDIA, with TSMC as the foundry for both, though the GTX 1070 Ti uses a 16 nm node. Transistor count is 10,300 million for AMD versus 7,200 million for NVIDIA, but die size is 251 mm² for AMD versus 314 mm² for NVIDIA. Transistor density is 41.0M per mm² for AMD versus 22.9M per mm² for NVIDIA.

Clock speeds diverge significantly. The RX 5700 XT has a base clock of 1605 MHz and a boost clock of 1905 MHz, plus a game clock of 1755 MHz. The GTX 1070 Ti has a base clock of 1607 MHz and a boost clock of 1683 MHz, with no game clock listed. Memory clocks are 1750 MHz (14 Gbps effective) for AMD and 2002 MHz (8 Gbps effective) for NVIDIA.

Memory type differs: GDDR6 for AMD versus GDDR5 for NVIDIA. Bandwidth is 448.0 GB/s for AMD versus 256.3 GB/s for NVIDIA. The RX 5700 XT has 2560 shading units, 160 TMUs, and 64 ROPs. The GTX 1070 Ti has 2432 shading units, 152 TMUs, and 64 ROPs. Pixel rate is 121.9 GPixel/s for AMD versus 107.7 GPixel/s for NVIDIA. Texture rate is 304.8 GTexel/s for AMD versus 255.8 GTexel/s for NVIDIA.

FP32 performance is 9.754 TFLOPS for AMD versus 8.186 TFLOPS for NVIDIA. FP16 performance is 19.51 TFLOPS (2:1) for AMD versus 127.9 GFLOPS (1:64) for NVIDIA. TDP is 225 W for AMD versus 180 W for NVIDIA. Power connectors are 1x 6-pin plus 1x 8-pin for AMD versus 1x 8-pin for NVIDIA. Suggested PSU is 550 W for AMD versus 450 W for NVIDIA. Bus interface is PCIe 4.0 x16 for AMD versus PCIe 3.0 x16 for NVIDIA.

Display outputs differ: AMD provides 1x HDMI 2.0b and 3x DisplayPort 1.4a, while NVIDIA provides 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a. Physical dimensions are 272 mm length, 111 mm height, and 36 mm width for AMD versus 267 mm length, 112 mm height, and 40 mm width for NVIDIA. Both cards are dual-slot designs. Both support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Neither card has ray tracing cores or tensor cores.

The release dates differ by roughly two years: AMD launched on 2019-07-06, NVIDIA on 2017-11-01. The AMD card's predecessor is Vega and successor is Navi II. The NVIDIA card's predecessor is GeForce 900 and successor is GeForce 20.

Architecture Differences

The RX 5700 XT uses the Navi 10 chip with RDNA 1.0 architecture, part of the Radeon RX 5000 series. The GTX 1070 Ti uses the GP104 chip with Pascal architecture, part of the GeForce 10-series. The architectural lineage alone explains many benchmark deltas: RDNA 1.0 was designed around higher clocks and higher efficiency per clock cycle compared to the older Pascal design. The 7 nm TSMC process gives AMD a density advantage of 41.0M transistors per mm² against 22.9M for the 16 nm Pascal chip.

Cache and compute resources also diverge. The RDNA 1.0 chip packs 2560 shading units and 160 TMUs, while Pascal has 2432 shading units and 152 TMUs. The ROP count is identical at 64. The AMD card doubles FP16 throughput with a 2:1 ratio, reaching 19.51 TFLOPS, while the NVIDIA card runs FP16 at a 1:64 ratio, just 127.9 GFLOPS. The memory subsystem is a key architectural difference: GDDR6 at 448.0 GB/s versus GDDR5 at 256.3 GB/s, a 74.8% bandwidth advantage for AMD.

PCIe generation differs, with AMD supporting PCIe 4.0 x16 and NVIDIA limited to PCIe 3.0 x16. Display output support is also distinct: AMD omits DVI while NVIDIA includes 1x DVI alongside HDMI and DisplayPort outputs. Both cards share the same API support set, including DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Power delivery requirements reflect the performance gap: 225 W TDP with a 550 W suggested PSU for AMD versus 180 W TDP with a 450 W suggested PSU for NVIDIA.

Where Each One Wins

The RX 5700 XT wins across the overwhelming majority of recorded tests. It dominates 3DMark Steel Nomad DX12 with a 45.1% lead, Geekbench Metal with a 236.6% lead, Geekbench Vulkan with a 12% lead, and all PassMark tests including DirectX 9, 10, 11, 12, G2D, G3D, and GPU Compute. Its 59th percentile ranking in the global GPU database, versus the GTX 1070 Ti's 56th percentile, reflects this broad superiority. The 10.9% lead in PassMark G3D and 25.5% lead in DirectX 12 make it the clear choice for modern gaming and D3D12 workloads.

The GTX 1070 Ti wins exactly one benchmark: Geekbench OpenCL. Its score of 49219 versus 9524 is a 416.8% advantage, and it is the only test where the NVIDIA card outperforms the AMD card. This result indicates that OpenCL compute tasks, such as certain physics simulations, video encoding pipelines, or scientific workloads that rely on that API, will run dramatically faster on the Pascal card. For all other purposes, including standard rasterization, DirectX 12, Vulkan, and even PassMark GPU Compute, the RX 5700 XT is the superior choice.

The data supports a conditional recommendation. If the workload is OpenCL-heavy, the GTX 1070 Ti wins outright. If the workload is anything else, the RX 5700 XT delivers the higher score, often by double-digit percentages.