AMD Radeon RX 470 vs NVIDIA GeForce RTX 2080 SUPER Comparison

Where Each One Wins

The benchmark distribution is decisively one-sided. The NVIDIA GeForce RTX 2080 SUPER wins all three recorded head-to-head tests, while the AMD Radeon RX 470 records no wins in the shared benchmark suite. This is not a close contest; the data shows a consistent margin favoring the Turing-based card across every workload that both GPUs were tested in.

The RTX 2080 SUPER demonstrates its largest advantage in compute-oriented workloads. In Geekbench OpenCL, it scores 99,226 against 33,568 for the RX 470, a gap of 65.7 points per thousand or roughly 66% higher. The Vulkan result follows a similar pattern: 111,284 versus 39,884, which is approximately 64% higher. These are not marginal differences; they indicate that the NVIDIA card sustains its lead even when the API changes, which points to a fundamental throughput advantage rather than an optimization quirk.

The 3DMark Steel Nomad DX12 test tells the same story at a different scale. The RTX 2080 SUPER posts 1,882, while the RX 470 manages 842. That is a 55.3% deficit for the AMD card. Steel Nomad is a modern DirectX 12 workload that stresses geometry, shading, and memory bandwidth simultaneously. The fact that the NVIDIA card nearly doubles the AMD score in this test suggests that the RX 470 struggles with the combined load, whereas the RTX 2080 SUPER has enough headroom in both compute and rasterization to absorb the demands.

For the RX 470, the positive takeaway is its standing relative to other GPUs in the database, not relative to this specific rival. Its average benchmark score is 28,996, which places it in the 74th percentile of all GPUs. That is a respectable position for a card from the Arctic Islands generation. The RTX 2080 SUPER, by contrast, has an average score of 24,170, placing it in the 69th percentile. This is an important nuance: despite winning every head-to-head test by wide margins, the RTX 2080 SUPER's overall average is dragged down by additional tests in its benchmark set, including several PassMark results that are low (140 in DirectX 10, 165 in DirectX 11, 75 in DirectX 12). The RX 470, with only four benchmark entries, does not carry those low-scoring legacy tests.

So the use-case split is clear. If the workload is a modern 3D game or a compute-heavy task using Vulkan or OpenCL, the RTX 2080 SUPER is the superior choice by a substantial margin. If the use case is broader, encompassing the full range of database benchmarks, the RX 470's average score is actually higher, and its percentile ranking is better. The RX 470 wins on consistency across its tested set; the RTX 2080 SUPER wins on peak performance.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon RX 470 has an average benchmark score of 28,996, which is higher than the NVIDIA GeForce RTX 2080 SUPER's average of 24,170. The RX 470 also sits at the 74th percentile of all GPUs, compared to the 69th percentile for the RTX 2080 SUPER.

Q: What is the largest performance gap in the head-to-head tests?

A: The largest gap is in Geekbench OpenCL, where the RTX 2080 SUPER scores 99,226 against the RX 470's 33,568, a 66.2% difference. The Vulkan test is close behind at 64.2%, and the 3DMark Steel Nomad DX12 test shows a 55.3% difference.

Q: Does the RX 470 win any benchmark against the RTX 2080 SUPER?

A: No. In the recorded head-to-head benchmarks, the RX 470 has zero wins. The RTX 2080 SUPER wins all three shared tests: 3DMark Steel Nomad DX12, Geekbench OpenCL, and Geekbench Vulkan.

Q: How do the two cards compare in memory capacity and bandwidth?

A: The RTX 2080 SUPER has 8 GB of GDDR6 memory with 495.9 GB/s bandwidth. The RX 470 has 4 GB of GDDR5 memory with 211.2 GB/s bandwidth. Both use a 256-bit memory bus.

Q: What are the DirectX and Vulkan API support levels for each card?

A: The RX 470 supports DirectX 12 (12_0) and Vulkan 1.3. The RTX 2080 SUPER supports DirectX 12 Ultimate (12_2) and Vulkan 1.4. Both support OpenGL 4.6.

Q: Which card has more shading units and texture mapping units?

A: The RTX 2080 SUPER has 3,072 shading units and 192 TMUs. The RX 470 has 2,048 shading units and 128 TMUs. The NVIDIA card also has 64 ROPs versus 32 for the AMD card.

Head-to-Head Benchmarks

The most telling result is the 3DMark Steel Nomad DX12 test. This is a modern, API-level stress test that combines ray-traced effects with traditional rasterization. The RTX 2080 SUPER scores 1,882, while the RX 470 scores 842. The delta is 55.3% in favor of NVIDIA. To put that in context, the RX 470's score is less than half of the RTX 2080 SUPER's. This is not a workload where the older GCN architecture can compensate with driver maturity or API tricks; the raw throughput difference is too large.

Geekbench OpenCL shows an even wider margin. The RTX 2080 SUPER posts 99,226, which is nearly triple the RX 470's 33,568. The 66.2% delta is the largest in the entire head-to-head set. OpenCL is a general-purpose compute API, and the RTX 2080 SUPER's Turing architecture with 384 tensor cores and 48 RT cores clearly accelerates these workloads far beyond what the RX 470's GCN 4.0 design can deliver. The RX 470's FP32 throughput is 4.940 TFLOPS, while the RTX 2080 SUPER reaches 11.15 TFLOPS, a 2.26x advantage. That raw compute gap explains the OpenCL result.

Geekbench Vulkan narrows the gap slightly but still favors NVIDIA decisively. The RTX 2080 SUPER scores 111,284, the RX 470 scores 39,884, and the delta is 64.2%. Vulkan is often considered a more efficient API for AMD hardware, yet the database shows that the RX 470 cannot close the gap. The RTX 2080 SUPER's higher memory bandwidth (495.9 GB/s versus 211.2 GB/s) and higher pixel rate (116.2 GPixel/s versus 38.59 GPixel/s) provide structural advantages that API selection does not erase.

One notable point is that the RX 470 does not have any benchmark where it wins, but its average score across all recorded tests is higher than the RTX 2080 SUPER's because the NVIDIA card's benchmark set includes low PassMark scores. The RTX 2080 SUPER's PassMark G3D score is 19,490, which is respectable, but its DirectX 9 score is 227, DirectX 10 is 140, DirectX 11 is 165, and DirectX 12 is 75. Those legacy tests drag its average down. The RX 470 has no PassMark results in the database, so its average is based solely on its stronger 3DMark and Geekbench scores.

Specification Differences

The two cards differ in nearly every measurable specification. The RTX 2080 SUPER uses a 12 nm process node from TSMC, while the RX 470 uses a 14 nm node from GlobalFoundries. The NVIDIA chip, TU104, contains 13,600 million transistors on a 545 mm² die. The AMD chip, Ellesmere, contains 5,700 million transistors on a 232 mm² die. Transistor density is similar (25.0M per mm² versus 24.6M per mm²), but the total transistor count is 2.4x higher on the NVIDIA card.

Memory is another major divergence. The RTX 2080 SUPER has 8 GB of GDDR6 running at 15.5 Gbps effective, yielding 495.9 GB/s bandwidth. The RX 470 has 4 GB of GDDR5 at 6.6 Gbps effective, yielding 211.2 GB/s. Both use a 256-bit bus, but the GDDR6 standard and higher clock rate give NVIDIA a 2.35x bandwidth advantage.

Clock speeds also favor NVIDIA. The RTX 2080 SUPER has a base clock of 1650 MHz and a boost clock of 1815 MHz. The RX 470 has a base clock of 926 MHz and a boost clock of 1206 MHz. The NVIDIA card's boost clock is 50% higher than the AMD card's boost clock. This clock advantage, combined with more shading units (3,072 versus 2,048) and more TMUs (192 versus 128), produces a large throughput gap.

Power requirements differ substantially. The RTX 2080 SUPER has a TDP of 250 W and requires a 600 W suggested PSU with 1x 6-pin and 1x 8-pin connectors. The RX 470 has a TDP of 120 W and requires a 300 W suggested PSU with a single 6-pin connector. The physical dimensions also differ: the RTX 2080 SUPER is 267 mm long and 116 mm tall, while the RX 470 is 240 mm long and 95 mm tall. Both are dual-slot cards with a width of 35 mm.

Display outputs are similar but not identical. Both have 1x HDMI 2.0 and 3x DisplayPort 1.4a. The RTX 2080 SUPER adds 1x USB Type-C, which the RX 470 lacks. The RTX 2080 SUPER supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the RX 470 supports DirectX 12 (12_0) and Vulkan 1.3.

Architecture Differences

The architectural divide is generational. The RX 470 is built on GCN 4.0, AMD's fourth iteration of its Graphics Core Next design, and belongs to the Arctic Islands generation (RX 400). The RTX 2080 SUPER is built on Turing, NVIDIA's architecture from the GeForce 20-series. Turing introduces dedicated hardware that GCN 4.0 lacks entirely: 48 RT cores for ray tracing and 384 tensor cores for AI acceleration. The RX 470 has neither.

The compute characteristics reflect this. The RX 470 offers FP32 at 4.940 TFLOPS and FP16 at the same 4.940 TFLOPS, a 1:1 ratio. The RTX 2080 SUPER offers FP32 at 11.15 TFLOPS and FP16 at 22.30 TFLOPS, a 2:1 ratio. The NVIDIA card's FP16 advantage is particularly pronounced: 22.30 TFLOPS versus 4.940 TFLOPS. This makes the RTX 2080 SUPER far more capable in mixed-precision workloads, which are common in machine learning and certain scientific computations.

Texture and pixel rates also diverge. The RTX 2080 SUPER achieves a texture rate of 348.5 GTexel/s and a pixel rate of 116.2 GPixel/s. The RX 470 achieves 154.4 GTexel/s and 38.59 GPixel/s. The NVIDIA card is 2.26x faster in texturing and 3.01x faster in pixel fill. These are structural limits of the GCN 4.0 design; more ROPs (64 versus 32) and more TMUs (192 versus 128) on the NVIDIA card directly enable these rates.

The transistor budgets tell the story of architectural ambition. The TU104 chip packs 13,600 million transistors into 545 mm², while the Ellesmere chip fits 5,700 million into 232 mm². The RTX 2080 SUPER's larger die accommodates the RT and tensor cores, which consume significant area. The RX 470's smaller die is a more modest design from 2016, optimized for mainstream gaming at lower power. The production status for both is end-of-life, but the release dates differ by roughly three years: the RX 470 launched in August 2016, and the RTX 2080 SUPER launched in July 2019.

In terms of API support, the RTX 2080 SUPER's DirectX 12 Ultimate (12_2) feature level is a full tier above the RX 470's DirectX 12 (12_0). This matters for modern games that use ray tracing and mesh shaders, which require the newer feature level. The Vulkan support also differs: 1.4 on NVIDIA versus 1.3 on AMD. The RX 470's OpenGL support matches at 4.6, but the overall API portfolio favors the newer card.