AMD Radeon RX 580 vs NVIDIA GeForce RTX 2060 Comparison

Head-to-Head Benchmarks

The recorded data shows a decisive overall victory for the NVIDIA GeForce RTX 2060, which wins 9 of the 10 shared benchmark comparisons. The only exception is the Passmark G2D test, where the AMD Radeon RX 580 edges ahead by 3.1%, scoring 769 against 745. That is a narrow margin, and it stands alone as the sole AMD win across the entire suite.

The largest single gap appears in the Passmark DirectX 10 test, where the RTX 2060 scores 98 versus 46 for the RX 580, a 113% lead. That is more than double the older card's result, and it suggests a fundamental difference in how the two architectures handle legacy DirectX 10 workloads. The DirectX 11 test tells a similar story: 110 for the RTX 2060 against 60 for the RX 580, an 83.3% advantage. In both cases, the NVIDIA card is not merely faster, it is operating in a different performance tier.

Compute workloads also favor the RTX 2060 heavily. In Geekbench OpenCL, the RTX 2060 posts 65,014 versus 37,453, a 73.6% lead. The Vulkan result is closer but still decisive: 65,846 against 45,173, a 45.8% advantage. These are substantial margins that point to the Turing architecture's efficiency in modern API and compute environments. The Passmark GPU Compute test reinforces this pattern, with the RTX 2060 at 5,488 and the RX 580 at 3,488, a 57.3% gap.

The overall 3D performance metric, Passmark G3D, shows the RTX 2060 at 14,111 against 8,813 for the RX 580, a 60.1% difference. That is the headline figure for traditional rasterization performance, and it is a commanding lead. The 3DMark Steel Nomad DX12 test, a more modern workload, shows a smaller but still clear 23.6% advantage for the RTX 2060 (1,242 versus 1,005). The DirectX 12 Passmark result is nearly identical in margin at 23.3% (53 versus 43). Even in DirectX 9, the oldest API tested, the RTX 2060 leads by 53.2% (190 versus 124).

The average benchmark score gap is also telling. The RTX 2060 averages 15,290 across all recorded tests, while the RX 580 averages 12,928. That places the RTX 2060 in the 58th percentile of all GPUs, versus the 53rd percentile for the RX 580. The RTX 2060's nearest rivals list includes the AMD Radeon RX 7600 with an average score of 15,271 and a delta of 0.8%, meaning the RTX 2060 is effectively tied with that newer card in aggregate performance. The RX 580, by contrast, sits near the NVIDIA GeForce GTX 1660 SUPER, which averages 12,986 with a delta of -0.4%.

For the RX 580, the data indicates that its most comparable peers are older NVIDIA designs like the GTX 1660 SUPER and the RTX 3050 Ti Mobile, all within 0.5% or less of its average score. The RTX 2060, meanwhile, rubs shoulders with the RX 7600 and the RTX 3050 OEM, which is a much higher performance neighborhood.

Architecture Differences

The two cards come from different manufacturing nodes and foundries. The RTX 2060 uses the TU106 chip built on a 12 nm process at TSMC, with 10,800 million transistors on a 445 mm² die, giving a transistor density of 24.3 million per mm². The RX 580 uses the Polaris 20 chip on a 14 nm process at GlobalFoundries, with 5,700 million transistors on a 232 mm² die, a density of 24.6 million per mm². The densities are nearly identical, which is interesting: the RTX 2060 packs almost twice the transistors into roughly twice the die area.

The memory subsystems diverge sharply. The RTX 2060 ships with 6 GB of GDDR6 on a 192 bit bus, delivering 336.0 GB/s of bandwidth. The RX 580 has 8 GB of GDDR5 on a wider 256 bit bus, but only reaches 256.0 GB/s. The RTX 2060 wins on raw bandwidth despite the narrower bus, a direct result of GDDR6's higher effective data rate: 14 Gbps versus 8 Gbps.

Compute unit counts tell a nuanced story. The RX 580 has more shading units (2,304 versus 1,920), more texture mapping units (144 versus 120), but fewer ROPs (32 versus 48). The RTX 2060 compensates with higher clock speeds: a base of 1365 MHz and boost of 1680 MHz, compared to 1257 MHz base and 1340 MHz boost for the RX 580. The result is that the RTX 2060 reaches 6.451 TFLOPS FP32 versus 6.175 TFLOPS for the RX 580, a modest advantage in raw floating point throughput.

The architectural feature set is where the two diverge most fundamentally. The RTX 2060 is a Turing part with 30 dedicated RT cores and 240 tensor cores, hardware that the RX 580 completely lacks. The RX 580 has no ray tracing cores and no tensor cores at all. This is a generational divide: the RTX 2060 supports DirectX 12 Ultimate (12_2), while the RX 580 is limited to DirectX 12 (12_0). Vulkan support also differs, with the RTX 2060 at version 1.4 versus 1.3 for the RX 580. Both support OpenGL 4.6.

FP16 performance further separates them. The RTX 2060 delivers 12.90 TFLOPS FP16 at a 2:1 ratio to FP32, meaning it can accelerate half-precision workloads for the tensor cores and AI tasks. The RX 580 does FP16 at a 1:1 ratio at 6.175 TFLOPS, offering no advantage over FP32 performance. The RX 580's pixel rate is 42.88 GPixel/s, less than the RTX 80.64 GPixel/s of the RTX 2060. Texture rate is close: 201.6 GTexel/s for NVIDIA versus 193.0 GTexel/s for AMD, a small gap given the similar TMU counts.

Where Each One Wins

The RTX 2060 wins everywhere except Passmark G2D. That single AMD victory is a 3.1% margin in a 2D graphics test, which is likely related to driver behavior or memory bandwidth characteristics rather than a fundamental hardware superiority. The RX 2060 leads in every 3D test, every DirectX test except G2D, and every compute test except G2D. The RX 580 cannot claim a single 3D or compute win in the recorded data. Its 8 GB frame buffer is larger than the RTX 2060's 6 GB, but the data does not show that translates into victories in the tested workloads.

The RTX 2060's strengths align with modern and legacy APIs alike. It dominates DirectX 10 and DirectX 11 by huge margins, which suggests legacy game compatibility is a clear strength. The DirectX 11 gap of 83.3% is one of the largest in the entire comparison, which matters for the many titles that still rely on that API. The DirectX 10 gap of 113% is even larger, and the DirectX 9 gap of 53.2% rounds out a clean sweep of older APIs. The RTX 2060 also wins the newer DirectX 12 and Vulkan workloads, though by smaller margins, which indicates that the RX 580 is comparatively more competitive in modern APIs than in legacy ones.

Compute is a one-sided affair. The RTX 2060's OpenCL score is 73.6% higher, and its Vulkan score is 45.8% higher. The Passmark GPU Compute test shows a 57.3% lead. The presence of tensor cores in the RTX 2060 likely explains some of this, though the FP32 advantage is only 4.5%, so the compute lead is far larger than the raw FLOPs difference. The RX 580's GCN 4.0 architecture simply does not scale as well in these workloads.

The Verdict

The RTX 2060 is the clear winner in every meaningful benchmark category. Its average score of 15,290 places it in the 58th percentile of all GPUs, while the RX 580 sits at 12,928 in the 53rd percentile. The RX 580's nearest rival, the RTX 3050 Ti Mobile, matches it at a -0.1% delta, showing that the RX 580 is firmly in the modern entry-level performance class. The RTX 2060, by contrast, matches the RX 7600 within 0.8%, a card from a much later generation.

For a user prioritizing raw 3D performance, the RTX 2060 is the only choice from this data. It leads by 60.1% in Passmark G3D, 23.6% in 3DMark Steel Nomad DX12, and 23.3% in Passmark DirectX 12. The RX 580's 2D lead of 3.1% is negligible in real-world terms. For compute tasks, the RTX 2060 is again superior across every single test in the database. The RX 580 does have 8 GB of VRAM versus 6 GB, and that may matter in specific memory-heavy scenarios, but the recorded benchmarks do not show a workload where that translates into a win.

The RTX 2060 also supports ray tracing and tensor cores, features absent from the RX 580 entirely. The RTX 2060 is end-of-life, as is the RX 580, and both use a single 8-pin power connector with a 450 W suggested PSU. The RTX 2060 has a lower TDP of 160 W versus 185 W for the RX 580, which is a meaningful efficiency difference. The RTX 2060 also supports DirectX 12 Ultimate versus DirectX 12 (12_0) for the RX 580, and has a higher Vulkan version.

There is no dataset scenario where the RX 580 wins in the measured performance categories. The only reason to pick it would be the larger 8 GB memory allocation, but the benchmarks provided do not show that capacity advantage delivering any performance benefit. The data points to the RTX 2060 as the stronger GPU in every tested dimension except a single 2D test.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce RTX 2060 averages 15,290 across all recorded tests, while the AMD Radeon RX 580 averages 12,928.

Q: Does the RX 580 win any benchmark tests?

A: Yes, the RX 580 wins the Passmark G2D test with 769 versus 745 for the RTX 2060, a 3.1% margin. This is its only win out of 10 shared tests.

Q: How do the two cards compare in ray tracing support?

A: The RTX 2060 includes 30 RT cores and 240 tensor cores, while the RX 580 has no ray tracing cores and no tensor cores.

Q: What is the memory configuration difference?

A: The RTX 2060 has 6 GB of GDDR6 on a 192 bit bus with 336.0 GB/s bandwidth. The RX 580 has 8 GB of GDDR5 on a 256 bit bus with 256.0 GB/s bandwidth.

Q: Which card has higher power consumption?

A: The RX 580 has a TDP of 185 W, while the RTX 2060 has a TDP of 160 W. Both use a single 8-pin power connector and a 450 W suggested PSU.

Q: What are the DirectX support levels?

A: The RTX 2060 supports DirectX 12 Ultimate (12_2), while the RX 580 supports DirectX 12 (12_0). Vulkan support is 1.4 for the RTX 2060 and 1.3 for the RX 580.

Specification Differences

| Specification | NVIDIA GeForce RTX 2060 | AMD Radeon RX 580 |

|---|---|---|

| Process node | 12 nm | 14 nm |

| Foundry | TSMC | GlobalFoundries |

| Transistors | 10,800 million | 5,700 million |

| Die size | 445 mm² | 232 mm² |

| Transistor density | 24.3M / mm² | 24.6M / mm² |

| Base clock | 1365 MHz | 1257 MHz |

| Boost clock | 1680 MHz | 1340 MHz |

| Memory clock | 14 Gbps effective | 8 Gbps effective |

| Memory size | 6 GB | 8 GB |

| Memory type | GDDR6 | GDDR5 |

| Memory bus width | 192 bit | 256 bit |

| Memory bandwidth | 336.0 GB/s | 256.0 GB/s |

| Shading units | 1920 | 2304 |

| TMUs | 120 | 144 |

| ROPs | 48 | 32 |

| RT cores | 30 | None |

| Tensor cores | 240 | None |

| Pixel rate | 80.64 GPixel/s | 42.88 GPixel/s |

| Texture rate | 201.6 GTexel/s | 193.0 GTexel/s |

| FP32 performance | 6.451 TFLOPS | 6.175 TFLOPS |

| FP16 performance | 12.90 TFLOPS (2:1) | 6.175 TFLOPS (1:1) |

| TDP | 160 W | 185 W |

| Display outputs | 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, 1x USB Type-C | 1x HDMI 2.0b, 3x DisplayPort 1.4a |

| DirectX | 12 Ultimate (12_2) | 12 (12_0) |

| Vulkan | 1.4 | 1.3 |

| Length | 229 mm (9 inches) | 241 mm (9.5 inches) |

| Height | 113 mm (4.4 inches) | Not specified |

| Width | 35 mm (1.4 inches) | Not specified |

| Release date | 2019-01-06 | 2017-04-17 |

| Predecessor | GeForce 10 | Arctic Islands |

| Successor | GeForce 30 | Vega |

| Launch MSRP | 349 USD | 229 USD |