AMD Radeon RX 570 vs AMD Radeon RX 590 Comparison

The AMD Radeon RX 590 and AMD Radeon RX 570 are both Polaris-generation cards, but the data shows a clear separation between them. The RX 590 wins both head-to-head benchmarks, with its largest advantage coming in the Geekbench Metal test, where it scores 59,832 against the RX 570’s 44,915 — a 33.2% delta. In the 3DMark Steel Nomad DX12 test, the RX 590 posts 1,081 points versus 880 for the RX 570, a 22.8% lead. Despite this, the average benchmark scores are remarkably close: the RX 590 averages 30,457 while the RX 570 averages 30,158, a difference of only 1% according to the nearestRivals deltaPct field. This suggests that while the RX 590 dominates in specific workloads, the broader benchmark picture is tighter than the head-to-head results imply.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 test is a modern DirectX 12 workload, and here the RX 590’s advantage is substantial. The RX 590 scores 1,081, while the RX 570 scores 880, yielding a 22.8% delta in the RX 590’s favor. This is not a marginal win; it reflects a meaningful gap in raw compute throughput under DX12 conditions. The RX 590’s shading units (2,304 versus 2,048) and higher clock speeds contribute to this outcome. The base clock of the RX 590 is 1469 MHz, with a boost of 1545 MHz, whereas the RX 570 operates at a 1168 MHz base and 1244 MHz boost. In a test that scales with shader count and frequency, the RX 590’s additional 256 shading units and roughly 25-30% higher clocks are the likely drivers of its 22.8% win.

The Geekbench Metal benchmark shows an even larger gap. The RX 590 achieves 59,832, while the RX 570 reaches 44,915 — a 33.2% delta. Metal is Apple’s graphics API, and while neither card is designed for macOS gaming, the test measures general compute and graphics throughput. The RX 590’s FP32 performance is listed at 7.119 TFLOPS, compared to the RX 570’s 5.095 TFLOPS, a raw compute advantage of about 39.7%. The observed 33.2% benchmark lead is consistent with that theoretical difference, though slightly lower, suggesting some overhead or memory bandwidth limitations. The RX 590’s memory bandwidth is 256.0 GB/s versus 224.0 GB/s for the RX 570, which helps but does not fully close the gap.

Looking at the nearest rivals, the RX 590’s average score of 30,457 places it nearly identical to the NVIDIA GeForce RTX 2080 Ti, which averages 30,459 (deltaPct 0%). That is a striking result for a Polaris card. The RX 590 also sits just 1% below the AMD Radeon RX 590 GME (30,755) and 1.2% above the NVIDIA GeForce RTX 5060 (30,109). The RX 570, meanwhile, averages 30,158, which puts it 0.2% above both the RTX 5060 and the AMD Radeon RX 7700 XT, and 0.6% above the NVIDIA Tesla M60. The RX 570 holds its own against much newer cards in average scores, but the head-to-head shows the RX 590’s specific benchmark wins are decisive.

Where Each One Wins

The RX 590 wins every head-to-head benchmark in the data, so the “where each one wins” section is lopsided. In 3DMark Steel Nomad DX12, the RX 590’s 22.8% lead indicates that it is the stronger choice for modern DX12 gaming workloads. The higher texture rate (222.5 GTexel/s versus 159.2 GTexel/s) and pixel rate (49.44 GPixel/s versus 39.81 GPixel/s) support this, as these directly affect fill-rate-bound scenarios. For users running DX12 titles at high resolutions, the RX 590’s advantage should translate to smoother frame pacing.

In Geekbench Metal, the RX 590’s 33.2% lead is even more pronounced. This suggests that for compute-heavy tasks, such as content creation or GPU-accelerated rendering that leverages Metal, the RX 590 is significantly more capable. The FP32 throughput of 7.119 TFLOPS versus 5.095 TFLOPS is the core reason. The RX 570 does not win any benchmark in the provided data, but its average benchmark score of 30,158 is close to the RX 590’s 30,457. This means that in a broader suite of applications — beyond the two head-to-head tests — the RX 570 is not far behind. The RX 570 also has a lower TDP of 150 W versus 175 W, which could make it a more thermally manageable option in constrained cases, though the data does not include temperature or power consumption test results.

For users prioritizing the two specific tests listed, the RX 590 is the clear winner. For users looking at a wider range of workloads, the RX 570’s 1% lower average score suggests that the performance gap narrows considerably outside of the Steel Nomad and Metal tests. The RX 570’s 4 GB memory capacity may be a limiting factor in modern games that exceed that amount, but the RX 590’s 8 GB capacity gives it more headroom for texture-heavy content.

Architecture Differences

Both cards use the GCN 4.0 architecture, but they are built on different process nodes. The RX 590 uses the Polaris 30 chip on a 12 nm process, while the RX 570 uses the Polaris 20 chip on a 14 nm process. Both are fabricated by GlobalFoundries. Despite the node difference, both chips have the same transistor count (5,700 million) and the same die size (232 mm²), resulting in identical transistor density of 24.6M/mm². The 12 nm node on the RX 590 allows for higher clock speeds at a modest TDP increase, which is the primary architectural benefit.

The RX 590 has 2,304 shading units, 144 texture mapping units (TMUs), and 32 render output units (ROPs). The RX 570 has 2,048 shading units, 128 TMUs, and the same 32 ROPs. This means the RX 590 has 12.5% more shading units and TMUs, but the same ROP count, so pixel fill-rate differences are driven entirely by clock speed. The RX 590’s pixel rate is 49.44 GPixel/s versus 39.81 GPixel/s for the RX 570, a 24.2% difference. The texture rate difference is larger: 222.5 GTexel/s versus 159.2 GTexel/s, a 39.8% gap, which mirrors the FP32 difference closely.

Memory configurations differ significantly. The RX 590 has 8 GB of GDDR5 memory on a 256-bit bus, with a memory clock of 2000 MHz (8 Gbps effective) and bandwidth of 256.0 GB/s. The RX 570 has 4 GB of GDDR5 on the same 256-bit bus, but a lower memory clock of 1750 MHz (7 Gbps effective), yielding 224.0 GB/s. The RX 590 thus has both more capacity and higher bandwidth. Both cards support DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. Neither has ray tracing or tensor cores. The RX 590 requires a 1x 8-pin power connector, while the RX 570 requires only a 1x 6-pin, reflecting the RX 590’s higher 175 W TDP versus 150 W. Both are dual-slot cards with the same 241 mm length and PCIe 3.0 x16 interface. Display outputs differ only in that the RX 570 adds a DVI port alongside the HDMI 2.0b and three DisplayPort 1.4a outputs found on both.

FAQ

Q: How much faster is the RX 590 in the 3DMark Steel Nomad DX12 test?

A: The RX 590 scores 1,081, which is 22.8% higher than the RX 570’s 880.

Q: What explains the RX 590’s larger lead in Geekbench Metal?

A: The RX 590’s FP32 throughput is 7.119 TFLOPS versus 5.095 TFLOPS for the RX 570, and the benchmark shows a 33.2% delta in favor of the RX 590.

Q: Do the average benchmark scores reflect the head-to-head dominance?

A: No. The RX 590 averages 30,457 and the RX 570 averages 30,158, a difference of only about 1% according to the deltaPct field, meaning the gap narrows outside the two specific head-to-head tests.

Q: Are these cards on the same manufacturing process?

A: No. The RX 590 uses a 12 nm process (Polaris 30), while the RX 570 uses a 14 nm process (Polaris 20), though both are made by GlobalFoundries with identical transistor counts and die sizes.

Q: What is the memory capacity difference?

A: The RX 590 has 8 GB of GDDR5, while the RX 570 has 4 GB. Both use a 256-bit bus, but the RX 590’s memory runs at 2000 MHz (8 Gbps effective) versus 1750 MHz (7 Gbps effective) on the RX 570.

Q: Do both cards support the same graphics APIs?

A: Yes, both support DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3, with no ray tracing or tensor cores on either.

The Verdict

The data points unequivocally to the RX 590 as the stronger performer in the two benchmarks provided. A 22.8% win in 3DMark Steel Nomad DX12 and a 33.2% win in Geekbench Metal are decisive margins. The RX 590’s higher shading unit count, faster clocks, and larger memory bandwidth make it the better card for anyone prioritizing these specific workloads. The 8 GB memory capacity also future-proofs it better than the RX 570’s 4 GB.

However, the average benchmark scores tell a different story. With the RX 590 averaging 30,457 and the RX 570 at 30,158, the delta is only 1%. This suggests that in a broad range of applications, the RX 570 is not far behind, and its lower TDP (150 W versus 175 W) and simpler power connector (1x 6-pin versus 1x 8-pin) make it a more accommodating card for older power supplies. The RX 570’s launch MSRP was 169 USD, while the RX 590’s was 279 USD, so the price gap was substantial at launch, though both are now end-of-life products.

For users who play DX12 titles or use Metal-based applications, the RX 590 is the rational choice based on its benchmark wins. For users with a wider software mix or tighter power constraints, the RX 570’s near-identical average score makes it a surprisingly competitive option. The RX 590’s 74th percentile ranking among all GPUs is matched by the RX 570, so neither card is a standout in the broader market. Ultimately, the RX 590 wins on raw performance in the tests shown, while the RX 570 wins on efficiency and cost — but the data does not support a recommendation for the RX 570 over the RX 590 in the two benchmarks where they are directly compared.

Specification Differences

| Specification | AMD Radeon RX 590 | AMD Radeon RX 570 |

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

| Chip | Polaris 30 | Polaris 20 |

| Process Node | 12 nm | 14 nm |

| Base Clock | 1469 MHz | 1168 MHz |

| Boost Clock | 1545 MHz | 1244 MHz |

| Memory Clock | 2000 MHz (8 Gbps effective) | 1750 MHz (7 Gbps effective) |

| Memory Size | 8 GB | 4 GB |

| Memory Bandwidth | 256.0 GB/s | 224.0 GB/s |

| Shading Units | 2304 | 2048 |

| TMUs | 144 | 128 |

| Pixel Rate | 49.44 GPixel/s | 39.81 GPixel/s |

| Texture Rate | 222.5 GTexel/s | 159.2 GTexel/s |

| FP32 | 7.119 TFLOPS | 5.095 TFLOPS |

| FP16 | 7.119 TFLOPS (1:1) | 5.095 TFLOPS (1:1) |

| TDP | 175 W | 150 W |

| Power Connectors | 1x 8-pin | 1x 6-pin |

| Display Outputs | 1x HDMI 2.0b, 3x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0b, 3x DisplayPort 1.4a |

| Release Date | 2018-11-14 | 2017-04-17 |

| Launch MSRP | 279 USD | 169 USD |