AMD Radeon RX 590 vs AMD Radeon RX 9070 Comparison

The AMD Radeon RX 590 and AMD Radeon RX 9070 represent two vastly different eras of AMD graphics technology, separated by more than six years of architectural evolution. The RX 590, built on the mature Polaris 30 chip with GCN 4.0, was a late-cycle refresh of the RX 500 series aimed at 1080p gaming. The RX 9070, however, is a current-generation RDNA 4.0 part from the Navi 48 chip, targeting high-refresh and high-resolution workloads. The data in the FACT PACK reveals a generational chasm in raw performance, feature support, and efficiency, yet the benchmark averages tell a surprisingly nuanced story when looking at percentile rankings and nearest rivals.

FAQ

Q: How do the two cards compare in the 3DMark Steel Nomad DX12 test?

A: The RX 9070 scores 6290, while the RX 590 scores 1081. The deltaPct of -82.8 indicates the RX 590 trails by 82.8% relative to the RX 9070, meaning the RX 9070 delivers roughly 5.8 times the frame rate in this specific DX12 workload.

Q: What is the average benchmark score for each card, and what does that imply?

A: The RX 590 has an average benchmark score of 24744, while the RX 9070 averages 23877. Despite the RX 9070’s massive win in the single head-to-head test, its overall average is actually 3.5% lower than the RX 590’s, suggesting the RX 9070’s average is pulled down by a broader mix of tests where it may not excel.

Q: Which card has a higher percentile ranking among all GPUs?

A: The RX 590 sits at the 70th percentile, while the RX 9070 is at the 69th percentile. This is a marginal difference, with the RX 590 barely ahead, indicating that when considering the entire GPU landscape, the older card’s average performance profile is comparable in rank to the newer one.

Q: What are the nearest rivals for each card based on average score?

A: For the RX 590, the nearest rivals are the NVIDIA RTX A5000 Mobile (avg 24763, deltaPct -0.1), Intel Arc A350M (avg 24647, deltaPct 0.4), AMD Radeon RX 6600 XT (avg 24442, deltaPct 1.2), and NVIDIA GeForce GTX 1630 (avg 24277, deltaPct 1.9). For the RX 9070, they are the AMD Radeon RX 6800S (avg 24063, deltaPct -0.8), NVIDIA GeForce RTX 2080 SUPER (avg 24170, deltaPct -1.2), NVIDIA GeForce GTX TITAN Z (avg 23736, deltaPct 0.6), and NVIDIA GeForce RTX 3080 Mobile (avg 23628, deltaPct 1.1).

Q: What are the launch MSRPs of the two cards?

A: The RX 590 had a launch MSRP of 279 USD, while the RX 9070 has a launch MSRP of 549 USD.

Q: Which card supports ray tracing, and which does not?

A: The RX 9070 includes 56 RT cores, enabling hardware-accelerated ray tracing. The RX 590 has no RT cores listed, meaning it lacks dedicated ray tracing hardware.

Architecture Differences

The architectural gap between these two GPUs is stark, starting with the manufacturing process. The RX 590 uses a 12 nm node at GlobalFoundries, while the RX 9070 is built on a 4 nm TSMC process. This process shrink allows the RX 9070 to pack 53,900 million transistors into a 357 mm² die, yielding a transistor density of 151.0M per mm². The RX 590, by contrast, has just 5,700 million transistors on a 232 mm² die, for a density of 24.6M per mm². That is a 6.1x increase in transistor density for the newer card.

The core architectures are fundamentally different: the RX 590 runs GCN 4.0, while the RX 9070 is RDNA 4.0. This shift is reflected in the compute resources. The RX 590 has 2304 shading units, 144 TMUs, and 32 ROPs. The RX 9070 scales this up to 3584 shading units, 224 TMUs, and 128 ROPs. More critically, the RX 9070 adds 56 RT cores, a feature entirely absent from the RX 590. The RX 9070 also supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, whereas the RX 590 is limited to DirectX 12 (12_0) and Vulkan 1.3. The memory subsystem sees a similar leap: the RX 590 uses GDDR5 with 8 GB and 256.0 GB/s bandwidth, while the RX 9070 uses GDDR6 with 16 GB and 644.6 GB/s bandwidth. The RX 9070 also moves to PCIe 5.0 x16, up from PCIe 3.0 x16 on the RX 590.

Clock behavior also diverges. The RX 590 has a base clock of 1469 MHz and a boost of 1545 MHz. The RX 9070 has a lower base of 1330 MHz but a much higher boost of 2520 MHz, plus a game clock of 2070 MHz. The FP32 throughput tells the performance story: the RX 590 delivers 7.119 TFLOPS, while the RX 9070 reaches 36.13 TFLOPS, a 5.1x increase. Pixel rate jumps from 49.44 GPixel/s to 322.6 GPixel/s, and texture rate from 222.5 GTexel/s to 564.5 GTexel/s.

Head-to-Head Benchmarks

The only direct head-to-head benchmark in the data is the 3dmark_3dmark_steel_nomad_dx12 test, and it is a decisive victory for the RX 9070. The RX 9070 scores 6290, while the RX 590 scores 1081. The deltaPct of -82.8 means the RX 590’s score is 82.8% lower than the RX 9070’s. This indicates that in a modern DX12 workload, the RX 9070 is over five times faster, a margin that reflects not just more compute units but also architectural efficiency and higher memory bandwidth.

However, the story is less one-sided when looking at the average benchmark scores. The RX 590’s average is 24744, while the RX 9070’s is 23877. This means that across a wider set of benchmarks, the RX 590 actually holds a 3.6% advantage in average score. The RX 9070’s individual benchmark results include a Passmark G3D score of 25381, which is higher than the RX 590’s average, but also a Passmark DirectX 12 score of just 74, which is exceptionally low. The RX 9070 also has a Passmark DirectX 10 score of 141 and a Passmark DirectX 9 score of 343, suggesting uneven performance across legacy APIs. The RX 590’s benchmarks are limited to a 3DMark Steel Nomad score of 1081 and a Geekbench Metal score of 48407, but the latter is a strong showing that likely props up its average.

This divergence between the single modern test and the broader average suggests the RX 9070 is heavily optimized for current-generation workloads, while the RX 590’s GCN architecture still holds its own in certain compute or legacy scenarios. The data shows no test where the RX 590 wins head-to-head against the RX 9070, but the average score discrepancy implies the RX 9070’s benchmark suite includes tests that penalize its newer architecture.

Specification Differences

The two cards differ across nearly every specification field. The RX 590 uses a 12 nm process from GlobalFoundries, while the RX 9070 uses a 4 nm process from TSMC. Transistor count is 5,700 million versus 53,900 million, and die size is 232 mm² versus 357 mm². The RX 590 has a base clock of 1469 MHz and a boost of 1545 MHz, with no game clock; the RX 9070 has a base of 1330 MHz, a boost of 2520 MHz, and a game clock of 2070 MHz. Memory speed is 2000 MHz (8 Gbps effective) on the RX 590 versus 2518 MHz (20.1 Gbps effective) on the RX 9070.

Memory capacity and type differ: 8 GB GDDR5 versus 16 GB GDDR6. Bandwidth is 256.0 GB/s versus 644.6 GB/s. Shading units are 2304 versus 3584, TMUs are 144 versus 224, and ROPs are 32 versus 128. The RX 9070 has 56 RT cores; the RX 590 has none. Pixel rate is 49.44 GPixel/s versus 322.6 GPixel/s. Texture rate is 222.5 GTexel/s versus 564.5 GTexel/s. FP32 is 7.119 TFLOPS versus 36.13 TFLOPS. TDP is 175 W versus 220 W. Power connectors are 1x 8-pin versus 2x 8-pin. Suggested PSU is 450 W versus 550 W. Bus interface is PCIe 3.0 x16 versus PCIe 5.0 x16. Display outputs are 1x HDMI 2.0b and 3x DisplayPort 1.4a versus 1x HDMI 2.1b and 3x DisplayPort 2.1a. DirectX support is 12 (12_0) versus 12 Ultimate (12_2). Vulkan support is 1.3 versus 1.4. The RX 590 is 241 mm long, while the RX 9070’s dimensions are not listed. Production status is end-of-life versus active. Release dates are 2018-11-14 versus 2025-03-05.

Where Each One Wins

The RX 9070 wins decisively in the only direct head-to-head benchmark, the 3DMark Steel Nomad DX12 test, with a score of 6290 versus 1081. This makes it the clear choice for modern DX12 gaming and any workload that leverages the newer API features like ray tracing, given its 56 RT cores. The RX 9070 also wins on raw specifications: it has more shading units, TMUs, ROPs, higher memory bandwidth, and over 5x the FP32 throughput. For any user running contemporary games or compute applications that scale with these resources, the RX 9070 is the superior performer.

The RX 590, however, holds a higher average benchmark score of 24744 versus 23877, and it sits at the 70th percentile versus the RX 9070’s 69th. This suggests that in a mixed workload of older DX10, DX11, and compute tests, the RX 590’s GCN architecture can be more consistent. The RX 590 also has a lower TDP of 175 W versus 220 W, and requires a less powerful PSU (450 W versus 550 W), making it more manageable in smaller builds. Its Geekbench Metal score of 48407 is a particular strength, indicating solid performance in Apple’s Metal API, which could be relevant for certain cross-platform compute tasks.

The Verdict

For anyone building a system for modern gaming or professional 3D rendering, the data overwhelmingly favors the RX 9070. Its 6290 score in the Steel Nomad DX12 test versus the RX 590’s 1081 is a 82.8% differential in the RX 9070’s favor, and it offers 16 GB of memory, ray tracing, and DirectX 12 Ultimate support. The RX 9070 is the only sensible choice for future-proofing, despite its higher TDP and power connector requirements.

The RX 590, despite being end-of-life, is not without merit based on the data. Its higher average benchmark score and 70th percentile ranking indicate that for legacy applications, older DX11 titles, or compute workloads that don’t utilize modern features, it can still perform competitively. The lower power draw and smaller physical footprint (241 mm versus unknown for the RX 9070) may appeal to users with constrained builds. However, with zero wins in the head-to-head benchmark and far lower memory bandwidth, the RX 590’s advantages are narrow and specific to older software. The verdict is clear: the RX 9070 is the performance leader for current and future workloads, while the RX 590’s data-driven strengths lie only in niche legacy scenarios and lower system requirements.