AMD Radeon Pro 570X vs NVIDIA GeForce RTX 3090 Comparison

Head-to-Head Benchmarks

The database records two direct benchmark comparisons between the AMD Radeon Pro 570X and the NVIDIA GeForce RTX 3090, and the results are not close. The NVIDIA card wins both recorded tests, with the largest gap appearing in the Geekbench OpenCL workload. There, the RTX 3090 scores 172,758 points against the Radeon Pro 570X’s 27,604 points, a delta of -84% from the AMD card’s perspective. That is not a marginal difference; it is a multiple of more than six times the AMD score. In practical terms, the OpenCL result reflects the RTX 3090’s massive advantage in raw compute throughput, which the specification data later confirms.

The second head-to-head test, Geekbench Vulkan, narrows the gap somewhat but still favors NVIDIA decisively. The RTX 3090 scores 53,927 points, while the Radeon Pro 570X scores 28,859 points, a delta of -46.5%. Even in a lower-level graphics API like Vulkan, where the AMD part’s GCN architecture might be expected to show relative strength, the NVIDIA card still nearly doubles the AMD score. The data shows no recorded test where the AMD Radeon Pro 570X comes out ahead; the win count stands at 0 for the AMD card and 2 for the NVIDIA card.

Looking at broader benchmark averages, the picture becomes more nuanced. The AMD Radeon Pro 570X has an average benchmark score of 32,176 across all recorded tests, which places it at the 76th percentile of all GPUs in the database. Its nearest rivals in that metric are the AMD FirePro S10000 at 32,388 (-0.7% delta), the AMD Radeon RX 7900 GRE at 32,456 (-0.9% delta), the Intel Arc Pro A30M at 31,894 (+0.9% delta), and the AMD FirePro S9300 X2 at 32,540 (-1.1% delta). These are tightly clustered scores, meaning the Radeon Pro 570X sits in a competitive band where no single rival is more than about 1% away in either direction.

The NVIDIA GeForce RTX 3090, by contrast, has an average benchmark score of 27,565, which places it at the 73rd percentile of all GPUs. Its nearest rivals are the NVIDIA GeForce RTX 4070 Mobile at 27,435 (+0.5% delta), the AMD Radeon RX 6700 XT at 27,425 (+0.5% delta), the AMD Radeon Pro Vega 20 at 27,839 (-1% delta), and the AMD Radeon RX 7800M at 27,883 (-1.1% delta). Interestingly, despite winning the head-to-head tests by large margins, the RTX 3090’s average score is lower than the AMD Radeon Pro 570X’s average. This occurs because the average benchmarks include a wider mix of tests, including older DirectX and PassMark workloads where the RTX 3090’s strengths do not translate into proportionally high scores. The PassMark DirectX 9 score for the RTX 3090 is 268, the DirectX 10 score is 182, the DirectX 11 score is 220, and the DirectX 12 score is 110, all relatively modest numbers compared to its Geekbench OpenCL result. The RTX 3090 also records a PassMark G3D score of 26,645 and a PassMark GPU compute score of 15,356.

Architecture Differences

The two GPUs come from fundamentally different architectural generations. The AMD Radeon Pro 570X uses the Ellesmere chip built on GCN 4.0, fabricated on a 14 nm process at GlobalFoundries. The die measures 232 mm² and contains 5,700 million transistors, yielding a transistor density of 24.6 million per square millimeter. The NVIDIA GeForce RTX 3090 uses the GA102 chip built on the Ampere architecture, fabricated on an 8 nm process at Samsung. Its die is far larger at 628 mm², with 28,300 million transistors, giving a transistor density of 45.1 million per square millimeter. The NVIDIA chip packs nearly five times the transistor count into under three times the die area, reflecting both the newer process node and the architectural complexity of Ampere.

The shading resources differ by an order of magnitude. The Radeon Pro 570X has 1,792 shading units, 112 texture mapping units, and 32 render output units. The RTX 3090 has 10,496 shading units, 328 TMUs, and 112 ROPs. The RTX 3090 also includes 82 dedicated ray tracing cores and 328 tensor cores, features entirely absent from the AMD part, which has no RT cores and no tensor cores. This explains the RTX 3090’s support for DirectX 12 Ultimate (12_2), while the Radeon Pro 570X is limited to DirectX 12 (12_0). The NVIDIA card also supports Vulkan 1.4, whereas the AMD part supports Vulkan 1.3; both support OpenGL 4.6.

Clock speeds tell a similar story. The Radeon Pro 570X runs at a base clock of 1000 MHz and boosts to 1105 MHz, with memory clocked at 1695 MHz (6.8 Gbps effective). The RTX 3090 has a base clock of 1395 MHz and a boost of 1695 MHz, with memory at 1219 MHz (19.5 Gbps effective). The memory subsystems are also vastly different: the AMD card has 4 GB of GDDR5 on a 256-bit bus, delivering 217.0 GB/s bandwidth, while the NVIDIA card has 24 GB of GDDR6X on a 384-bit bus, delivering 936.2 GB/s bandwidth. That is more than four times the memory capacity and more than four times the bandwidth.

The resulting throughput figures reinforce the architectural gap. The Radeon Pro 570X achieves a pixel rate of 35.36 GPixel/s and a texture rate of 123.8 GTexel/s, with FP32 and FP16 both rated at 3.960 TFLOPS (1:1 ratio). The RTX 3090 achieves 189.8 GPixel/s and 556.0 GTexel/s, with FP32 and FP16 both rated at 35.58 TFLOPS (1:1 ratio). The NVIDIA card is roughly 5.4 times faster in pixel rate, 4.5 times faster in texture rate, and exactly 9 times faster in FP32 compute.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon Pro 570X has an average benchmark score of 32,176, which is higher than the NVIDIA GeForce RTX 3090’s average of 27,565. However, the RTX 3090 wins both head-to-head tests recorded in the database.

Q: What is the largest recorded performance gap between the two cards?

A: The largest gap is in Geekbench OpenCL, where the RTX 3090 scores 172,758 versus the Radeon Pro 570X’s 27,604, a delta of -84% from the AMD card’s perspective.

Q: Does the AMD card support ray tracing or tensor cores?

A: No. The Radeon Pro 570X has no ray tracing cores and no tensor cores. The RTX 3090 has 82 RT cores and 328 tensor cores.

Q: How do the memory specifications compare?

A: The Radeon Pro 570X has 4 GB of GDDR5 on a 256-bit bus with 217.0 GB/s bandwidth. The RTX 3090 has 24 GB of GDDR6X on a 384-bit bus with 936.2 GB/s bandwidth.

Q: What DirectX versions do the two cards support?

A: The Radeon Pro 570X supports DirectX 12 (12_0), while the RTX 3090 supports DirectX 12 Ultimate (12_2), which includes additional features such as hardware ray tracing.

Q: Which card has a higher pixel rate?

A: The RTX 3090 has a pixel rate of 189.8 GPixel/s, compared to 35.36 GPixel/s for the Radeon Pro 570X.

Specification Differences

The two cards differ across nearly every core specification in the database. The process node is 14 nm for the AMD part versus 8 nm for the NVIDIA part. The foundries differ: GlobalFoundries for AMD, Samsung for NVIDIA. Transistor counts are 5,700 million for AMD versus 28,300 million for NVIDIA. Die sizes are 232 mm² for AMD versus 628 mm² for NVIDIA. Transistor density is 24.6 million per mm² for AMD versus 45.1 million per mm² for NVIDIA.

Base clocks are 1000 MHz for the AMD card versus 1395 MHz for the NVIDIA card. Boost clocks are 1105 MHz versus 1695 MHz. Memory clocks are 1695 MHz (6.8 Gbps effective) versus 1219 MHz (19.5 Gbps effective). Memory capacity is 4 GB versus 24 GB. Memory type is GDDR5 versus GDDR6X. Bus width is 256-bit versus 384-bit. Memory bandwidth is 217.0 GB/s versus 936.2 GB/s.

Shading units are 1,792 versus 10,496. TMUs are 112 versus 328. ROPs are 32 versus 112. RT cores are absent on the AMD card versus 82 on the NVIDIA card. Tensor cores are absent versus 328. Pixel rates are 35.36 GPixel/s versus 189.8 GPixel/s. Texture rates are 123.8 GTexel/s versus 556.0 GTexel/s. FP32 is 3.960 TFLOPS versus 35.58 TFLOPS. FP16 is 3.960 TFLOPS versus 35.58 TFLOPS.

TDP is 150 W for the AMD card versus 350 W for the NVIDIA card. The AMD card is an IGP with no power connectors, while the NVIDIA card is triple-slot with a 1x 12-pin connector and a suggested PSU of 750 W. The bus interface is PCIe 3.0 x16 for AMD versus PCIe 4.0 x16 for NVIDIA. Display outputs are portable device dependent for AMD versus 1x HDMI 2.1 and 3x DisplayPort 1.4a for NVIDIA. The NVIDIA card has physical dimensions of 336 mm length, 140 mm height, and 61 mm width; the AMD card has no recorded dimensions. Release dates are March 2019 for the AMD card versus August 2020 for the NVIDIA card. The NVIDIA card has a predecessor (GeForce 20) and a successor (GeForce 40); the AMD card has neither. The NVIDIA card has a recorded launch MSRP of 1,499 USD; the AMD card has none. Both are end-of-life products.

The Verdict

The data points to two very different products serving different roles. The AMD Radeon Pro 570X is a low-power, integrated-class GPU with modest compute resources, a 150 W TDP, and no external power connectors. Its average benchmark score of 32,176 is respectable for its class, placing it near peers like the FirePro S10000 and RX 7900 GRE within about 1% in either direction. It is a 76th percentile GPU overall, meaning it outperforms roughly three-quarters of all recorded GPUs in the database, despite its limited memory and older architecture.

The NVIDIA GeForce RTX 3090 is a high-end discrete card in the GeForce 30-series, with a 350 W TDP, triple-slot cooler, and a 1x 12-pin power connector. It wins both head-to-head benchmark tests decisively, with an 84% lead in OpenCL and a 46.5% lead in Vulkan. Its raw specifications are overwhelmingly superior: 24 GB of GDDR6X memory, 936.2 GB/s bandwidth, 35.58 TFLOPS FP32, and hardware ray tracing and tensor cores. However, its average benchmark score of 27,565 is lower than the AMD card’s average, and it sits at the 73rd percentile, slightly below the AMD card’s 76th percentile. This discrepancy stems from the RTX 3090’s relatively low scores in older DirectX and PassMark tests, where its architectural strengths do not translate into proportionally high results.

For a user whose workload aligns with modern, compute-heavy APIs like OpenCL or Vulkan, the RTX 3090 is the clear choice based on the recorded benchmarks. It delivers more than six times the OpenCL performance and nearly double the Vulkan performance of the Radeon Pro 570X. For a user whose workload involves legacy DirectX titles or whose system constraints demand low power draw and no additional power connectors, the Radeon Pro 570X may be the more appropriate fit, as its higher average score and higher percentile suggest it handles a broader range of older benchmarks more consistently. The RTX 3090 also requires a 750 W suggested PSU and a triple-slot chassis, while the Radeon Pro 570X is an IGP with no power connectors, making it suitable for compact or integrated systems. The choice ultimately depends on whether the priority is maximum performance in modern workloads or compatibility with constrained hardware environments.