AMD Radeon Pro 5700 vs NVIDIA GeForce RTX 2060 SUPER Comparison

The AMD Radeon Pro 5700 and NVIDIA GeForce RTX 2060 SUPER are two end-of-life graphics cards from different eras and design philosophies, yet they land in nearly the same performance tier. The aggregate benchmark data shows the Radeon Pro 5700 with an average score of 18,189, while the RTX 2060 SUPER averages 18,093, a razor-thin 0.5% difference that places them as direct rivals. However, the composition of those scores reveals a stark divergence: the NVIDIA card wins all nine head-to-head benchmark comparisons, often by wide margins, while the AMD card’s overall average is buoyed by a single strong result in one test. This creates an unusual dynamic where the two cards are statistically inseparable in overall terms, but behave very differently across individual workloads.

FAQ

Q: Which card has the higher aggregate benchmark score?

A: The AMD Radeon Pro 5700 has a slightly higher average benchmark score of 18,189 compared to the NVIDIA GeForce RTX 2060 SUPER’s 18,093, a difference of 0.5% in AMD’s favor.

Q: How many head-to-head benchmark comparisons does each card win?

A: The NVIDIA GeForce RTX 2060 SUPER wins all 9 head-to-head comparisons listed. The AMD Radeon Pro 5700 wins zero of those comparisons.

Q: What is the most significant performance gap between the two cards?

A: The largest delta is in PassMark DirectX 11, where the RTX 2060 SUPER scores 130 versus the Radeon Pro 5700’s 78, a 40% advantage for NVIDIA.

Q: Do both cards support the same DirectX version?

A: No. The AMD card supports DirectX 12 (12_1), while the NVIDIA card supports DirectX 12 Ultimate (12_2), which is a higher feature level.

Q: Which card has dedicated ray tracing hardware?

A: Only the NVIDIA GeForce RTX 2060 SUPER includes 34 RT cores. The AMD Radeon Pro 5700 has no RT cores listed.

Q: Are both cards the same physical size?

A: No. The NVIDIA card is a dual-slot design measuring 229 mm in length, 113 mm in height, and 35 mm in width. The AMD card is listed as an IGP (integrated graphics processor) with no dimensions or display outputs.

Architecture Differences

The two cards are built on fundamentally different architectures and process nodes. The AMD Radeon Pro 5700 uses the Navi 10 chip based on RDNA 1.0 architecture, fabricated on a 7 nm process at TSMC. In contrast, the NVIDIA GeForce RTX 2060 SUPER uses the TU106 chip based on the older Turing architecture, fabricated on a 12 nm process, also at TSMC. This process difference is significant: AMD’s 7 nm node allows for a die size of 251 mm² containing 10,300 million transistors, yielding a density of 41.0 million transistors per mm². NVIDIA’s 12 nm node requires a much larger 445 mm² die to house 10,800 million transistors, resulting in a lower density of 24.3 million transistors per mm².

The architectural feature sets diverge sharply. The NVIDIA card includes 34 RT cores and 272 tensor cores, enabling hardware-accelerated ray tracing and AI-assisted features, while the AMD card lists no such dedicated units. This is reflected in their DirectX support: the RTX 2060 SUPER achieves DirectX 12 Ultimate (12_2), whereas the Radeon Pro 5700 only reaches DirectX 12 (12_1). Both cards support OpenGL 4.6 and Vulkan 1.4.

Memory configuration is similar in capacity and type—both have 8 GB of GDDR6 on a 256-bit bus—but the implementation differs. The AMD card runs memory at 1500 MHz (12 Gbps effective) producing 384.0 GB/s of bandwidth, while the NVIDIA card runs at 1750 MHz (14 Gbps effective) for 448.0 GB/s. The bus interface also differs: AMD uses PCIe 4.0 x16, while NVIDIA uses PCIe 3.0 x16.

Where Each One Wins

The benchmark data presents a clear-cut split: the NVIDIA GeForce RTX 2060 SUPER dominates every measured workload, while the AMD Radeon Pro 5700’s sole claim to superiority is its slightly higher aggregate average score. In practical terms, the RTX 2060 SUPER is the stronger performer across the board, with advantages ranging from 5.7% in the lightweight PassMark G2D test to a commanding 40% in PassMark DirectX 11.

The AMD card’s closest relative performance is in the PassMark G2D test, where it scores 805 versus NVIDIA’s 854, a modest 5.7% deficit. This suggests the two are nearly comparable in 2D graphics operations. However, in 3D and compute workloads, NVIDIA’s lead widens substantially. The RTX 2060 SUPER excels in DirectX 11 and DirectX 9 legacy titles, showing 40% and 34.4% advantages respectively. It also leads in modern APIs, with a 21.3% edge in DirectX 12 and a 33.7% edge in Vulkan.

For users prioritizing compute performance, the NVIDIA card again wins, taking PassMark GPU Compute by 26.2% and GeekBench OpenCL by 31.4%. The only category where AMD comes close to parity is the aggregate average, but this is misleading, as the Radeon Pro 5700’s average is pulled up by its GeekBench Metal score of 60,126, a test that the NVIDIA card does not have a result for. Without that Metal result, the AMD card’s average would be considerably lower.

Specification Differences

The most glaring difference is the form factor and intended use case. The AMD Radeon Pro 5700 is listed as an IGP (integrated graphics processor) with no display outputs, no power connectors, and a 130 W TDP. It is designed to be embedded in a system, likely a Mac, rather than installed as a discrete card. The NVIDIA GeForce RTX 2060 SUPER, by contrast, is a dual-slot discrete card requiring a 1x 8-pin power connector, with a 175 W TDP and a suggested 450 W power supply. It measures 229 mm by 113 mm by 35 mm and provides a full suite of outputs: 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C.

Clock speeds favor NVIDIA significantly. The RTX 2060 SUPER has a base clock of 1470 MHz and a boost clock of 1650 MHz, while the Radeon Pro 5700 operates at 1243 MHz base and 1350 MHz boost. This clock advantage contributes to NVIDIA’s higher theoretical throughput: 7.181 TFLOPS FP32 versus AMD’s 6.221 TFLOPS, and 14.36 TFLOPS FP16 versus AMD’s 12.44 TFLOPS (both at 2:1 ratio). The pixel rate is 105.6 GPixel/s for NVIDIA versus 86.40 GPixel/s for AMD, and texture rates are 224.4 GTexel/s versus 194.4 GTexel/s.

The compute unit counts differ slightly. The AMD card has 2304 shading units, 144 TMUs, and 64 ROPs. The NVIDIA card has fewer shading units at 2176, fewer TMUs at 136, but the same 64 ROPs. NVIDIA compensates for the lower core count with higher clocks and the addition of 34 RT cores and 272 tensor cores. The release dates are about a year apart, with the RTX 2060 SUPER launching in July 2019 and the Radeon Pro 5700 in August 2020.

Head-to-Head Benchmarks

The head-to-head results are uniformly in favor of the NVIDIA GeForce RTX 2060 SUPER, making for a one-sided comparison. The largest victory comes in PassMark DirectX 11, where NVIDIA scores 130 against AMD’s 78, a 40% delta. This is a legacy API test, but the margin indicates a fundamental efficiency advantage for Turing in this workload. The DirectX 9 test shows a similar pattern, with NVIDIA at 218 versus AMD’s 143, a 34.4% lead.

Modern API results also favor NVIDIA decisively. In GeekBench Vulkan, the RTX 2060 SUPER scores 77,402 versus the Radeon Pro 5700’s 51,289, a 33.7% gap. The DirectX 12 test in PassMark shows NVIDIA ahead at 61 versus 48, a 21.3% margin. The GeekBench OpenCL test gives NVIDIA a 76,957 score against AMD’s 52,787, a 31.4% advantage.

Compute and 3D rendering tests continue the trend. PassMark GPU Compute shows NVIDIA at 6,721 versus AMD’s 4,961, a 26.2% lead. PassMark G3D, which is a more general 3D performance metric, has NVIDIA at 16,462 versus AMD’s 11,583, a 29.6% gap. Even the 2D test, PassMark G2D, goes to NVIDIA with 854 versus 805, though the 5.7% delta is the closest of any comparison.

The only test where AMD shows any competitive strength is GeekBench Metal, where it scores 60,126. However, this test has no corresponding result for the NVIDIA card, so it cannot be compared directly. The RTX 2060 SUPER also has a 3DMark Steel Nomad DX12 score of 2,011, which has no AMD counterpart. These exclusive results explain why the AMD card’s aggregate average score (18,189) is only 0.5% higher than NVIDIA’s (18,093), despite losing every shared benchmark. In the shared workload set, NVIDIA’s average is dramatically higher, and the data clearly shows that for any application using OpenCL, Vulkan, DirectX 9-12, or general compute, the RTX 2060 SUPER is the faster card by margins that range from substantial to overwhelming.