AMD Radeon Pro 5700 vs NVIDIA GeForce RTX 2060 Comparison

Where Each One Wins

The recorded benchmark data splits cleanly along workload type. The AMD Radeon Pro 5700 takes a single win out of nine head-to-head tests, and that win is in 2D graphics. The Passmark G2D score of 805 for the AMD card beats the NVIDIA GeForce RTX 2060's 745 by 8.1%. This suggests the AMD part has an edge in desktop composition, 2D rendering, or other non-3D tasks, which aligns with its integrated design and no display outputs.

Every other benchmark in the comparison goes to the NVIDIA GeForce RTX 2060. The NVIDIA card wins all 3D DirectX tests, both Geekbench compute tests, and the Passmark 3D and compute suites. The largest margins come in older DirectX APIs: the RTX 2060 leads by 29.1% in DirectX 11, 27.6% in DirectX 10, and 24.7% in DirectX 9. These are substantial gaps, indicating that legacy 3D workloads favor the NVIDIA architecture heavily.

For modern compute workloads, the NVIDIA card still leads but by smaller margins. In Geekbench OpenCL, the RTX 2060 scores 65014 against 52787 for the AMD card, a difference of 18.8%. In Geekbench Vulkan, the lead is 22.1% with scores of 65846 versus 51289. The Passmark GPU Compute test shows a narrower 9.6% advantage for NVIDIA, with scores of 5488 versus 4961.

The overall average benchmark score tells a similar story. The AMD Radeon Pro 5700 has an average score of 18189 across all recorded tests, placing it at the 62nd percentile of all GPUs in the database. The NVIDIA GeForce RTX 2060 has an average score of 15290, placing it at the 58th percentile. This is counterintuitive given the head-to-head results, but it reflects the different benchmark suites each card was tested with. The AMD card's average includes its strong Geekbench Metal score of 60126, which is not part of the head-to-head comparison. The RTX 2060's average is dragged down by its relatively low Passmark DirectX scores, which are all below 200 points.

Architecture Differences

The two cards come from different architectural generations and fabrication nodes. The AMD Radeon Pro 5700 uses the Navi 10 chip built on RDNA 1.0 architecture, manufactured on TSMC's 7 nm process. The NVIDIA GeForce RTX 2060 uses the TU106 chip based on Turing architecture, also made by TSMC but on a 12 nm process.

The transistor counts are close: the AMD chip has 10,300 million transistors on a 251 mm² die, while the NVIDIA chip has 10,800 million transistors on a much larger 445 mm² die. The 7 nm process gives AMD a transistor density of 41.0 million per mm², compared to NVIDIA's 24.3 million per mm² on 12 nm.

The compute configurations differ significantly. The AMD card has 2304 shading units, 144 texture mapping units, and 64 raster operation units. The NVIDIA card has fewer of each: 1920 shading units, 120 TMUs, and 48 ROPs. However, the NVIDIA card includes dedicated hardware that AMD lacks: 30 RT cores for ray tracing and 240 tensor cores for AI workloads. These are entirely absent from the AMD specification.

Clock speeds favor NVIDIA. The RTX 2060 has a base clock of 1365 MHz and a boost clock of 1680 MHz, while the Radeon Pro 5700 runs at 1243 MHz base and 1350 MHz boost. Despite the lower clock speeds and higher shading unit count, the AMD card's FP32 throughput is 6.221 TFLOPS versus 6.451 TFLOPS for NVIDIA, meaning the RTX 2060 actually delivers slightly more raw floating-point performance. FP16 performance follows the same pattern: 12.44 TFLOPS for AMD versus 12.90 TFLOPS for NVIDIA, both at a 2:1 ratio.

Memory configurations also differ. The AMD card has 8 GB of GDDR6 on a 256-bit bus, yielding 384.0 GB/s of bandwidth. The NVIDIA card has 6 GB of GDDR6 on a 192-bit bus, yielding 336.0 GB/s. The AMD card has more memory and more bandwidth, but the NVIDIA card runs its memory at a higher effective speed: 14 Gbps versus 12 Gbps.

The power and physical specs reflect their different intended roles. The Radeon Pro 5700 is an integrated GPU (IGP) with no power connectors and a 130 W TDP, suggesting it is designed for a Mac system. The RTX 2060 is a dual-slot card with one 8-pin power connector, a 160 W TDP, and a 229 mm length. The AMD card has no display outputs, while the NVIDIA card has a full complement including DVI, HDMI 2.0, two DisplayPort 1.4a outputs, and a USB Type-C port.

Head-to-Head Benchmarks

The biggest single margin in the comparison is in Passmark DirectX 11, where the NVIDIA card scores 110 against the AMD card's 78, a 29.1% advantage. This is followed by DirectX 10, where scores of 98 versus 71 give NVIDIA a 27.6% lead. These older APIs show NVIDIA's Turing architecture handling legacy workloads much more efficiently.

DirectX 9 shows a 24.7% gap, with scores of 190 for NVIDIA and 143 for AMD. This is the highest absolute score in any Passmark DirectX test for either card, suggesting both handle the older API well but NVIDIA does so considerably better. DirectX 12 narrows the gap to 9.4%, with scores of 53 for NVIDIA and 48 for AMD. The modern API reduces the architectural advantage, though NVIDIA still wins.

Geekbench Vulkan shows a 22.1% lead for NVIDIA, with scores of 65846 versus 51289. Geekbench OpenCL shows an 18.8% lead, with scores of 65014 versus 52787. These compute-oriented tests demonstrate that NVIDIA's combination of higher clocks and dedicated tensor cores translates to meaningful performance advantages in general-purpose GPU workloads.

The Passmark G3D score, which aggregates 3D gaming performance, gives NVIDIA a 17.9% advantage with scores of 14111 versus 11583. This is the most comprehensive 3D test in the comparison and reinforces the pattern. The Passmark GPU Compute test is the closest 3D-adjacent result, with NVIDIA leading by 9.6% (5488 versus 4961).

The single AMD win comes in Passmark G2D, where the Radeon Pro 5700 scores 805 versus 745 for NVIDIA, an 8.1% advantage. This is the only test where AMD leads, and it is a 2D workload rather than a 3D or compute workload.

FAQ

Q: Which card has better raw compute performance in OpenCL?

A: The NVIDIA GeForce RTX 2060 wins in Geekbench OpenCL with a score of 65014, which is 18.8% higher than the AMD Radeon Pro 5700's 52787.

Q: Does the AMD card win any benchmark against the RTX 2060?

A: Yes, the AMD Radeon Pro 5700 wins the Passmark G2D test with a score of 805, which is 8.1% higher than the NVIDIA card's 745.

Q: How do the two cards compare in DirectX 12 performance?

A: The NVIDIA card leads in Passmark DirectX 12 with a score of 53 versus 48 for AMD, a margin of 9.4%. This is the smallest gap in any DirectX test.

Q: What is the difference in memory capacity and bandwidth?

A: The AMD card has 8 GB of GDDR6 memory on a 256-bit bus with 384.0 GB/s bandwidth. The NVIDIA card has 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth.

Q: Which card has dedicated ray tracing hardware?

A: Only the NVIDIA GeForce RTX 2060 has dedicated ray tracing hardware, with 30 RT cores. The AMD Radeon Pro 5700 has no RT cores listed.

Q: What are the average benchmark scores and percentiles for each card?

A: The AMD card has an average score of 18189 and sits at the 62nd percentile of all GPUs. The NVIDIA card has an average score of 15290 and sits at the 58th percentile.

The Verdict

The data points to a clear overall winner for 3D and compute workloads. The NVIDIA GeForce RTX 2060 wins 8 of 9 head-to-head benchmarks, with margins ranging from 9.4% in DirectX 12 to 29.1% in DirectX 11. For anyone running 3D applications, older DirectX titles, or GPU compute tasks, the RTX 2060 is the stronger choice based on these measurements.

The AMD Radeon Pro 5700 has a single strength: 2D graphics. Its 8.1% lead in Passmark G2D suggests it handles desktop rendering or 2D compositing tasks better. It also has advantages in memory capacity (8 GB versus 6 GB) and memory bandwidth (384.0 GB/s versus 336.0 GB/s), which are not directly reflected in the benchmark results but matter for large datasets or high-resolution textures.

The RTX 2060 has additional hardware features that the AMD card lacks: 30 RT cores and 240 tensor cores. These enable ray tracing and AI-accelerated workloads, although the benchmark data does not include specific tests for those features. The NVIDIA card also has a higher boost clock (1680 MHz versus 1350 MHz) and higher FP32 throughput (6.451 TFLOPS versus 6.221 TFLOPS).

The average benchmark scores tell a different story, with the AMD card at 18189 versus 15290 for NVIDIA. This discrepancy comes from the AMD card's strong Geekbench Metal score of 60126, a test not present in the head-to-head comparison. Users on macOS or Metal-based applications may find the AMD card performs better in that specific API.

For most use cases, the RTX 2060 is the data-backed recommendation. Its wins span every 3D API, both Geekbench compute tests, and the comprehensive Passmark G3D score. The only scenario where the AMD card appears preferable is 2D-focused work or environments where its larger memory pool and higher bandwidth are decisive.

Specification Differences

| Specification | AMD Radeon Pro 5700 | NVIDIA GeForce RTX 2060 |

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

| Architecture | RDNA 1.0 | Turing |

| Process Node | 7 nm | 12 nm |

| Die Size | 251 mm² | 445 mm² |

| Transistor Density | 41.0M / mm² | 24.3M / mm² |

| Base Clock | 1243 MHz | 1365 MHz |

| Boost Clock | 1350 MHz | 1680 MHz |

| Memory Size | 8 GB | 6 GB |

| Memory Bus Width | 256 bit | 192 bit |

| Memory Bandwidth | 384.0 GB/s | 336.0 GB/s |

| Shading Units | 2304 | 1920 |

| TMUs | 144 | 120 |

| ROPs | 64 | 48 |

| RT Cores | None | 30 |

| Tensor Cores | None | 240 |

| FP32 Performance | 6.221 TFLOPS | 6.451 TFLOPS |

| FP16 Performance | 12.44 TFLOPS (2:1) | 12.90 TFLOPS (2:1) |

| TDP | 130 W | 160 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 8-pin |

| Suggested PSU | 300 W | 450 W |

| Bus Interface | PCIe 4.0 x16 | PCIe 3.0 x16 |

| Display Outputs | No outputs | 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, 1x USB Type-C |

| DirectX Support | 12 (12_1) | 12 Ultimate (12_2) |

| Release Date | 2020-08-03 | 2019-01-06 |