GPU Comparison

The NVIDIA Quadro M6000 and AMD Radeon Pro 580X represent two distinct approaches to professional graphics, separated by process node, architecture generation, and target platform. The data shows a clear overall winner in raw compute benchmarks, with the Quadro M6000 taking both head-to-head tests. However, the Radeon Pro 580X holds its own in specific API workloads and offers a different feature set tied to its Apple-centric design. This analysis breaks down where each card excels, the architectural reasons behind those differences, and the exact benchmark deltas that separate them.

Where Each One Wins

The benchmark results are unambiguous in their overall verdict: the NVIDIA Quadro M6000 wins both head-to-head tests conducted. It takes the Geekbench OpenCL test with a score of 39,688 against the Radeon Pro 580X’s 36,426, a 9% advantage. The gap widens considerably in the Vulkan test, where the Quadro M6000 scores 46,913 versus 40,115, a 16.9% lead. This makes the Quadro the superior choice for general compute workloads and for Vulkan-based applications, where its architecture appears to scale more effectively.

The AMD Radeon Pro 580X, while losing both direct comparisons, does not lack merit. Its only uniquely recorded benchmark is Geekbench Metal, where it scores 39,577. This is a workload the Quadro M6000 cannot contest, as it has no corresponding Metal score in the data. For users in the Apple ecosystem, specifically those running the Radeon Pro Mac generation, this Metal performance is the relevant metric. The card’s overall average benchmark score of 38,706 is lower than the Quadro’s 43,301, but its percentile ranking at 82 versus the Quadro’s 84 shows that both sit in a similar performance tier relative to all GPUs. The Radeon’s win condition is not raw speed, but platform compatibility and API-specific optimization for Metal.

Architecture Differences

The two cards come from different manufacturing and design philosophies. The NVIDIA Quadro M6000 uses the GM200 chip built on a 28 nm process at TSMC, packing 8,000 million transistors into a 601 mm² die. This yields a transistor density of 13.3M per mm². In contrast, the AMD Radeon Pro 580X uses the Ellesmere chip on a more advanced 14 nm process at GlobalFoundries, with 5,700 million transistors on a much smaller 232 mm² die, achieving a higher density of 24.6M per mm². The smaller, denser die suggests AMD’s process advantage, but the Quadro’s larger physical footprint allows for more raw hardware resources.

The Quadro M6000 features 3,072 shading units, 192 texture mapping units, and 96 ROPs. The Radeon Pro 580X counters with 2,304 shading units, 144 TMUs, and only 32 ROPs. This ROP disparity is significant: the Quadro’s pixel rate is 106.9 GPixel/s, nearly three times the Radeon’s 38.40 GPixel/s. The texture rate also favors NVIDIA, at 213.9 GTexel/s versus 172.8 GTexel/s. In floating-point compute, the Quadro delivers 6.844 TFLOPS in FP32, while the Radeon provides 5.530 TFLOPS. Notably, the Radeon matches its FP32 rate in FP16 (5.530 TFLOPS at 1:1), while the Quadro lists no FP16 capability, suggesting a potential advantage for the AMD card in mixed-precision workloads.

Memory configurations diverge sharply. The Quadro M6000 carries 12 GB of GDDR5 on a 384-bit bus, yielding 317.4 GB/s of bandwidth. The Radeon Pro 580X has 8 GB of GDDR5 on a 256-bit bus, producing 218.9 GB/s. Clock speeds are higher on the AMD part, base 1100 MHz and boost 1200 MHz versus 988 MHz and 1114 MHz for NVIDIA, but the wider bus and memory size of the Quadro overcome this in bandwidth-bound tasks. Power and form factor also differ: the Quadro is a 250 W dual-slot card requiring a single 8-pin connector and a 600 W power supply, while the Radeon is an integrated graphics processor (IGP) with a 185 W TDP, no power connectors, and no suggested PSU, reflecting its designed integration into Apple’s MPX bus interface.

Head-to-Head Benchmarks

The Geekbench OpenCL test shows a 9% win for the NVIDIA Quadro M6000, with scores of 39,688 against 36,426. This margin is modest but consistent with the Quadro’s higher FP32 throughput and memory bandwidth. The 8,000 million transistor count and 384-bit memory bus provide a solid foundation for compute-heavy OpenCL workloads, where the Radeon’s narrower 256-bit bus becomes a bottleneck despite its higher clocks.

The Vulkan test is where the Quadro M6000 decisively separates itself. Its score of 46,913 beats the Radeon Pro 580X’s 40,115 by 16.9%. This larger delta suggests that NVIDIA’s Maxwell 2.0 architecture handles Vulkan’s low-level draw call and command buffer management more efficiently than AMD’s GCN 4.0. The Quadro also lists Vulkan 1.4 API support versus the Radeon’s Vulkan 1.3, which may contribute to better driver-level optimization. In both tests, the Quadro’s higher ROP count (96 versus 32) and pixel rate likely aid in geometry-heavy scenes, even though these are compute benchmarks, as memory access patterns benefit from the larger cache hierarchy implied by the bigger die.

FAQ

Q: Which card has a higher average benchmark score?

A: The NVIDIA Quadro M6000 has an average benchmark score of 43,301, while the AMD Radeon Pro 580X averages 38,706. This places the Quadro in the 84th percentile of all GPUs, compared to the Radeon’s 82nd percentile.

Q: Does the Radeon Pro 580X win any benchmark tests?

A: The Radeon Pro 580X has a recorded Geekbench Metal score of 39,577, which is a test the Quadro M6000 does not have a score for. However, in the two head-to-head tests (OpenCL and Vulkan), the Quadro wins both.

Q: What is the memory bandwidth difference between the two?

A: The Quadro M6000 offers 317.4 GB/s of bandwidth from its 12 GB GDDR5 memory on a 384-bit bus. The Radeon Pro 580X provides 218.9 GB/s from 8 GB GDDR5 on a 256-bit bus.

Q: How do their power requirements compare?

A: The Quadro M6000 has a 250 W TDP, is dual-slot, and requires a single 8-pin power connector with a 600 W suggested PSU. The Radeon Pro 580X has a 185 W TDP and is an IGP with no power connectors or suggested PSU.

Q: Which card supports a higher DirectX version?

A: The Quadro M6000 supports DirectX 12 (12_1), while the Radeon Pro 580X supports DirectX 12 (12_0). Both support OpenGL 4.6, but the Quadro lists Vulkan 1.4 versus the Radeon’s Vulkan 1.3.

Q: What is the process node and die size difference?

A: The Quadro M6000 uses a 28 nm process with a 601 mm² die and 8,000 million transistors. The Radeon Pro 580X uses a 14 nm process with a 232 mm² die and 5,700 million transistors.

Specification Differences

The following table lists only the fields where the two cards differ, based on the provided data.

| Specification | NVIDIA Quadro M6000 | AMD Radeon Pro 580X |

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

| Chip | GM200 | Ellesmere |

| Architecture | Maxwell 2.0 | GCN 4.0 |

| Generation | Quadro Maxwell (Mx000) | Radeon Pro Mac (500X Series) |

| Process Node | 28 nm | 14 nm |

| Foundry | TSMC | GlobalFoundries |

| Transistors | 8,000 million | 5,700 million |

| Die Size | 601 mm² | 232 mm² |

| Transistor Density | 13.3M / mm² | 24.6M / mm² |

| Base Clock | 988 MHz | 1100 MHz |

| Boost Clock | 1114 MHz | 1200 MHz |

| Memory Clock | 1653 MHz (6.6 Gbps effective) | 1710 MHz (6.8 Gbps effective) |

| Memory Size | 12 GB | 8 GB |

| Memory Bus Width | 384 bit | 256 bit |

| Memory Bandwidth | 317.4 GB/s | 218.9 GB/s |

| Shading Units | 3072 | 2304 |

| TMUs | 192 | 144 |

| ROPs | 96 | 32 |

| Pixel Rate | 106.9 GPixel/s | 38.40 GPixel/s |

| Texture Rate | 213.9 GTexel/s | 172.8 GTexel/s |

| FP32 | 6.844 TFLOPS | 5.530 TFLOPS |

| FP16 | Not listed | 5.530 TFLOPS (1:1) |

| TDP | 250 W | 185 W |

| Slot Width | Dual-slot | IGP |

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

| Suggested PSU | 600 W | None |

| Bus Interface | PCIe 3.0 x16 | Apple MPX |

| Display Outputs | 1x DVI, 4x DisplayPort 1.2 | 2x HDMI 2.0b |

| DirectX | 12 (12_1) | 12 (12_0) |

| Vulkan | 1.4 | 1.3 |

| Dimensions | 267 mm (10.5 inches) long, 111 mm (4.4 inches) high | Not listed |

| Release Date | 2015-03-20 | 2019-03-17 |

| Predecessor | Quadro Kepler | Not listed |

| Successor | Quadro Pascal | Not listed |

| Geekbench OpenCL | 39,688 | 36,426 |

| Geekbench Vulkan | 46,913 | 40,115 |

| Geekbench Metal | Not listed | 39,577 |

| Average Benchmark Score | 43,301 | 38,706 |

| Percentile vs All GPUs | 84 | 82 |