AMD Instinct MI100 vs NVIDIA Quadro P6000 Comparison

The Verdict

The benchmark database clearly separates these two professional workstation cards. The AMD Instinct MI100 is the compute-focused performer, delivering a dominant lead in the only shared benchmark. The NVIDIA Quadro P6000, while an older design, remains a capable option for those needing display outputs and broader API support.

For raw compute workloads measured by the database, the AMD Instinct MI100 is the clear choice. Its Geekbench OpenCL score of 139,035 is more than double the Quadro P6000's 66,382. The MI100 sits at the 96th percentile among all GPUs, while the P6000 sits at the 90th. This gap is significant and consistent with their respective architectures.

However, the Quadro P6000 is not without merit. It offers display outputs (1x DVI and 4x DisplayPort 1.4a), which the MI100 completely lacks. The P6000 also supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, whereas the MI100 reports N/A for all three APIs. For users needing a GPU that can drive monitors and run graphics applications, the P6000 is the only viable option from this pair.

The verdict depends entirely on workflow. Compute-heavy, headless servers favor the MI100. Workstations requiring displays and graphics API support favor the Quadro P6000. The data shows a performance chasm in compute, but the P6000 fills a functional gap the MI100 cannot.

Architecture Differences

The two cards come from different architectural generations. The AMD Instinct MI100 uses the Arcturus chip built on CDNA 1.0 architecture, manufactured on a 7 nm process at TSMC. The NVIDIA Quadro P6000 uses the GP102 chip on the older Pascal architecture, manufactured on a 16 nm process, also at TSMC.

Transistor counts differ substantially. The MI100 packs 25,600 million transistors on a 750 mm² die, giving a transistor density of 34.1 million per mm². The P6000 has 11,800 million transistors on a 471 mm² die, with a density of 25.1 million per mm². The MI100's newer process node allows for significantly more transistors and higher density.

Memory architectures are fundamentally different. The MI100 uses 32 GB of HBM2 with a 4096-bit bus and 1.23 TB/s of bandwidth. The P6000 uses 24 GB of GDDR5X with a 384-bit bus and 432.8 GB/s of bandwidth. The MI100's memory subsystem offers nearly three times the bandwidth, which is critical for compute workloads.

Compute unit counts also favor the MI100. It has 7,680 shading units, 480 TMUs, and 64 ROPs. The P6000 has 3,840 shading units, 240 TMUs, and 96 ROPs. The MI100 has double the shading units and TMUs, while the P6000 has 50% more ROPs. Clock speeds differ: the MI100 runs at 1000 MHz base and 1502 MHz boost, while the P6000 runs at 1506 MHz base and 1645 MHz boost. The P6000's higher clocks help its pixel rate (157.9 GPixel/s vs 96.13 GPixel/s), but the MI100's massive shading unit count drives its texture rate (721.0 GTexel/s vs 394.8 GTexel/s) and FP32 throughput (23.07 TFLOPS vs 12.63 TFLOPS).

Power and connectivity also differ. The MI100 has a 300 W TDP with 2x 8-pin connectors and a suggested PSU of 700 W. The P6000 has a 250 W TDP with 1x 8-pin and a suggested PSU of 600 W. The MI100 uses PCIe 4.0 x16, while the P6000 uses PCIe 3.0 x16.

Where Each One Wins

The AMD Instinct MI100 wins decisively in compute throughput. Its FP32 performance of 23.07 TFLOPS nearly doubles the P6000's 12.63 TFLOPS. FP16 performance shows an even larger gap: 46.14 TFLOPS for the MI100 versus just 197.4 GFLOPS for the P6000. The MI100's FP16 advantage is a 2:1 ratio over FP32, while the P6000's FP16 is a 1:64 ratio, making it effectively negligible.

Memory bandwidth is another MI100 victory. The 1.23 TB/s of HBM2 bandwidth versus 432.8 GB/s of GDDR5X means the MI100 can feed its compute units far more effectively. This is crucial for large datasets and memory-bound workloads.

The NVIDIA Quadro P6000 wins in display and graphics functionality. It has 1x DVI and 4x DisplayPort 1.4a outputs, enabling multi-monitor setups. The MI100 has no display outputs at all. The P6000 also supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, making it usable for professional graphics applications. The MI100 lists N/A for all these APIs.

The P6000 also has a higher pixel rate at 157.9 GPixel/s versus 96.13 GPixel/s, driven by its 96 ROPs and higher boost clock. This suggests better fill-rate performance for rasterization tasks, though the MI100's compute dominance overshadows this in most professional workloads.

FAQ

Q: Which card has better compute performance?

A: The AMD Instinct MI100. Its Geekbench OpenCL score is 139,035, versus 66,382 for the NVIDIA Quadro P6000. The MI100 also delivers 23.07 TFLOPS FP32 and 46.14 TFLOPS FP16, compared to 12.63 TFLOPS FP32 and 197.4 GFLOPS FP16 for the P6000.

Q: Can either card connect to a monitor?

A: Only the NVIDIA Quadro P6000. It has 1x DVI and 4x DisplayPort 1.4a outputs. The AMD Instinct MI100 has no display outputs.

Q: What is the memory bandwidth difference?

A: The MI100 has 1.23 TB/s from 32 GB of HBM2 on a 4096-bit bus. The P6000 has 432.8 GB/s from 24 GB of GDDR5X on a 384-bit bus. The MI100 offers roughly three times the bandwidth.

Q: Which card supports newer graphics APIs?

A: The NVIDIA Quadro P6000 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The AMD Instinct MI100 reports N/A for all three.

Q: How do their power requirements compare?

A: The MI100 has a 300 W TDP and requires a 700 W suggested PSU with 2x 8-pin connectors. The P6000 has a 250 W TDP and requires a 600 W suggested PSU with 1x 8-pin.

Q: Which card ranks higher overall?

A: The MI100 sits at the 96th percentile among all GPUs, while the P6000 sits at the 90th percentile. The MI100's average benchmark score is 139,035, versus 69,986 for the P6000.

Head-to-Head Benchmarks

The database contains one shared benchmark: Geekbench OpenCL. The AMD Instinct MI100 scores 139,035, while the NVIDIA Quadro P6000 scores 66,382. This gives the MI100 a 109.4% advantage, meaning it delivers more than double the compute performance.

This result aligns with the nearest rivals data. The MI100's closest competitors are the NVIDIA Tesla V100 PCIe 16 GB (138,063, just 0.7% behind), the Tesla V100 SXM2 32 GB (137,731, 0.9% behind), the AMD Radeon PRO V620 (136,472, 1.9% behind), and the AMD Radeon Pro W6800X Duo (135,774, 2.4% behind). The MI100 leads all of them, but by a narrow margin.

The P6000's nearest rivals are much closer to its own score. The NVIDIA CMP 90HX (69,000) is 1.4% behind, the AMD Radeon Pro WX 8200 (69,870) is 0.2% behind, the NVIDIA RTX A3000 Mobile (70,140) is 0.2% ahead, and the AMD Radeon RX 6600 LE (70,829) is 1.2% ahead. The P6000's 66,382 average is competitive with these cards, but it sits far below the MI100's tier.

The head-to-head delta of 109.4% is the largest in this comparison. No other benchmark in the database narrows this gap. The MI100's architecture, with double the shading units and nearly three times the memory bandwidth, explains this result. The P6000's higher clocks and ROP count cannot compensate for the compute resource disparity.

Specification Differences

The two cards differ across nearly every specification field.

| Specification | AMD Instinct MI100 | NVIDIA Quadro P6000 |

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

| Architecture | CDNA 1.0 | Pascal |

| Process node | 7 nm | 16 nm |

| Transistors | 25,600 million | 11,800 million |

| Die size | 750 mm² | 471 mm² |

| Transistor density | 34.1M / mm² | 25.1M / mm² |

| Base clock | 1000 MHz | 1506 MHz |

| Boost clock | 1502 MHz | 1645 MHz |

| Memory size | 32 GB | 24 GB |

| Memory type | HBM2 | GDDR5X |

| Memory bus | 4096 bit | 384 bit |

| Memory bandwidth | 1.23 TB/s | 432.8 GB/s |

| Shading units | 7680 | 3840 |

| TMUs | 480 | 240 |

| ROPs | 64 | 96 |

| Pixel rate | 96.13 GPixel/s | 157.9 GPixel/s |

| Texture rate | 721.0 GTexel/s | 394.8 GTexel/s |

| FP32 | 23.07 TFLOPS | 12.63 TFLOPS |

| FP16 | 46.14 TFLOPS (2:1) | 197.4 GFLOPS (1:64) |

| TDP | 300 W | 250 W |

| Power connectors | 2x 8-pin | 1x 8-pin |

| Suggested PSU | 700 W | 600 W |

| Bus interface | PCIe 4.0 x16 | PCIe 3.0 x16 |

| Display outputs | No outputs | 1x DVI, 4x DisplayPort 1.4a |

| DirectX | N/A | 12 (12_1) |

| OpenGL | N/A | 4.6 |

| Vulkan | N/A | 1.4 |

| Release date | 2020-11-15 | 2016-09-30 |

| Predecessor | Radeon Instinct | Quadro Maxwell |

| Successor | None | Quadro Volta |

| Launch MSRP | None | 5,999 USD |

Both cards are dual-slot, end-of-life products with identical physical dimensions: 267 mm (10.5 inches) in length and 111 mm (4.4 inches) in height. The MI100's 25,600 million transistors on a 750 mm² die represent the newer, larger design, while the P6000's 11,800 million transistors on 471 mm² reflect its older generation. The MI100 also uses PCIe 4.0, doubling the interconnect bandwidth of the P6000's PCIe 3.0 interface.