AMD FirePro M6100 vs NVIDIA P106-090 Comparison

Head-to-Head Benchmarks

The only direct benchmark comparison available between the NVIDIA P106-090 and the AMD FirePro M6100 is the Geekbench OpenCL test, and the result is decisive. The NVIDIA P106-090 scores 21,304 points, while the AMD FirePro M6100 manages 13,354 points. This translates to a 59.5% advantage for the NVIDIA part. That is a substantial performance gap, placing the P106-090 in a different performance tier entirely for compute workloads.

To contextualize this result, the average benchmark score across all recorded tests for the P106-090 is 13,470, which is 0.5% higher than the AMD Radeon Pro 555 (13,407) and 0.3% lower than the NVIDIA GeForce GTX 570 (13,515). The FirePro M6100’s average score is 13,354, which sits 0.2% below the AMD Radeon HD 8950M (13,376) and 0.3% above the AMD Radeon Pro 555X (13,321). Both cards occupy the 54th percentile of all GPUs, meaning they are statistically mid-pack performers, but the OpenCL result shows the NVIDIA silicon has significantly more compute headroom in at least one major API.

The delta in OpenCL performance is not marginal; it is a landslide. A 59.5% improvement in a raw compute benchmark suggests fundamental architectural and execution efficiency differences. The P106-090’s score of 21,304 is closer to the Geekbench Vulkan result of 18,596 for the same card, indicating that its compute throughput is consistently high across different API workloads. The FirePro M6100 has no Vulkan benchmark recorded, so the OpenCL result remains the sole point of direct comparison, but it is a telling one.

Where Each One Wins

The NVIDIA P106-090 wins the only head-to-head benchmark, claiming a clean sweep with 1 win versus 0 for the AMD FirePro M6100. The advantage is most pronounced in raw compute throughput, as evidenced by the 59.5% lead in Geekbench OpenCL. This suggests the P106-090 is better suited for tasks that leverage general-purpose GPU compute, such as rendering, simulation, or data-parallel workloads.

The FirePro M6100, despite its loss, is not without merit in other contexts. Its architecture is designed for mobile professional workstations, as indicated by its MXM Module form factor and "Portable Device Dependent" display outputs. This means it is intended for laptops or compact systems where space and power are constrained, whereas the P106-090 is a dual-slot, 250mm-long desktop card with a 1x 6-pin power connector. The FirePro’s lack of a recorded TDP and its "None" power connector requirement suggest it is a lower-power, more thermally conservative solution for integrated mobile platforms.

In terms of benchmark percentiles, both cards are tied at the 54th percentile, meaning neither is a standout performer in the broader GPU landscape. However, the P106-090’s 59.5% OpenCL lead shows that within this specific workload, it is the clear victor. The FirePro M6100’s only recorded benchmark is its OpenCL score, so there is no data to suggest it wins in any other category. The P106-090 also has a Geekbench Vulkan score of 18,596 and a 3DMark Steel Nomad DX12 score of 509, giving it a more complete benchmark profile.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The NVIDIA P106-090 uses the GP106 chip based on the Pascal architecture, manufactured on TSMC’s 16 nm process. It integrates 4,400 million transistors on a 200 mm² die, yielding a transistor density of 22.0 million per mm². In contrast, the AMD FirePro M6100 uses the Emerald chip based on GCN 2.0, built on TSMC’s 28 nm process. It contains 2,080 million transistors on a 160 mm² die, with a transistor density of 13.0 million per mm². The newer 16 nm node gives NVIDIA a significant density advantage, allowing more transistors in a larger die area while potentially improving power efficiency.

The memory subsystems also diverge sharply. The P106-090 has 3 GB of GDDR5 memory on a 192-bit bus, delivering 192.2 GB/s of bandwidth. The FirePro M6100 has 2 GB of GDDR5 on a 128-bit bus, yielding only 96.00 GB/s. This is a 2x difference in bandwidth, which directly impacts performance in memory-intensive workloads. The P106-090’s memory clock is 2002 MHz (8 Gbps effective), while the FirePro’s is 1500 MHz (6 Gbps effective). Higher clock speed and wider bus combine to give NVIDIA a decisive memory throughput advantage.

Compute resources tell a more nuanced story. The FirePro M6100 has more shading units (896 vs. 768) and more texture mapping units (56 vs. 48), but the P106-090 has far more raster operation units (48 vs. 16). This explains the pixel rate disparity: the P106-090 achieves 73.49 GPixel/s versus the FirePro’s 17.60 GPixel/s. The texture rates are closer, with 73.49 GTexel/s for NVIDIA and 61.60 GTexel/s for AMD. In raw FP32 compute, the P106-090 reaches 2.352 TFLOPS, while the FirePro reaches 1.971 TFLOPS. The P106-090 also supports FP16 at 36.74 GFLOPS (1:64 ratio), whereas the FirePro has no recorded FP16 capability.

API support differs as well. Both support DirectX 12 and OpenGL 4.6, but the P106-090 supports DirectX 12_1 and Vulkan 1.4, while the FirePro supports DirectX 12_0 and Vulkan 1.2.170. The newer Vulkan version on NVIDIA may offer better performance in modern titles that leverage the latest API features. The bus interface also contrasts: the P106-090 uses PCIe 1.0 x1 (a legacy limitation for a mining card), while the FirePro uses MXM-B (3.0), which is a mobile module standard.

FAQ

Q: Which GPU has higher raw compute performance in OpenCL?

A: The NVIDIA P106-090 scores 21,304 in Geekbench OpenCL, which is 59.5% higher than the AMD FirePro M6100’s score of 13,354.

Q: Do the two cards have the same memory bandwidth?

A: No. The NVIDIA P106-090 has 192.2 GB/s bandwidth from 3 GB GDDR5 on a 192-bit bus, while the AMD FirePro M6100 has 96.00 GB/s from 2 GB GDDR5 on a 128-bit bus.

Q: Which GPU has more shading units?

A: The AMD FirePro M6100 has 896 shading units, compared to 768 on the NVIDIA P106-090. Despite this, the NVIDIA card achieves higher FP32 throughput (2.352 TFLOPS vs. 1.971 TFLOPS).

Q: What is the process node difference between the two GPUs?

A: The NVIDIA P106-090 is manufactured on a 16 nm process, while the AMD FirePro M6100 uses a 28 nm process. Both are fabricated by TSMC.

Q: Which card has better pixel fill rate?

A: The NVIDIA P106-090 has a pixel rate of 73.49 GPixel/s, which is over 4x higher than the AMD FirePro M6100’s 17.60 GPixel/s, due to its 48 ROPs versus 16.

Q: Are both GPUs still in production?

A: No. Both the NVIDIA P106-090 and the AMD FirePro M6100 are listed as end-of-life products.

Specification Differences

| Specification | NVIDIA P106-090 | AMD FirePro M6100 |

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

| Architecture | Pascal | GCN 2.0 |

| Process Node | 16 nm | 28 nm |

| Transistors | 4,400 million | 2,080 million |

| Die Size | 200 mm² | 160 mm² |

| Transistor Density | 22.0M / mm² | 13.0M / mm² |

| Memory Size | 3 GB | 2 GB |

| Memory Bus Width | 192 bit | 128 bit |

| Memory Bandwidth | 192.2 GB/s | 96.00 GB/s |

| Memory Clock | 2002 MHz (8 Gbps effective) | 1500 MHz (6 Gbps effective) |

| Shading Units | 768 | 896 |

| TMUs | 48 | 56 |

| ROPs | 48 | 16 |

| Pixel Rate | 73.49 GPixel/s | 17.60 GPixel/s |

| Texture Rate | 73.49 GTexel/s | 61.60 GTexel/s |

| FP32 Performance | 2.352 TFLOPS | 1.971 TFLOPS |

| FP16 Performance | 36.74 GFLOPS (1:64) | Not specified |

| TDP | 75 W | Not specified |

| Slot Width | Dual-slot | MXM Module |

| Power Connectors | 1x 6-pin | None |

| Suggested PSU | 250 W | Not specified |

| Bus Interface | PCIe 1.0 x1 | MXM-B (3.0) |

| Display Outputs | No outputs | Portable Device Dependent |

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

| Vulkan Support | 1.4 | 1.2.170 |

| Release Date | 2017-07-30 | 2014-05-26 |

| Predecessor | Not specified | FirePro Mobility |

| Successor | Not specified | Radeon Pro Mobile |