NVIDIA P106-090 vs NVIDIA RTX PRO 6000D Blackwell Max-Q Comparison

Where Each One Wins

The recorded data splits these two NVIDIA parts into completely different performance classes, with essentially no overlap in their intended workloads. The NVIDIA P106-090 is a Pascal-generation mining GPU with no display outputs, built exclusively for compute tasks that do not require a video signal. Its benchmark presence is modest, and its three recorded scores place it near the middle of the database’s GPU distribution. The NVIDIA RTX PRO 6000D Blackwell Max-Q, by contrast, is an active workstation part with full display outputs and a much larger memory pool, and it dominates the single shared benchmark by a wide margin.

In the only head-to-head test available, 3DMark Steel Nomad DX12, the RTX PRO 6000D Blackwell Max-Q scores 11088 versus the P106-090’s 509. That is a 95.4% deficit for the older card, meaning the Blackwell part delivers roughly 21.8 times the score in that specific workload. For any DirectX 12 gaming or rendering scenario represented by Steel Nomad, the RTX PRO 6000D Blackwell Max-Q is the clear winner.

However, the P106-090 does have its own measured strengths outside the head-to-head. In Geekbench OpenCL it scores 21304, and in Geekbench Vulkan it scores 18596. These are not directly compared to the RTX PRO 6000D Blackwell Max-Q in the database, but they show the mining card is functional in general-purpose compute APIs. The RTX PRO 6000D Blackwell Max-Q has no recorded Geekbench results, so the P106-090 holds the only data points in those tests. If the task is OpenCL or Vulkan compute where the newer card has not been measured, the P106-090 at least provides a reference number, though its overall performance percentile is only 54 versus the RTX PRO’s 50.

The real split is architectural and generational. The P106-090 belongs to the Mining GPUs generation, launched in 2017, and is end-of-life. The RTX PRO 6000D Blackwell Max-Q belongs to the Blackwell PRO W (x000) generation, released in 2025, and remains active. The P106-090 wins on the metric of having no display outputs, which makes it unsuitable for desktop use, while the RTX PRO 6000D Blackwell Max-Q provides four DisplayPort 2.1b outputs. For pure compute density, the Blackwell part wins. For legacy OpenCL or Vulkan workloads where only the older card has data, the P106-090 is the only measured option.

FAQ

Q: Which GPU is faster in the shared 3DMark Steel Nomad DX12 benchmark?

A: The NVIDIA RTX PRO 6000D Blackwell Max-Q scores 11088, while the NVIDIA P106-090 scores 509. The RTX PRO is ahead by 95.4% in that test, meaning it delivers about 21.8 times the score.

Q: Does the P106-090 have any benchmark wins over the RTX PRO 6000D Blackwell Max-Q?

A: In the head-to-head data, the P106-090 wins 0 tests and the RTX PRO 6000D Blackwell Max-Q wins 1 test. The P106-090 does have Geekbench OpenCL and Vulkan scores, but the RTX PRO has no recorded results in those tests, so no direct comparison exists.

Q: What are the memory specifications of each card?

A: The P106-090 has 3 GB of GDDR5 memory on a 192-bit bus, with 192.2 GB/s bandwidth. The RTX PRO 6000D Blackwell Max-Q has 96 GB of GDDR7 memory on a 512-bit bus, with 1.79 TB/s bandwidth.

Q: Which card supports display outputs?

A: The P106-090 has no display outputs at all, making it unusable for display output. The RTX PRO 6000D Blackwell Max-Q has 4x DisplayPort 2.1b outputs.

Q: What is the production status of each GPU?

A: The P106-090 is end-of-life, released on 2017-07-30. The RTX PRO 6000D Blackwell Max-Q is active, released on 2025-03-17.

Q: What is the launch MSRP of the RTX PRO 6000D Blackwell Max-Q?

A: The launch MSRP of the RTX PRO 6000D Blackwell Max-Q is 8,565 USD, stated once here per database rules.

Architecture Differences

The two GPUs come from entirely different eras of NVIDIA’s roadmap, and the data shows a generational leap in nearly every architectural metric. The P106-090 uses the GP106 chip on a 16 nm TSMC process, integrating 4,400 million transistors on a 200 mm² die, which yields a transistor density of 22.0 million per square millimeter. The RTX PRO 6000D Blackwell Max-Q uses the GB202 chip on a 5 nm TSMC process, with 92,200 million transistors on a 750 mm² die, for a density of 122.9 million per square millimeter. That is a density increase of roughly 5.6 times, and a transistor count increase of about 21 times.

The core configurations differ dramatically. The P106-090 has 768 shading units, 48 texture mapping units, and 48 ROPs. The RTX PRO 6000D Blackwell Max-Q has 24,064 shading units, 752 TMUs, and 192 ROPs. The Blackwell part also includes 188 RT cores and 752 tensor cores, while the Pascal part has none. Clock speeds are higher on the newer card: the P106-090 runs at 1354 MHz base and 1531 MHz boost, while the RTX PRO runs at 1590 MHz base and 2288 MHz boost, a boost advantage of 757 MHz. Memory clocks also differ, with the P106-090 at 2002 MHz (8 Gbps effective) and the RTX PRO at 1750 MHz (28 Gbps effective), the latter using faster GDDR7 versus GDDR5.

Compute throughput shows a massive gap. The P106-090 delivers 2.352 TFLOPS FP32 and 36.74 GFLOPS FP16 (with a 1:64 ratio), while the RTX PRO delivers 110.1 TFLOPS FP32 and 110.1 TFLOPS FP16 (with a 1:1 ratio). Pixel rate is 73.49 GPixel/s for the P106-090 versus 439.3 GPixel/s for the RTX PRO, and texture rate is 73.49 GTexel/s versus 1,720.6 GTexel/s. The RTX PRO’s FP16 performance is 2,996 times higher, reflecting a fundamental shift toward balanced FP16 compute rather than the Pascal-era crippled rate.

Interface and power delivery also diverge. The P106-090 uses a PCIe 1.0 x1 bus interface, which is an extreme bottleneck for data transfer, while the RTX PRO uses PCIe 5.0 x16. The P106-090 has a 75 W TDP with a 1x 6-pin power connector and a 250 W suggested PSU, while the RTX PRO has a 300 W TDP with a 1x 16-pin connector and a 700 W suggested PSU. Physical dimensions are similar in length, 250 mm versus 267 mm, but the RTX PRO is taller at 111 mm and thicker at 40 mm, both dual-slot designs. The P106-090 has no display outputs, while the RTX PRO has 4x DisplayPort 2.1b. API support is also newer on the Blackwell part, with DirectX 12 Ultimate (12_2) versus DirectX 12 (12_1), both supporting OpenGL 4.6 and Vulkan 1.4.

Head-to-Head Benchmarks

The database contains exactly one direct comparison between these two cards, the 3DMark Steel Nomad DX12 test. In that workload, the RTX PRO 6000D Blackwell Max-Q scores 11088, and the P106-090 scores 509, with a delta of -95.4% for the older card. In relative terms, the RTX PRO delivers a score that is approximately 21.8 times higher than the P106-090. This is the largest recorded gap between the two, and it reflects the combined effect of 31 times more shading units, 15.7 times more TMUs, 4 times more ROPs, 46.8 times higher FP32 throughput, and 9.3 times higher memory bandwidth.

The RTX PRO’s nearest rivals in the database include the NVIDIA RTX PRO 6000 Blackwell Max-Q with an identical average score of 11088 and 0% delta, the AMD Radeon RX 550 at 11075 with a 0.1% delta, the NVIDIA GeForce GTX 1650 SUPER at 11047 with a 0.4% delta, and the AMD FirePro W4300 at 11225 with a -1.2% delta. This places the RTX PRO 6000D in a cluster of cards scoring around 11,000 in Steel Nomad, suggesting its performance is comparable to those parts in this specific test, despite its much larger memory and compute capacity.

The P106-090’s nearest rivals are the NVIDIA GeForce GTX 570 at 13515 with a -0.3% delta, the AMD Radeon Pro 555 at 13407 with a 0.5% delta, the AMD Radeon HD 7770M at 13536 with a -0.5% delta, and the AMD Radeon RX 9070 XT at 13543 with a -0.5% delta. The P106-090’s average benchmark score of 13470 sits within 0.5% of all four rivals, indicating that its overall performance is roughly on par with those older and mid-range parts, even though its Steel Nomad score is much lower than its average. The P106-090’s average score is actually higher than its Steel Nomad result by a factor of about 26, meaning the OpenCL and Vulkan scores dominate its average, while the RTX PRO’s average is solely its Steel Nomad score.

The wins are entirely one-sided in the head-to-head. The RTX PRO wins the only shared test by a huge margin, and the P106-090 has no recorded wins against the RTX PRO. The P106-090’s Geekbench scores, while present, are not part of the head-to-head, so they do not constitute a win in the direct comparison framework.

The Verdict

From the recorded data, the NVIDIA RTX PRO 6000D Blackwell Max-Q is the superior choice for any workload that the shared benchmark represents. Its 11088 Steel Nomad score versus the P106-090’s 509 means a 95.4% advantage, and it carries newer architecture, massively more memory, higher compute rates, and full display support. The RTX PRO is active, has a launch MSRP of 8,565 USD, and is designed for workstation use with 96 GB of GDDR7 and 188 RT cores. Anyone needing DirectX 12 performance, large memory capacity, or display outputs should select the RTX PRO.

The P106-090, on the other hand, is a legacy mining part with no display outputs and end-of-life status. Its scores in Geekbench OpenCL (21304) and Geekbench Vulkan (18596) are the only measurements where it stands alone, but they do not compete with the RTX PRO in any direct test. Its average benchmark score of 13470 is close to its nearest rivals, all within 0.5%, which suggests it performs at a level similar to cards like the GeForce GTX 570 or Radeon RX 9070 XT in aggregate. For a user with a specific legacy OpenCL or Vulkan compute workload that matches the P106-090’s architecture, the card might still function, but it offers no measured advantage over the RTX PRO and lacks modern features such as RT cores, tensor cores, and a PCIe 5.0 interface.

The verdict is unambiguous. The RTX PRO 6000D Blackwell Max-Q wins the only head-to-head test by a factor of 21.8, and it wins on every architectural metric recorded. The P106-090 is not a competitor in any scenario that involves the shared benchmark, and its only unique data points are Geekbench results that the RTX PRO does not have recorded. Choose the RTX PRO for any modern compute, rendering, or display-capable task. Choose the P106-090 only if the workload is specifically tuned to Pascal-era OpenCL or Vulkan, and even then, its lack of display outputs and PCIe 1.0 x1 interface will limit usability. The data favors the Blackwell part in all direct comparisons.