NVIDIA P106-100 vs NVIDIA RTX PRO 4000 Blackwell Comparison

Where Each One Wins

The performance split between these two NVIDIA parts is not close. The NVIDIA RTX PRO 4000 Blackwell wins every benchmark category where both cards have recorded data, and the margins are substantial. The RTX PRO 4000 Blackwell takes the win in the 3DMark Steel Nomad DX12 test with a score of 4648 against the P106-100's 899, a 417% advantage. In the Geekbench Vulkan test, the RTX PRO 4000 Blackwell scores 194168 versus 32897 for the P106-100, a 490.2% lead.

The P106-100 does not win any head-to-head comparison in the database. Its only recorded benchmarks are 3DMark Steel Nomad DX12, Geekbench OpenCL, and Geekbench Vulkan. Of those, only two overlap with the RTX PRO 4000 Blackwell's test suite, and it loses both. The P106-100's Geekbench OpenCL score of 35951 has no corresponding OpenCL result for the RTX PRO 4000 Blackwell, so it cannot be directly compared in that test.

The RTX PRO 4000 Blackwell also holds a decisive advantage in the broader benchmark suite. Its Passmark G3D score is 28427, while the P106-100 has no recorded Passmark G3D result. The RTX PRO 4000 Blackwell's Passmark GPU Compute score is 14805, again with no equivalent P106-100 data point. The P106-100's 68th percentile ranking among all GPUs places it near the RTX PRO 4000 Blackwell's 72nd percentile, but that percentile similarity masks the enormous absolute performance gap between the two cards. The average benchmark score tells a clearer story: 27135 for the RTX PRO 4000 Blackwell versus 23249 for the P106-100.

Architecture Differences

The two cards come from entirely different architectural eras. The RTX PRO 4000 Blackwell uses the GB203 chip built on Blackwell 2.0 architecture, fabricated on a 5 nm process at TSMC. The P106-100 uses the GP106 chip on the older Pascal architecture, fabricated on a 16 nm process, also at TSMC. The transistor counts reflect this generational leap: the GB203 packs 45,600 million transistors on a 378 mm² die, while the GP106 has 4,400 million transistors on a 200 mm² die. Transistor density is 120.6 million per mm² for the RTX PRO 4000 Blackwell versus 22.0 million per mm² for the P106-100.

The compute resources differ by an order of magnitude. The RTX PRO 4000 Blackwell has 8960 shading units, 280 texture mapping units, and 96 ROPs. The P106-100 has 1280 shading units, 80 TMUs, and 48 ROPs. The RTX PRO 4000 Blackwell also includes 70 ray tracing cores and 280 tensor cores, while the P106-100 has no ray tracing cores and no tensor cores at all. The FP32 throughput is 36.83 TFLOPS for the RTX PRO 4000 Blackwell versus 4.375 TFLOPS for the P106-100. FP16 performance shows an even larger split: 36.83 TFLOPS (1:1 ratio) for the RTX PRO 4000 Blackwell versus 68.36 GFLOPS (1:64 ratio) for the P106-100.

Memory configurations also diverge sharply. The RTX PRO 4000 Blackwell has 24 GB of GDDR7 memory on a 192-bit bus, with 672.0 GB/s of bandwidth. The P106-100 has 6 GB of GDDR5 memory on a 192-bit bus, with 192.2 GB/s of bandwidth. Both cards use a 192-bit memory interface, but the newer memory technology and larger capacity give the RTX PRO 4000 Blackwell a 3.5x bandwidth advantage. Clock speeds tell a more nuanced story: the P106-100 actually has a higher base clock at 1506 MHz versus 1230 MHz for the RTX PRO 4000 Blackwell, but the RTX PRO 4000 Blackwell boosts to 2055 MHz versus 1709 MHz for the P106-100. The memory clocks are 1750 MHz (28 Gbps effective) for the RTX PRO 4000 Blackwell versus 2002 MHz (8 Gbps effective) for the P106-100.

The RTX PRO 4000 Blackwell supports DirectX 12 Ultimate (12_2), while the P106-100 caps at DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. The RTX PRO 4000 Blackwell connects via PCIe 5.0 x16, while the P106-100 uses PCIe 1.0 x16, a significant interface generation gap. The RTX PRO 4000 Blackwell has four DisplayPort 2.1b outputs, while the P106-100 has no display outputs at all, reflecting its mining-focused design. The P106-100 is end-of-life, while the RTX PRO 4000 Blackwell is still active production. The RTX PRO 4000 Blackwell is a single-slot card with a 1x 16-pin power connector and 140 W TDP; the P106-100 is dual-slot with a 1x 6-pin connector and 120 W TDP. Both suggest a 300 W power supply.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 test is the most demanding workload in the shared suite. The RTX PRO 4000 Blackwell scores 4648, which is 417% higher than the P106-100's 899. This test stresses DX12 feature sets, and the RTX PRO 4000 Blackwell's Blackwell 2.0 architecture with DirectX 12 Ultimate support gives it a structural advantage over the Pascal-era P106-100 with its DirectX 12 (12_1) feature level. A 417% delta is not a marginal improvement; it represents a completely different performance class.

The Geekbench Vulkan benchmark shows an even larger relative gap. The RTX PRO 4000 Blackwell posts 194168, while the P106-100 manages 32897. The 490.2% delta suggests the Vulkan workload scales exceptionally well with the RTX PRO 4000 Blackwell's 8960 shading units and modern driver optimization. The P106-100's Vulkan support exists, but the architecture was not designed for the compute-heavy Vulkan paths that modern benchmarks exercise.

The RTX PRO 4000 Blackwell's average benchmark score of 27135 places it just below the NVIDIA GeForce RTX 3090, which averages 27565, a 1.6% gap. It also sits close to the AMD Radeon RX 6700 XT at 27425 (1.1% behind) and the NVIDIA GeForce RTX 4070 Mobile at 27435 (1.1% behind). Against the NVIDIA RTX A4000, the RTX PRO 4000 Blackwell is 1.7% ahead with 26683 for the A4000. The P106-100's average of 23249 ties with the AMD Radeon Pro Vega 16 at 23250, and it sits 0.1% behind the AMD Radeon RX 6600M at 23273 and the AMD Radeon R9 M290X at 23276. The P106-100 is also 0.3% behind the AMD Radeon AI PRO R9700 at 23315.

The pixel and texture rates reinforce the compute gap. The RTX PRO 4000 Blackwell achieves 197.3 GPixel/s and 575.4 GTexel/s, while the P106-100 achieves 82.03 GPixel/s and 136.7 GTexel/s. These are not directly comparable benchmark scores but rather derived specifications, and they show the same pattern: the RTX PRO 4000 Blackwell has roughly 2.4x the pixel throughput and 4.2x the texture throughput.

The Verdict

The data supports only one conclusion for compute-intensive workloads: the NVIDIA RTX PRO 4000 Blackwell is in a different performance tier entirely. It wins both shared benchmarks by margins of 417% and 490.2%, and its architecture provides features the P106-100 simply does not have, including ray tracing cores, tensor cores, and DirectX 12 Ultimate support. The 24 GB GDDR7 memory with 672.0 GB/s bandwidth dwarfs the P106-100's 6 GB GDDR5 with 192.2 GB/s. For any modern workload that can use these resources, the RTX PRO 4000 Blackwell is the only rational choice.

The P106-100 is a mining-oriented card with no display outputs, so it cannot serve as a primary graphics solution for a workstation. Its Pascal architecture lacks the hardware acceleration features that modern professional applications increasingly require. Its 68th percentile ranking is respectable for a card from 2017, and its average score of 23249 is only 14% behind the RTX PRO 4000 Blackwell's 27135 in percentage terms, but that average is skewed by the P106-100's limited benchmark coverage. In the tests where both cards have results, the gap is enormous.

Buyers should choose the RTX PRO 4000 Blackwell if they need professional rendering, AI workloads, or any modern graphics application that leverages DirectX 12 Ultimate, Vulkan, or compute shaders. The P106-100 makes sense only for legacy compute tasks that do not require display output and where the 120 W power draw and older architecture are acceptable. The production status difference matters too: the RTX PRO 4000 Blackwell is active, while the P106-100 is end-of-life. From a pure performance-per-watt perspective, the RTX PRO 4000 Blackwell delivers 36.83 TFLOPS at 140 W, while the P106-100 delivers 4.375 TFLOPS at 120 W, a roughly 7.2x efficiency advantage for the newer card.

FAQ

Q: How much faster is the RTX PRO 4000 Blackwell in 3DMark Steel Nomad DX12?

A: The RTX PRO 4000 Blackwell scores 4648 versus 899 for the P106-100, a 417% advantage.

Q: Does the P106-100 have ray tracing or tensor cores?

A: No. The P106-100 has no ray tracing cores and no tensor cores. The RTX PRO 4000 Blackwell has 70 ray tracing cores and 280 tensor cores.

Q: What memory configurations do the two cards use?

A: The RTX PRO 4000 Blackwell has 24 GB of GDDR7 on a 192-bit bus with 672.0 GB/s bandwidth. The P106-100 has 6 GB of GDDR5 on a 192-bit bus with 192.2 GB/s bandwidth.

Q: Can the P106-100 be used as a display adapter?

A: No. The P106-100 has no display outputs. The RTX PRO 4000 Blackwell has 4x DisplayPort 2.1b outputs.

Q: How does the RTX PRO 4000 Blackwell compare to the NVIDIA GeForce RTX 3090?

A: The RTX PRO 4000 Blackwell has an average benchmark score of 27135, which is 1.6% behind the RTX 3090's 27565.

Q: What is the FP32 compute difference between the two cards?

A: The RTX PRO 4000 Blackwell delivers 36.83 TFLOPS FP32, while the P106-100 delivers 4.375 TFLOPS FP32, a roughly 8.4x difference.