NVIDIA GeForce RTX 5090 D V2 vs NVIDIA RTX PRO 6000 Blackwell Comparison

Head-to-Head Benchmarks

The single benchmark in the head-to-head comparison is 3DMark Steel Nomad (DX12), and it produces an extremely tight result. The NVIDIA GeForce RTX 5090 D V2 scores 16,504 points, while the NVIDIA RTX PRO 6000 Blackwell scores 16,408 points. The delta is a mere 0.6% in favor of the GeForce card, which is nearly within run-to-run variance. The data shows a statistical tie for practical purposes, yet the GeForce RTX 5090 D V2 takes the win on paper. This is remarkable because the two cards are built on the same GB202 chip, the same Blackwell 2.0 architecture, and the same 5nm TSMC process node with identical transistor counts. The 0.6% advantage in this synthetic workload is far smaller than the specification gaps between the two cards would suggest. The RTX PRO 6000 Blackwell carries significantly higher compute resources—more shading units, more texture mapping units, more render output units, more ray tracing cores, and more tensor cores—yet it still trails marginally in this DX12 test. The GeForce card's higher base clock of 2017 MHz compared to 1590 MHz on the RTX PRO helps close the gap, and its boost clock of 2407 MHz versus 2617 MHz means the workstation card actually reaches higher peak frequencies under load. The data indicates that clock speed distribution and power delivery characteristics matter as much as raw core counts in this particular workload. In the nearestRivals comparison, the RTX 5090 D V2 sits 0.5% above the AMD Radeon PRO W7500 and 0.9% above the AMD Radeon RX 5700 XT, while the RTX PRO 6000 Blackwell is 0.3% above the RX 5700 XT and 0.4% above the AMD Radeon Pro 5600M. Both cards occupy the 59th percentile among all GPUs, confirming that they perform at essentially the same level in this specific benchmark despite their architectural differences.

Architecture Differences

Both cards share the same fundamental architecture: Blackwell 2.0 on the GB202 chip, fabricated by TSMC on a 5nm process node. The die size is identical at 750 mm², and both pack 92,200 million transistors, yielding a transistor density of 122.9M per mm². The memory type is also the same—GDDR7 with an effective data rate of 28 Gbps and a memory clock of 1750 MHz. The architectural divergence appears in the execution units and memory subsystem. The RTX PRO 6000 Blackwell features 24,064 shading units, 752 texture mapping units, and 192 render output units, compared to 21,760 shading units, 680 TMUs, and 176 ROPs on the GeForce RTX 5090 D V2. That represents a 10.6% advantage in shading units, a 10.6% advantage in TMUs, and a 9.1% advantage in ROPs for the workstation card. Ray tracing cores number 188 on the RTX PRO versus 170 on the GeForce, an increase of 10.6%. Tensor cores follow the same pattern: 752 on the RTX PRO versus 680 on the GeForce, again a 10.6% difference. The memory bus width is a major differentiator: the RTX PRO 6000 Blackwell uses a 512-bit interface with 96 GB of GDDR7, while the GeForce card uses a 384-bit interface with 24 GB of GDDR7. Memory bandwidth scales accordingly: 1.79 TB/s on the RTX PRO versus 1.34 TB/s on the GeForce, a 33.6% advantage. The compute rates reflect these differences. FP32 performance is 126.0 TFLOPS on the RTX PRO versus 104.8 TFLOPS on the GeForce, a 20.2% advantage. FP16 performance follows the same 1:1 ratio: 126.0 TFLOPS versus 104.8 TFLOPS. Pixel rate is 502.5 GPixel/s on the RTX PRO versus 423.6 GPixel/s on the GeForce, a 18.6% advantage. Texture rate is 1,968.0 GTexel/s versus 1,636.8 GTexel/s, a 20.2% advantage. The base clocks differ substantially—2017 MHz on the GeForce versus 1590 MHz on the RTX PRO—but boost clocks reverse the order: 2407 MHz on the GeForce versus 2617 MHz on the RTX PRO. Both cards are dual-slot designs with a 16-pin power connector, but the RTX PRO has a 600 W TDP and a 1000 W suggested PSU, while the GeForce has a 575 W TDP and a 950 W suggested PSU. Display outputs differ: the GeForce offers 1x HDMI 2.1b and 3x DisplayPort 2.1b, while the RTX PRO offers 4x DisplayPort 2.1b with no HDMI. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, and both use a PCIe 5.0 x16 bus interface. Physical dimensions are nearly identical in length and height—304 mm and 137 mm for both—but the GeForce is 48 mm wide while the RTX PRO is 40 mm wide.

FAQ

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

A: The NVIDIA GeForce RTX 5090 D V2 scores 16,504 points versus 16,408 points for the NVIDIA RTX PRO 6000 Blackwell, giving the GeForce card a 0.6% lead in this test.

Q: Do both cards use the same GPU chip?

A: Yes, both are built on the GB202 chip with Blackwell 2.0 architecture, fabricated by TSMC on a 5nm process node with 92,200 million transistors and a 750 mm² die size.

Q: How much memory does each card have, and what is the bandwidth difference?

A: The RTX PRO 6000 Blackwell has 96 GB of GDDR7 on a 512-bit bus with 1.79 TB/s bandwidth, while the RTX 5090 D V2 has 24 GB of GDDR7 on a 384-bit bus with 1.34 TB/s bandwidth. The RTX PRO has 33.6% more bandwidth.

Q: Which card has more shading units and ray tracing cores?

A: The RTX PRO 6000 Blackwell has 24,064 shading units and 188 ray tracing cores, compared to 21,760 shading units and 170 ray tracing cores on the RTX 5090 D V2—a 10.6% advantage in both categories.

Q: What are the boost clock speeds of the two cards?

A: The RTX PRO 6000 Blackwell boosts to 2617 MHz, while the RTX 5090 D V2 boosts to 2407 MHz. The base clocks are 1590 MHz and 2017 MHz respectively.

Q: What is the TDP and suggested PSU for each card?

A: The RTX PRO 6000 Blackwell has a 600 W TDP with a 1000 W suggested PSU, while the RTX 5090 D V2 has a 575 W TDP with a 950 W suggested PSU.

Specification Differences

| Specification | NVIDIA GeForce RTX 5090 D V2 | NVIDIA RTX PRO 6000 Blackwell |

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

| Base Clock | 2017 MHz | 1590 MHz |

| Boost Clock | 2407 MHz | 2617 MHz |

| Memory Size | 24 GB | 96 GB |

| Memory Bus Width | 384 bit | 512 bit |

| Memory Bandwidth | 1.34 TB/s | 1.79 TB/s |

| Shading Units | 21,760 | 24,064 |

| TMUs | 680 | 752 |

| ROPs | 176 | 192 |

| Ray Tracing Cores | 170 | 188 |

| Tensor Cores | 680 | 752 |

| Pixel Rate | 423.6 GPixel/s | 502.5 GPixel/s |

| Texture Rate | 1,636.8 GTexel/s | 1,968.0 GTexel/s |

| FP32 Performance | 104.8 TFLOPS | 126.0 TFLOPS |

| FP16 Performance | 104.8 TFLOPS | 126.0 TFLOPS |

| TDP | 575 W | 600 W |

| Suggested PSU | 950 W | 1000 W |

| Card Width | 48 mm | 40 mm |

| Display Outputs | 1x HDMI 2.1b, 3x DisplayPort 2.1b | 4x DisplayPort 2.1b |

| Release Date | 2025-08-14 | 2025-03-17 |

| Launch MSRP | 2,299 USD | 8,565 USD |

The Verdict

The benchmark data shows that the GeForce RTX 5090 D V2 and the RTX PRO 6000 Blackwell are virtually indistinguishable in the 3DMark Steel Nomad DX12 test, with a 0.6% delta favoring the GeForce card. Both sit at the 59th percentile among all GPUs. However, the RTX PRO 6000 Blackwell offers dramatically more memory capacity (96 GB versus 24 GB), a wider memory bus (512 bit versus 384 bit), higher bandwidth (1.79 TB/s versus 1.34 TB/s), and substantially more compute resources across every category—shading units, TMUs, ROPs, ray tracing cores, and tensor cores all exceed the GeForce by roughly 10%. The FP32 and FP16 compute rates are 20.2% higher on the RTX PRO. The GeForce card counters with a higher base clock (2017 MHz versus 1590 MHz) and a lower TDP (575 W versus 600 W), which explains how it keeps pace in the synthetic benchmark despite fewer execution units. The RTX PRO also has a higher boost clock (2617 MHz versus 2407 MHz). The data suggests that the GeForce card is optimized for the specific characteristics of this DX12 gaming workload, while the RTX PRO's additional resources may shine in workloads not captured by this single benchmark. The RTX PRO was released earlier, on 2025-03-17, while the GeForce arrived on 2025-08-14. Both cards are currently active in production, use the same 16-pin power connector, and share identical PCIe 5.0 x16 interfaces. The launch MSRP for the RTX PRO is substantially higher at 8,565 USD compared to 2,299 USD for the GeForce.

Where Each One Wins

The NVIDIA GeForce RTX 5090 D V2 wins in the 3DMark Steel Nomad DX12 benchmark, taking a 0.6% lead. It also holds advantages in base clock speed (2017 MHz versus 1590 MHz), lower TDP (575 W versus 600 W), lower suggested PSU requirement (950 W versus 1000 W), and a slimmer profile in terms of power draw despite being wider physically (48 mm versus 40 mm). The GeForce card also offers an HDMI 2.1b output in addition to three DisplayPort 2.1b connections, which may be relevant for consumer display setups.

The NVIDIA RTX PRO 6000 Blackwell wins in every raw compute category: FP32 performance (126.0 TFLOPS versus 104.8 TFLOPS), FP16 performance (126.0 TFLOPS versus 104.8 TFLOPS), pixel rate (502.5 GPixel/s versus 423.6 GPixel/s), and texture rate (1,968.0 GTexel/s versus 1,636.8 GTexel/s). It offers 4x the memory capacity (96 GB versus 24 GB), 33.6% more memory bandwidth (1.79 TB/s versus 1.34 TB/s), and a wider 512-bit bus. The RTX PRO also has more shading units (24,064 versus 21,760), more TMUs (752 versus 680), more ROPs (192 versus 176), more ray tracing cores (188 versus 170), more tensor cores (752 versus 680), and a higher boost clock (2617 MHz versus 2407 MHz). For workloads that scale with memory capacity or raw compute throughput—such as large dataset processing, high-resolution rendering, or AI inference—the RTX PRO's specifications provide a clear margin. The single benchmark result does not capture these advantages, so users with memory-intensive or compute-bound tasks should weigh the specification differences heavily. The data indicates that the GeForce card is tuned for gaming-style DX12 workloads, while the RTX PRO is built for professional compute scenarios that leverage its larger memory pool and higher throughput rates.