Intel Arc B390 vs NVIDIA RTX PRO 5000 72 GB Blackwell Comparison

Intel Arc B390 and NVIDIA RTX PRO 5000 72 GB Blackwell occupy opposite ends of the graphics spectrum, with the database showing a single head-to-head benchmark where the NVIDIA part dominates. The Arc B390 is an integrated processor graphics solution built for efficiency, while the RTX PRO 5000 is a professional workstation card with 72 GB of GDDR7 memory. The recorded data shows a 546.2% score advantage for the NVIDIA card in 3DMark Steel Nomad DX12, a gap that reflects fundamental differences in positioning, hardware resources, and power envelopes.

Where Each One Wins

The benchmark data records only one common test, 3DMark Steel Nomad DX12, and the NVIDIA RTX PRO 5000 72 GB Blackwell wins it decisively. The NVIDIA card scores 9,579.5 points, while the Intel Arc B390 scores 1,482 points. This translates to an 84.5% deficit for the Intel part in that specific workload. The Arc B390 has zero wins across all recorded head-to-head benchmarks, while the RTX PRO 5000 takes one win.

The Intel Arc B390, as an integrated processor graphics solution, shows a percentile rank of 9 against all GPUs in the database, placing it below the vast majority of recorded graphics hardware. The NVIDIA RTX PRO 5000, by contrast, sits at the 37th percentile, a stronger relative position, though the database places its nearest rivals in an unusual range. The RTX PRO 5000's nearest rivals include the NVIDIA GeForce GTX 460 SE, which scores 6,389 and trails by 0.3%, and the AMD Radeon Vega 10 Mobile, which scores 6,476 and leads by 1.1%. The Intel Arc B390's nearest rivals are all older low-end NVIDIA parts, including the GeForce GT 520MX at 1,463 (1.3% behind) and the GeForce GT 710 at 1,443 (2.7% behind).

The use-case split is clear from the numbers. The RTX PRO 5000 wins the recorded graphics workload outright, and its additional benchmark entries in the database, such as PassMark G3D at 24,310 and PassMark GPU Compute at 15,667, confirm broad applicability across different test types. The Arc B390 has a single recorded benchmark, limiting its demonstrated scope to that one DX12 test.

Architecture Differences

The two parts share the same DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4 API support, but the underlying silicon could not be more different. The Intel Arc B390 uses the Panther Lake chip with Xe3-LPG architecture, built on a 3 nm process at Intel's foundry. The NVIDIA RTX PRO 5000 uses the GB202 chip with Blackwell 2.0 architecture, fabricated on a 5 nm process at TSMC. The NVIDIA part contains 92,200 million transistors on a 750 mm² die, with a transistor density of 122.9M per mm². The Intel part's transistor count and die size are recorded as unknown.

The Intel Arc B390 has 1,536 shading units, 48 texture mapping units, 24 ROPs, and 12 ray tracing cores. The NVIDIA RTX PRO 5000 has 14,080 shading units, 440 TMUs, 160 ROPs, and 110 ray tracing cores. The NVIDIA part also includes 440 tensor cores, a feature class that the Intel part does not list. The clock behavior differs sharply: the Arc B390 runs at a 300 MHz base and a 2,500 MHz boost, while the RTX PRO 5000 runs at a 1,740 MHz base and a 2,377 MHz boost. The Intel integrated part boosts higher but starts far lower, while the NVIDIA discrete card maintains a much higher floor.

Memory architecture is the most visible split. The Arc B390 uses system shared memory, with bandwidth described as system dependent and no dedicated VRAM. The RTX PRO 5000 carries 72 GB of GDDR7 on a 384-bit bus, delivering 1.34 TB/s of bandwidth and a 1,750 MHz memory clock (28 Gbps effective). The Intel part's memory clock is listed as system shared, reinforcing its role as an integrated solution that borrows from the host.

Power and physical design diverge as well. The Arc B390 has an 80 W TDP, is an IGP with no power connectors, and uses an integrated bus interface. The RTX PRO 5000 has a 300 W TDP, requires a single 16-pin power connector, uses a dual-slot cooler, and connects via PCIe 5.0 x16. The NVIDIA card measures 267 mm in length, 111 mm in height, and 40 mm in width, while the Intel part has no recorded dimensions.

FAQ

Q: Which GPU has the higher 3DMark Steel Nomad DX12 score?

A: The NVIDIA RTX PRO 5000 72 GB Blackwell scores 9,579.5 points, while the Intel Arc B390 scores 1,482 points. The NVIDIA part wins by 84.5%.

Q: What memory configurations do the two GPUs use?

A: The Intel Arc B390 uses system shared memory with system dependent bandwidth. The NVIDIA RTX PRO 5000 uses 72 GB of GDDR7 memory on a 384-bit bus with 1.34 TB/s bandwidth.

Q: How do the shading unit counts compare?

A: The NVIDIA RTX PRO 5000 has 14,080 shading units, compared to 1,536 on the Intel Arc B390. The NVIDIA part also has 440 TMUs and 160 ROPs against 48 TMUs and 24 ROPs on the Intel part.

Q: What is the TDP difference between the two cards?

A: The Intel Arc B390 has an 80 W TDP and uses no power connectors. The NVIDIA RTX PRO 5000 has a 300 W TDP and requires a single 16-pin power connector, with a suggested PSU of 700 W.

Q: Where does each GPU rank among all GPUs in the database?

A: The Intel Arc B390 sits at the 9th percentile, while the NVIDIA RTX PRO 5000 sits at the 37th percentile.

Q: What process nodes and architectures do they use?

A: The Intel Arc B390 uses the Panther Lake chip with Xe3-LPG architecture on a 3 nm process at Intel. The NVIDIA RTX PRO 5000 uses the GB202 chip with Blackwell 2.0 architecture on a 5 nm process at TSMC.

Specification Differences

The Intel Arc B390 and NVIDIA RTX PRO 5000 differ across nearly every recorded specification. The Intel part uses the Panther Lake chip with Xe3-LPG architecture, while the NVIDIA part uses GB202 with Blackwell 2.0. Process nodes are 3 nm for Intel versus 5 nm for NVIDIA. The NVIDIA card has 92,200 million transistors on a 750 mm² die with 122.9M transistors per mm²; the Intel card's figures are unknown.

Clock speeds show the Intel part boosting to 2,500 MHz from a 300 MHz base, while the NVIDIA part boosts to 2,377 MHz from a 1,740 MHz base. The memory systems are entirely different: system shared for Intel versus 72 GB GDDR7, 384-bit bus, and 1.34 TB/s for NVIDIA. Core counts favor NVIDIA heavily: 14,080 shading units versus 1,536, 440 TMUs versus 48, 160 ROPs versus 24, and 110 ray tracing cores versus 12. Tensor cores exist only on the NVIDIA part at 440.

Rates follow the core counts. The NVIDIA card delivers 380.3 GPixel/s pixel rate and 1,045.9 GTexel/s texture rate, while the Intel card delivers 60.00 GPixel/s and 120.0 GTexel/s. FP32 compute is 66.94 TFLOPS for NVIDIA versus 7.680 TFLOPS for Intel. FP16 is 66.94 TFLOPS (1:1) for NVIDIA versus 15.36 TFLOPS (2:1) for Intel.

Physical and power specifications differ completely. The Intel part is an IGP with 80 W TDP, no power connectors, and no dimensions recorded. The NVIDIA part is a dual-slot card at 300 W, uses one 16-pin connector, requires a 700 W PSU, uses PCIe 5.0 x16, and measures 267 mm by 111 mm by 40 mm. Display outputs are portable device dependent for Intel, while NVIDIA provides 4x DisplayPort 2.1b. Release dates also differ: the Intel part launched on 2026-01-26, while the NVIDIA part launched on 2025-10-20.

Head-to-Head Benchmarks

The database contains one head-to-head benchmark between these two GPUs: 3DMark Steel Nomad DX12. The NVIDIA RTX PRO 5000 scores 9,579.5 points, and the Intel Arc B390 scores 1,482 points. The delta is 84.5% in favor of NVIDIA, meaning the Intel part achieves only 15.5% of the NVIDIA score. This is not a close contest by any metric.

Beyond the head-to-head, the database records additional benchmarks for the NVIDIA part that illustrate the breadth of its performance. PassMark G3D shows 24,310 points, PassMark GPU Compute shows 15,667 points, and PassMark G2D shows 914 points. Legacy DirectX tests on the NVIDIA card include 311 in DirectX 9, 219 in DirectX 11, and 175 in DirectX 10, with DirectX 12 at 78. The Intel Arc B390 has no such additional entries, so its measured capability rests entirely on the single 3DMark result.

Rival comparisons in the database frame each card differently. The Arc B390's nearest rival is the NVIDIA GeForce GT 520MX at 1,463 points, just 1.3% behind. The GT 710 at 1,443 trails by 2.7%. These are ancient low-end parts, and the Arc B390's 9th percentile rank places it in their company. The RTX PRO 5000's nearest rivals include the GeForce GTX 460 SE at 6,389 (0.3% behind) and the Radeon Vega 10 Mobile at 6,476 (1.1% ahead). Despite its much higher raw score, the RTX PRO 5000 sits only at the 37th percentile because the database's distribution includes many faster cards.

The 3DMark result is the only direct comparison available, and it leaves no ambiguity. The NVIDIA card outclasses the Intel integrated part by a factor of roughly 6.46 in that workload. Every architectural advantage, from shading units to memory bandwidth, contributes to that margin.

The Verdict

The recorded data supports only one conclusion for those who need maximum graphics throughput: the NVIDIA RTX PRO 5000 72 GB Blackwell wins the measured workload outright, with an 84.5% lead in 3DMark Steel Nomad DX12. Its 14,080 shading units, 110 ray tracing cores, 440 tensor cores, 72 GB of GDDR7, and 1.34 TB/s of bandwidth place it in a professional class that the Intel part cannot approach. The NVIDIA card also holds a 37th percentile rank versus the 9th percentile for Intel, and its additional PassMark results show consistent performance across compute and legacy graphics tests.

The Intel Arc B390 is not a competitor to the RTX PRO 5000 in the benchmark data. It is an integrated processor graphics solution with an 80 W TDP, shared system memory, and a 3 nm process, designed for efficiency rather than raw performance. Its nearest rivals are parts like the GeForce GT 520MX and GT 710, which it slightly exceeds. Users who rely on the single recorded workload and need the highest score should select the NVIDIA card. Users constrained to an integrated solution with no discrete card option might consider the Arc B390, but the data shows no scenario where it matches the NVIDIA part's performance.

The verdict follows the numbers. For professional rendering, compute, or any workload represented by 3DMark Steel Nomad DX12, the RTX PRO 5000 delivers over six times the score of the Arc B390. For minimal-power integrated graphics, the Arc B390 exists in a different category entirely, with a 9th percentile rank and rivals from over a decade ago. The database records one head-to-head result, and that result is decisive.