Intel Arc Pro B390 vs NVIDIA H200 NVL Comparison

Head-to-Head Benchmarks

The benchmark database contains no direct head-to-head comparison results between the Intel Arc Pro B390 and the NVIDIA H200 NVL. The H200 NVL has a recorded Geekbench OpenCL score of 334,891, while the Arc Pro B390 has no benchmark entries in the database. The H200 NVL sits at the 100th percentile among all GPUs, while the Arc Pro B390 is at the 50th percentile.

The H200 NVL's measured score places it within a competitive server GPU field. It trails the NVIDIA B200 by 3.1%, with the B200 scoring 345,482. The H200 NVL leads the AMD Instinct MI300X by 5.3%, which scores 317,994. Against the NVIDIA B300 SXM6 AC, the H200 NVL is 9.4% behind, with that part scoring 369,831. The H200 NVL is 13.2% ahead of the NVIDIA L40S, which scores 295,763.

These deltas show the H200 NVL operating in the upper tier of compute accelerators, not at the absolute top but close enough that the differences matter mainly for workloads that scale tightly with raw compute throughput. The Arc Pro B390, by contrast, has no measured benchmark data available, so its relative standing cannot be quantified from the database.

FAQ

Q: What is the recorded benchmark score for the NVIDIA H200 NVL?

A: The H200 NVL has a Geekbench OpenCL score of 334,891, which places it at the 100th percentile among all GPUs in the database.

Q: How does the H200 NVL compare to the nearest rivals in the database?

A: The H200 NVL is 3.1% behind the NVIDIA B200 (345,482), 5.3% ahead of the AMD Instinct MI300X (317,994), 9.4% behind the NVIDIA B300 SXM6 AC (369,831), and 13.2% ahead of the NVIDIA L40S (295,763).

Q: Does the Intel Arc Pro B390 have any benchmark results in the database?

A: No. The Arc Pro B390 has an empty benchmark array and an average benchmark score of 0, with no nearest rivals listed.

Q: What is the production status of both GPUs?

A: Both the Intel Arc Pro B390 and the NVIDIA H200 NVL have a production status of "Active."

Q: What are the release dates for these parts?

A: The Intel Arc Pro B390 was released on January 26, 2026. The NVIDIA H200 NVL was released on November 17, 2024.

Q: Which GPU has the higher FP32 compute throughput?

A: The NVIDIA H200 NVL delivers 60.32 TFLOPS FP32, compared to 7.680 TFLOPS for the Intel Arc Pro B390.

Architecture Differences

The Intel Arc Pro B390 uses the Panther Lake chip with the Xe3-LPG architecture, built on a 3 nm process at Intel's foundry. It is part of the Arc Graphics-WM (Panther Lake) generation. The NVIDIA H200 NVL uses the GH100 chip with the Hopper architecture, fabricated on a 5 nm process at TSMC. It belongs to the Server Hopper (Hxx) generation.

The H200 NVL contains 80,000 million transistors on an 814 mm² die, with a transistor density of 98.3M / mm². The Arc Pro B390 has unknown transistor count and die size in the database.

The H200 NVL features 16896 shading units, 528 TMUs, 24 ROPs, and 528 tensor cores. The Arc Pro B390 has 1536 shading units, 48 TMUs, 24 ROPs, and 12 ray tracing cores. The H200 NVL has no listed ray tracing cores or DirectX support, while the Arc Pro B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The H200 NVL has no API support listed for DirectX, OpenGL, or Vulkan.

The H200 NVL is a dual-slot card with an 8-pin EPS power connector and a suggested PSU of 1000 W. It uses a PCIe 5.0 x16 bus interface and has no display outputs. The Arc Pro B390 is an integrated graphics processor (IGP) with no power connectors, no slot width beyond the IGP designation, and display outputs described as "Portable Device Dependent."

The H200 NVL carries 141 GB of HBM3e memory on a 6144-bit bus, delivering 4.89 TB/s bandwidth. The Arc Pro B390 uses system shared memory with a system dependent bandwidth, meaning its memory characteristics are tied to the host platform.

The Verdict

The data indicates two fundamentally different products. The NVIDIA H200 NVL is a server accelerator with 141 GB of HBM3e memory, 60.32 TFLOPS FP32 performance, and a measured benchmark score at the 100th percentile. The Intel Arc Pro B390 is an integrated GPU with 7.680 TFLOPS FP32, no benchmark results, and a 50th percentile ranking.

The H200 NVL is positioned for compute-heavy server workloads where memory bandwidth and tensor throughput matter. Its 4.89 TB/s bandwidth and 528 tensor cores support large-scale matrix operations. The Arc Pro B390, with its 12 ray tracing cores and DirectX 12 Ultimate support, targets graphics rendering on mobile or portable devices, but its lack of benchmark data prevents any quantitative comparison.

The H200 NVL's nearest rivals show it competing within a narrow performance band. The 3.1% gap to the B200 and 5.3% lead over the MI300X place it in the same performance class as those accelerators. The 13.2% lead over the L40S widens the separation from lower-tier server parts.

The Arc Pro B390's 80 W TDP and integrated design contrast sharply with the H200 NVL's 600 W TDP and dual-slot footprint. The power envelope difference reflects the intended deployment: one as a low-power integrated solution, the other as a dedicated server accelerator.

Specification Differences

| Specification | Intel Arc Pro B390 | NVIDIA H200 NVL |

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

| Process Node | 3 nm | 5 nm |

| Foundry | Intel | TSMC |

| Transistors | Unknown | 80,000 million |

| Die Size | Unknown | 814 mm² |

| Base Clock | 300 MHz | 1365 MHz |

| Boost Clock | 2500 MHz | 1785 MHz |

| Memory Size | System Shared | 141 GB |

| Memory Type | System Shared | HBM3e |

| Memory Bus Width | System Shared | 6144 bit |

| Memory Bandwidth | System Dependent | 4.89 TB/s |

| Shading Units | 1536 | 16896 |

| TMUs | 48 | 528 |

| ROPs | 24 | 24 |

| Tensor Cores | None | 528 |

| FP32 Performance | 7.680 TFLOPS | 60.32 TFLOPS |

| FP16 Performance | 15.36 TFLOPS (2:1) | 120.6 TFLOPS (2:1) |

| TDP | 80 W | 600 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 8-pin EPS |

| Bus Interface | IGP | PCIe 5.0 x16 |

| Display Outputs | Portable Device Dependent | No outputs |

| DirectX Support | 12 Ultimate (12_2) | N/A |

| OpenGL Support | 4.6 | N/A |

| Vulkan Support | 1.4 | N/A |

| Release Date | 2026-01-26 | 2024-11-17 |

Where Each One Wins

The NVIDIA H200 NVL wins decisively in raw compute throughput. Its FP32 output of 60.32 TFLOPS is roughly eight times the 7.680 TFLOPS of the Arc Pro B390. FP16 performance shows a similar gap: 120.6 TFLOPS versus 15.36 TFLOPS. The H200 NVL's pixel rate of 42.84 GPixel/s trails the Arc Pro B390's 60.00 GPixel/s, but its texture rate of 942.5 GTexel/s far exceeds the 120.0 GTexel/s of the integrated part.

Memory capacity and bandwidth favor the H200 NVL overwhelmingly. The 141 GB HBM3e pool with 4.89 TB/s bandwidth supports large model weights and high-throughput data movement. The Arc Pro B390's system shared memory is dependent on the host platform, providing no fixed bandwidth guarantee.

The Arc Pro B390 wins in graphics API support. It offers DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, while the H200 NVL lists N/A for all three. The Arc Pro B390 also includes 12 ray tracing cores, a feature absent from the H200 NVL's specification list. Power consumption favors the Arc Pro B390 at 80 W versus 600 W, and its integrated nature eliminates the need for external power connectors or a dedicated PSU recommendation.

The H200 NVL's transistor count of 80,000 million on an 814 mm² die indicates a massive compute-oriented silicon, while the Arc Pro B390's smaller footprint is designed for energy-constrained portable devices. The H200 NVL's dual-slot form factor and PCIe 5.0 x16 interface suit rack-mounted server environments. The Arc Pro B390's IGP bus interface and portable-device-dependent display outputs target mobile platforms.

For compute workloads such as large-scale inference, training, or scientific simulation, the H200 NVL is the clear choice based on the recorded data. For graphics rendering with modern API support on low-power mobile systems, the Arc Pro B390 is the only option with those capabilities in this comparison.