Intel Arc Pro B390 vs NVIDIA GeForce RTX 4090 Max-Q Comparison

Head-to-Head Benchmarks

The recorded data shows no direct benchmark comparisons between the Intel Arc Pro B390 and the NVIDIA GeForce RTX 4090 Max-Q. Both entries sit at the 50th percentile against all GPUs in the database, and neither has an average benchmark score recorded. The head-to-head benchmark table is empty, with zero wins recorded for either part.

What can be established from the raw specifications is the scale of the gap. The RTX 4090 Max-Q delivers 28.31 TFLOPS of FP32 compute, while the Arc Pro B390 delivers 7.680 TFLOPS. That is roughly 3.7 times higher raw floating-point throughput for the NVIDIA part. The texture rate tells a similar story: 442.3 GTexel/s versus 120.0 GTexel/s, a 3.7x advantage. Pixel throughput is 163.0 GPixel/s against 60.00 GPixel/s, a 2.7x lead for the RTX part.

The shading unit count reinforces the compute disparity. NVIDIA packs 9728 shading units into the AD103 chip, compared to 1536 for Intel's Panther Lake integrated graphics. That is a 6.3x difference in raw shader count. The RTX 4090 Max-Q also fields 304 tensor cores and 76 ray tracing cores, while the Arc Pro B390 lists 12 ray tracing cores and no tensor core count at all. The memory subsystem widens the gap further. The RTX part uses 16 GB of GDDR6 on a 256-bit bus, delivering 576.0 GB/s of bandwidth. The Intel part relies on System Shared memory with bandwidth listed as System Dependent, meaning its performance scales with whatever memory configuration the host portable device provides.

These are not competing in the same performance class. The Arc Pro B390 is an integrated graphics processor on Intel's own Panther Lake die, whereas the RTX 4090 Max-Q is a discrete-class mobile GPU. The 80 W TDP is shared by both, but the execution resources behind that power envelope are vastly different.

Where Each One Wins

The Arc Pro B390 has no benchmark wins recorded in the database. There are no wins listed for the NVIDIA part either, because the head-to-head table is empty. The comparison must be made on architectural capability and specification deltas.

The RTX 4090 Max-Q wins every measurable compute category: shading units, texture rate, pixel rate, FP32 throughput, FP16 throughput, memory capacity, memory bandwidth, ray tracing cores, and tensor cores. It also uses a dedicated GDDR6 memory pool with 576.0 GB/s of bandwidth, which removes the dependency on system RAM performance. For workloads that stress raw throughput, rasterization, ray tracing, or AI acceleration via tensor cores, the NVIDIA part holds the advantage by a wide margin.

The Arc Pro B390 wins the clock speed race in one narrow sense: its boost clock is 2500 MHz versus 1455 MHz for the RTX part. Its base clock of 300 MHz is far lower than the NVIDIA base of 930 MHz, but the Intel part reaches a higher peak frequency. This does not translate into a performance win given the 6.3x shader count disadvantage, but it does indicate the Intel architecture is tuned for higher clock ramps within its 80 W envelope.

The Intel part also uses a 3 nm process node from Intel's own foundry, compared to the 5 nm TSMC node for NVIDIA. Smaller process geometry does not offset the massive resource disparity. The Arc Pro B390 is a system-integrated GPU with no power connectors and no separate bus interface, meaning it operates entirely within the host processor package. The RTX 4090 Max-Q connects via PCIe 4.0 x16, giving it a dedicated high-bandwidth path to the system.

Architecture Differences

The Arc Pro B390 is built on Intel's Panther Lake chip using the Xe3-LPG architecture. It belongs to the Arc Graphics-WM generation and is listed as the successor to HD Graphics-WM. The process node is 3 nm, fabricated at Intel's own foundry. Transistor count and die size are both listed as unknown in the database. The architecture supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The RTX 4090 Max-Q uses the AD103 chip with Ada Lovelace architecture, part of the GeForce 40-series and the GeForce 40 Mobile generation. It is fabricated on a 5 nm process at TSMC, with 45,900 million transistors on a 379 mm² die. The transistor density is 121.1M per mm². This part succeeds GeForce 30 Mobile and is succeeded by GeForce 50 Mobile in the database lineage.

Memory architecture separates the two completely. The Intel part shares system memory, with no dedicated VRAM, no memory type specified beyond System Shared, and bandwidth that depends on the host platform. The NVIDIA part has 16 GB of GDDR6 on a 256-bit bus with 576.0 GB/s of bandwidth. The memory clock is listed as 2250 MHz with 18 Gbps effective data rate.

The core layouts diverge sharply. Intel uses 1536 shading units, 48 texture mapping units, 24 ROPs, and 12 ray tracing cores. NVIDIA uses 9728 shading units, 304 TMUs, 112 ROPs, 76 ray tracing cores, and 304 tensor cores. The FP16 output tells a different architectural story: Intel achieves 15.36 TFLOPS with a 2:1 ratio to FP32, while NVIDIA lists 28.31 TFLOPS at a 1:1 ratio. This indicates Intel's Xe3-LPG can double throughput on half-precision workloads, while Ada Lovelace maintains the same rate for both formats.

Both parts share the same 80 W TDP, and both are listed with IGP slot width, no power connectors, and Portable Device Dependent display outputs. The bus interface differs: Intel uses IGP, NVIDIA uses PCIe 4.0 x16. The Intel part has no tensor core count listed, while NVIDIA includes 304. The release dates are far apart: the RTX 4090 Max-Q launched on 2023-01-02, and the Arc Pro B390 on 2026-01-26. Both are marked as Active production status.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The NVIDIA GeForce RTX 4090 Max-Q delivers 28.31 TFLOPS of FP32, compared to 7.680 TFLOPS for the Intel Arc Pro B390. The NVIDIA part holds a roughly 3.7x advantage.

Q: How much memory does each GPU use, and what type?

A: The RTX 4090 Max-Q uses 16 GB of GDDR6 on a 256-bit bus with 576.0 GB/s bandwidth. The Arc Pro B390 uses System Shared memory with System Dependent bandwidth, meaning it relies on the host system's RAM.

Q: Do both GPUs support the same graphics APIs?

A: Yes. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What is the process node for each chip?

A: The Intel Arc Pro B390 is manufactured on a 3 nm process at Intel's foundry. The NVIDIA GeForce RTX 4090 Max-Q is manufactured on a 5 nm process at TSMC.

Q: Which GPU has more ray tracing cores?

A: The RTX 4090 Max-Q has 76 ray tracing cores. The Arc Pro B390 has 12 ray tracing cores.

Q: Are both GPUs rated at the same power draw?

A: Yes, both are listed with an 80 W TDP. Neither uses external power connectors, and both are classified as IGP slot width.

The Verdict

The database shows a decisive specification gap between these two parts. The RTX 4090 Max-Q outclasses the Arc Pro B390 in every measured throughput category: FP32 is 3.7x higher, texture rate is 3.7x higher, pixel rate is 2.7x higher, shading units are 6.3x more numerous, and memory bandwidth is 576.0 GB/s versus a system-dependent figure. The NVIDIA part also brings 304 tensor cores and 76 ray tracing cores to the table, while Intel lists 12 ray tracing cores and no tensor core count.

The Arc Pro B390 does have a higher boost clock at 2500 MHz versus 1455 MHz, and it uses a smaller 3 nm process node. It also carries the advantage of being an integrated GPU with no separate memory pool to manage, which can simplify system design in portable devices. But those advantages do not offset the raw resource disparity. A 1536-shader GPU with shared memory cannot compete with a 9728-shader GPU with 16 GB of dedicated GDDR6 at 576.0 GB/s.

Users who need maximum graphics throughput, ray tracing performance, or AI acceleration should look to the RTX 4090 Max-Q. Users constrained to an integrated GPU solution on a Panther Lake platform, where the Arc Pro B390 is the only option, will get a functional DirectX 12 Ultimate part with a 2:1 FP16 advantage over its own FP32 rate, but the performance ceiling is far lower. The data does not support any scenario where the Intel part wins a direct performance comparison.

Specification Differences

| Specification | Intel Arc Pro B390 | NVIDIA GeForce RTX 4090 Max-Q |

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

| Architecture | Xe3-LPG | Ada Lovelace |

| Chip | Panther Lake | AD103 |

| Process Node | 3 nm | 5 nm |

| Foundry | Intel | TSMC |

| Transistors | unknown | 45,900 million |

| Die Size | unknown | 379 mm² |

| Base Clock | 300 MHz | 930 MHz |

| Boost Clock | 2500 MHz | 1455 MHz |

| Memory Size | System Shared | 16 GB |

| Memory Type | System Shared | GDDR6 |

| Memory Bus Width | System Shared | 256 bit |

| Memory Bandwidth | System Dependent | 576.0 GB/s |

| Shading Units | 1536 | 9728 |

| TMUs | 48 | 304 |

| ROPs | 24 | 112 |

| Ray Tracing Cores | 12 | 76 |

| Tensor Cores | null | 304 |

| FP32 Performance | 7.680 TFLOPS | 28.31 TFLOPS |

| FP16 Performance | 15.36 TFLOPS (2:1) | 28.31 TFLOPS (1:1) |

| Pixel Rate | 60.00 GPixel/s | 163.0 GPixel/s |

| Texture Rate | 120.0 GTexel/s | 442.3 GTexel/s |

| TDP | 80 W | 80 W |

| Bus Interface | IGP | PCIe 4.0 x16 |

| Release Date | 2026-01-26 | 2023-01-02 |

| Predecessor | HD Graphics-WM | GeForce 30 Mobile |

| Successor | null | GeForce 50 Mobile |

| Percentile vs All GPUs | 50 | 50 |

| Average Benchmark Score | 0 | 0 |