Intel Arc Pro B390 vs NVIDIA GeForce RTX 5080 Mobile Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear hierarchy: the NVIDIA GeForce RTX 5080 Mobile outperforms the Intel Arc Pro B390 across every measurable category. The RTX 5080 Mobile holds an 81st percentile ranking among all GPUs in the database, while the Arc Pro B390 sits at the 50th percentile. This gap is reflected in the average benchmark score of 38,349 for the NVIDIA part, compared to a recorded average score of zero for the Intel part, which has no benchmark entries in the database.

The most decisive advantage comes in raw compute throughput. The RTX 5080 Mobile delivers 23.04 TFLOPS of FP32 performance, exactly three times the 7.680 TFLOPS offered by the Arc Pro B390. In FP16 workloads, the NVIDIA part maintains 23.04 TFLOPS at a 1:1 ratio, while the Intel part reaches 15.36 TFLOPS through a 2:1 shader packing scheme. This means the NVIDIA part leads by 50% even in half-precision tasks where the Intel architecture attempts to compensate through its dual-rate execution.

Memory bandwidth is another area of substantial divergence. The RTX 5080 Mobile accesses its dedicated 16 GB GDDR7 memory across a 256-bit bus, yielding 896.0 GB/s of bandwidth. The Arc Pro B390 relies on system shared memory with bandwidth described as system dependent, which places it at a structural disadvantage in any bandwidth-sensitive workload. Texture and pixel throughput follow the same pattern: the NVIDIA part produces 360.0 GTexel/s and 144.0 GPixel/s, while the Intel part manages 120.0 GTexel/s and 60.00 GPixel/s. The NVIDIA part is three times faster in texturing and 2.4 times faster in pixel fill.

The compute unit counts reinforce this disparity. The RTX 5080 Mobile carries 7,680 shading units, 240 texture mapping units, and 96 ROPs. The Arc Pro B390 has 1,536 shading units, 48 TMUs, and 24 ROPs. In each category, the NVIDIA part holds a fivefold advantage in shaders, a fivefold advantage in TMUs, and a fourfold advantage in ROPs. Ray tracing resources follow suit, with 60 RT cores on the NVIDIA side versus 12 on the Intel side, and 240 tensor cores on the NVIDIA part compared to no tensor core count listed for the Intel part.

Clock speeds tell a more nuanced story. The Intel Arc Pro B390 boosts to 2500 MHz from a 300 MHz base, while the RTX 5080 Mobile boosts to 1500 MHz from a 975 MHz base. The Intel part achieves a higher peak clock by 1,000 MHz, but this does not compensate for the massive difference in execution resources. The NVIDIA GPU relies on its wider architecture to deliver superior results despite the lower clock rate.

Where Each One Wins

The RTX 5080 Mobile wins in every benchmark category recorded in the database. Its PassMark G3D score of 27,711 represents the strongest synthetic gaming result, with PassMark GPU Compute at 12,134, GeekBench Vulkan at 169,754, and GeekBench OpenCL at 166,986. The 3DMark Steel Nomad DX12 score of 4,952 confirms DirectX 12 performance capability. PassMark legacy tests also favor the NVIDIA part, with DirectX 9 at 314, DirectX 11 at 253, DirectX 10 at 171, and DirectX 12 at 116. The PassMark G2D score of 1,095 indicates solid 2D acceleration.

The Arc Pro B390 has no benchmark scores in the database, meaning the data shows no workload category where the Intel part wins. Its wins are limited to architectural attributes: a higher boost clock, a smaller process node, and a lower base clock that suggests more aggressive power gating. The 3 nm Intel process node versus the 5 nm TSMC node for NVIDIA gives the Intel part a density advantage in theory, though the database records the NVIDIA transistor count at 45,600 million on a 378 mm² die, which yields a density of 120.6M per mm². The Intel transistor count and die size are listed as unknown, so no direct density comparison is possible.

The use-case split is therefore defined by what each part can service. The RTX 5080 Mobile, with its 16 GB dedicated GDDR7 memory and 896.0 GB/s bandwidth, targets high-end gaming and compute workloads that demand sustained memory throughput. The Arc Pro B390, with system shared memory and a 80 W TDP identical to the NVIDIA part, is positioned for integrated graphics duty in portable devices, where its 7.680 TFLOPS of FP32 throughput still represents a meaningful compute resource despite the deficit.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce RTX 5080 Mobile has an average benchmark score of 38,349, while the Intel Arc Pro B390 has no recorded benchmark scores, giving it an average of zero.

Q: How do the FP32 compute performances compare?

A: The RTX 5080 Mobile delivers 23.04 TFLOPS of FP32 throughput, which is exactly three times the 7.680 TFLOPS provided by the Arc Pro B390.

Q: What memory configurations do the two GPUs use?

A: The RTX 5080 Mobile uses 16 GB of dedicated GDDR7 memory on a 256-bit bus with 896.0 GB/s bandwidth. The Arc Pro B390 uses system shared memory with system dependent bandwidth.

Q: Are there any benchmark categories where the Intel GPU wins?

A: The database contains no benchmark scores for the Intel Arc Pro B390, so the recorded data shows no wins for the Intel part in any test category.

Q: What is the ray tracing resource difference?

A: The RTX 5080 Mobile has 60 RT cores and 240 tensor cores, while the Arc Pro B390 has 12 RT cores and no tensor core count listed.

Q: How do the process nodes differ?

A: The Intel Arc Pro B390 uses a 3 nm process from Intel, while the RTX 5080 Mobile uses a 5 nm process from TSMC.

Specification Differences

The two GPUs differ across nearly every recorded specification. The RTX 5080 Mobile uses the GB203 chip with Blackwell 2.0 architecture, while the Arc Pro B390 uses the Panther Lake chip with Xe3-LPG architecture. The NVIDIA part comes from the GeForce 50-series generation, whereas the Intel part belongs to the Arc Graphics-WM generation. Process nodes differ at 5 nm for NVIDIA and 3 nm for Intel, with the NVIDIA die measuring 378 mm² and holding 45,600 million transistors.

Memory configurations are entirely divergent. The RTX 5080 Mobile has 16 GB of GDDR7 on a 256-bit bus with 896.0 GB/s bandwidth, while the Arc Pro B390 uses system shared memory with no dedicated allocation. Clock speeds show a 975 MHz base and 1500 MHz boost for the NVIDIA part, compared to a 300 MHz base and 2500 MHz boost for the Intel part. The memory clock for NVIDIA is 1750 MHz with 28 Gbps effective speed, while the Intel memory clock is listed as system shared.

The compute unit counts favor NVIDIA decisively: 7,680 shading units versus 1,536, 240 TMUs versus 48, 96 ROPs versus 24, 60 RT cores versus 12, and 240 tensor cores versus none listed. FP32 throughput sits at 23.04 TFLOPS for NVIDIA and 7.680 TFLOPS for Intel, with FP16 at 23.04 TFLOPS for NVIDIA and 15.36 TFLOPS for Intel. The bus interface differs, with PCIe 5.0 x16 for NVIDIA and IGP for Intel. Both parts share an 80 W TDP, IGP slot width, and no power connectors.

Architecture Differences

The architectural split is fundamental. NVIDIA's Blackwell 2.0 architecture on the GB203 chip uses a 5 nm TSMC process, while Intel's Xe3-LPG architecture on the Panther Lake chip uses a 3 nm Intel process. The NVIDIA architecture is built around a large discrete GPU design with dedicated memory, evidenced by the 45,600 million transistors on a 378 mm² die. The Intel design integrates graphics into a system-on-chip arrangement, given its IGP bus interface and shared memory model.

The RTX 5080 Mobile includes 240 tensor cores, which enable AI-accelerated workloads, while the Intel part lists no tensor core count. Ray tracing hardware exists on both, but at different scales: 60 RT cores for NVIDIA versus 12 for Intel. The NVIDIA architecture supports a 1:1 FP16 ratio, meaning FP16 throughput equals FP32 throughput at 23.04 TFLOPS. The Intel architecture uses a 2:1 FP16 ratio, doubling its FP32 rate to reach 15.36 TFLOPS in half precision.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The release dates differ, with the RTX 5080 Mobile launched on April 1, 2025, and the Arc Pro B390 on January 26, 2026. The NVIDIA part succeeds the GeForce 40 Mobile series, while the Intel part succeeds the HD Graphics-WM. Display outputs are portable device dependent for both, reflecting their mobile and integrated positioning.

The Verdict

The data supports a decisive conclusion: the NVIDIA GeForce RTX 5080 Mobile is the superior performer in every recorded benchmark and specification category. Its 23.04 TFLOPS of FP32 compute, 896.0 GB/s memory bandwidth, and 16 GB of dedicated GDDR7 memory position it firmly for demanding workloads. The 81st percentile ranking and average score of 38,349 confirm its standing among all GPUs.

The Intel Arc Pro B390 offers a different value proposition based on its architectural characteristics. Its 3 nm process node represents a more advanced manufacturing technology, and its 2500 MHz boost clock is notably higher than the NVIDIA part's 1500 MHz. The 7.680 TFLOPS of FP32 throughput, while one third of the NVIDIA capability, still provides meaningful compute for an integrated part. The 12 RT cores and 15.36 TFLOPS of FP16 performance indicate capability in ray tracing and half-precision workloads, albeit at reduced scale.

Users requiring maximum graphics performance should select the RTX 5080 Mobile based on the benchmark evidence. The passmark G3D score of 27,711, GeekBench Vulkan score of 169,754, and 3DMark Steel Nomad score of 4,952 demonstrate strong real-world performance across multiple test suites. Users constrained to integrated graphics in portable devices will find the Arc Pro B390 offers a solid baseline, but the recorded data shows no scenario where it matches the NVIDIA part's output. The 80 W TDP shared by both parts means the performance gap cannot be attributed to power allocation; the RTX 5080 Mobile simply uses its resources more effectively within the same power envelope.