Intel Arc Pro B390 vs NVIDIA N1X 40SM Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark results between the Intel Arc Pro B390 and the NVIDIA N1X 40SM. Both entries list zero benchmark scores, zero wins for either part, and no nearest rivals with comparative delta percentages. Consequently, the analysis below relies solely on the architectural and specification data recorded for each product.

The Intel Arc Pro B390 posts a peak FP32 throughput of 7.680 TFLOPS, while the NVIDIA N1X 40SM delivers 24.02 TFLOPS. That is a 3.13x advantage for the NVIDIA part in raw single-precision compute. In FP16, the Intel part reaches 15.36 TFLOPS via a 2:1 ratio, while the NVIDIA part sustains 24.02 TFLOPS at a 1:1 ratio, making the NVIDIA part 1.56x faster in half-precision work.

Texture throughput tells a similar story. The NVIDIA N1X 40SM reaches 750.7 GTexel/s versus 120.0 GTexel/s for the Intel Arc Pro B390, a 6.26x margin. Pixel fill rate favors NVIDIA as well: 93.84 GPixel/s against 60.00 GPixel/s, a 1.56x lead. The NVIDIA part also carries 5120 shading units, 320 TMUs, and 40 ROPs, compared to 1536 shading units, 48 TMUs, and 24 ROPs on the Intel side.

Ray tracing hardware favors NVIDIA decisively. The N1X 40SM includes 40 RT cores and 160 tensor cores, while the Intel Arc Pro B390 has 12 RT cores and no tensor core count listed. The NVIDIA part also accesses a dedicated 128 GB LPDDR5X memory pool over a 256-bit bus, delivering 273.2 GB/s of bandwidth. The Intel part uses system shared memory with system-dependent bandwidth, meaning its memory performance cannot be fixed in the database.

Clock behavior differs meaningfully. The Intel part boosts to 2500 MHz from a 300 MHz base, a relatively wide frequency range. The NVIDIA part boosts to 2346 MHz from a 741 MHz base. Despite the lower boost clock, the NVIDIA part's larger execution resource pool produces far higher aggregate throughput.

Both entries occupy the 50th percentile among all GPUs in the database, and both carry an average benchmark score of zero, reflecting the absence of run data. The recorded specifications, however, indicate a substantial performance gap in compute, texture, pixel, and ray tracing workloads.

The Verdict

From the recorded data alone, the NVIDIA N1X 40SM holds a commanding lead in every measurable compute category. Its FP32 output is 3.13x higher, its FP16 output is 1.56x higher, its texture rate is 6.26x higher, and its pixel rate is 1.56x higher. Any application that scales with shading units, TMUs, or RT cores will favor the NVIDIA part by a wide margin.

The Intel Arc Pro B390 does have advantages in certain specific areas. Its base clock of 300 MHz is far lower than the NVIDIA part's 741 MHz, which may indicate a wider dynamic power range for idle or low-load states, though the database does not record power figures for the NVIDIA part. The Intel part also reports a 3 nm process node from Intel's own foundry, versus a 5 nm node from TSMC for the NVIDIA part, which suggests a denser transistor implementation per area, though transistor counts are listed as unknown for both.

For users prioritizing raw compute throughput, memory bandwidth, or ray tracing capability, the NVIDIA N1X 40SM is the clear choice based on the recorded specifications. For users who require a processor with a lower base clock and a newer process node, the Intel Arc Pro B390 presents a different trade-off, but the database shows no benchmark wins for either part to validate real-world performance.

The absence of DirectX, OpenGL, and Vulkan API entries for the NVIDIA part is notable. The Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, while the NVIDIA part lists N/A for all three. This suggests the NVIDIA part may target a different software ecosystem, possibly one that does not rely on conventional graphics APIs, though the database does not specify which workloads it is designed for.

Architecture Differences

The Intel Arc Pro B390 uses the Xe3-LPG architecture on the Panther Lake chip, belonging to the Arc Graphics-WM (Panther Lake) generation. It is fabricated on a 3 nm process at Intel's foundry. The NVIDIA N1X 40SM uses the Blackwell 2.0 architecture on the GB20B chip, belonging to the Blackwell IGP (N1x) generation, fabricated on a 5 nm process at TSMC. The NVIDIA die measures 382 mm², while the Intel die size is listed as unknown.

The Intel part integrates 1536 shading units, 48 TMUs, 24 ROPs, and 12 RT cores. It does not list tensor cores. The NVIDIA part carries 5120 shading units, 320 TMUs, 40 ROPs, 40 RT cores, and 160 tensor cores. The NVIDIA part's tensor core count is more than 13x the Intel part's RT core count, and its shading unit count is 3.33x higher.

Memory architecture diverges sharply. The Intel Arc Pro B390 uses system shared memory with a system-dependent bus width and bandwidth. The NVIDIA N1X 40SM has a dedicated 128 GB LPDDR5X pool on a 256-bit bus, delivering a fixed 273.2 GB/s. The NVIDIA memory clock is listed as 1067 MHz with 8.5 Gbps effective transfer, while the Intel memory clock is listed as system shared.

Power delivery differs as well. The Intel part has a TDP of 80 W, while the NVIDIA part's TDP is unknown. Both use IGP slot widths and no power connectors. The Intel part's bus interface is IGP, while the NVIDIA part uses PCIe 5.0 x16. Display outputs differ: the Intel part lists portable device dependent outputs, while the NVIDIA part lists a single HDMI port.

The NVIDIA part's API support is entirely absent, with DirectX, OpenGL, and Vulkan all listed as N/A. The Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. This difference likely reflects the NVIDIA part's IGP classification for a specific platform, possibly a workstation or embedded use case, rather than a general-purpose consumer graphics adapter.

Release dates differ by several months. The Intel Arc Pro B390 entered production status with a release date of January 26, 2026. The NVIDIA N1X 40SM has a release date of May 31, 2026. Both parts are listed as active production. The Intel part lists its predecessor as HD Graphics-WM, while the NVIDIA part has no predecessor recorded.

FAQ

Q: Which GPU has higher raw compute performance?

A: The NVIDIA N1X 40SM delivers 24.02 TFLOPS FP32, which is 3.13x the Intel Arc Pro B390's 7.680 TFLOPS. In FP16, the NVIDIA part also leads at 24.02 TFLOPS versus 15.36 TFLOPS.

Q: Does the Intel Arc Pro B390 support standard graphics APIs?

A: Yes, the Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The NVIDIA N1X 40SM lists N/A for all three APIs in the database.

Q: What memory configuration does each GPU use?

A: The Intel Arc Pro B390 uses system shared memory with system-dependent bandwidth. The NVIDIA N1X 40SM has 128 GB of LPDDR5X memory on a 256-bit bus, providing 273.2 GB/s of bandwidth.

Q: How do the process nodes compare?

A: The Intel Arc Pro B390 is fabricated on a 3 nm node at Intel, while the NVIDIA N1X 40SM uses a 5 nm node at TSMC. The NVIDIA die size is recorded as 382 mm², while the Intel die size is unknown.

Q: Which GPU has more ray tracing hardware?

A: The NVIDIA N1X 40SM has 40 RT cores and 160 tensor cores. The Intel Arc Pro B390 has 12 RT cores and no tensor core count listed.

Q: What are the boost clocks for each part?

A: The Intel Arc Pro B390 boosts to 2500 MHz from a 300 MHz base. The NVIDIA N1X 40SM boosts to 2346 MHz from a 741 MHz base.

Where Each One Wins

The NVIDIA N1X 40SM wins in every category where the database records comparable specifications. Its FP32 throughput of 24.02 TFLOPS is over three times the Intel part's 7.680 TFLOPS, making it the stronger choice for compute-heavy workloads such as simulation, rendering, or scientific processing that rely on single-precision floating point. Its FP16 output of 24.02 TFLOPS also exceeds the Intel part's 15.36 TFLOPS, though the margin narrows to 1.56x.

Texture-heavy workloads favor the NVIDIA part overwhelmingly. With 750.7 GTexel/s versus 120.0 GTexel/s, the NVIDIA part sustains more than six times the texture fill rate. This translates to advantages in tasks that sample many textures per frame or per compute pass. Its pixel rate of 93.84 GPixel/s is also 1.56x higher than the Intel part's 60.00 GPixel/s, benefiting fill-bound operations like high-resolution rasterization.

Ray tracing is strictly an NVIDIA advantage in this comparison. The N1X 40SM's 40 RT cores and 160 tensor cores provide hardware acceleration for ray-traced effects and tensor-based workloads, respectively. The Intel Arc Pro B390 offers 12 RT cores and no tensor core count, leaving it without an equivalent tensor compute path.

Memory bandwidth is another decisive NVIDIA win. The 273.2 GB/s from 128 GB of LPDDR5X on a 256-bit bus is a fixed, dedicated resource. The Intel part's system shared memory has no fixed bandwidth in the database, meaning its effective throughput depends entirely on the host platform. For workloads that move large data sets between compute and memory, the NVIDIA part's predictable bandwidth is a structural advantage.

The Intel Arc Pro B390 holds advantages in a few narrow areas. Its 3 nm process node is smaller than the NVIDIA part's 5 nm node, which may imply better transistor density per area, though transistor counts are unknown for both. Its base clock of 300 MHz is substantially lower than the NVIDIA part's 741 MHz, potentially enabling lower idle power draw, though the NVIDIA part's TDP is unknown and cannot be compared directly. The Intel part also supports a full set of conventional graphics APIs, while the NVIDIA part lists none, making the Intel part the only one of the two with recorded compatibility for DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4.

The Intel part's 80 W TDP is a recorded power figure, while the NVIDIA part's TDP is not listed. This makes the Intel part the only one with a known power envelope, which could matter for system-level integration decisions. However, without a TDP for the NVIDIA part, no direct efficiency comparison is possible from the data.

The NVIDIA part's 1x HDMI display output is concrete, while the Intel part's display outputs are listed as portable device dependent. For fixed installations requiring a single HDMI connection, the NVIDIA part is explicitly supported; for portable or embedded use, the Intel part's outputs depend on the host device.

In summary, the NVIDIA N1X 40SM wins on compute, texture, pixel, ray tracing, tensor, and memory bandwidth. The Intel Arc Pro B390 wins on process node size, base clock flexibility, API compatibility, and a known TDP. The database records zero benchmark wins for either part, so all conclusions here derive from the recorded specification sheet.