Intel Arc B390 vs NVIDIA Rubin GPU Comparison

The Verdict

The Intel Arc B390 and NVIDIA Rubin GPU occupy entirely different segments of the hardware landscape, and the recorded data confirms they should never be cross-shopped. The Arc B390 is an integrated graphics processor (IGP) built for portable devices, drawing just 80 W and sharing system memory. The Rubin GPU is a 2300 W server-class accelerator in an SXM Module form factor, equipped with 288 GB of HBM4 memory and a 16384-bit bus. Benchmark results indicate the Arc B390 produces a 3DMark Steel Nomad DX12 score of 1482, placing it in the 9th percentile of all GPUs. The Rubin GPU has no recorded benchmark scores in the database, and its 50th percentile placement reflects an average score of 0, meaning no direct performance comparison is possible from measured data.

The decision between these two products is dictated by platform and workload, not by raw performance. The Arc B390 is the only option for a compact, low-power integrated solution with display outputs dependent on the portable device. The Rubin GPU is a discrete, PCIe 6.0 x16 accelerator with no display outputs, designed exclusively for server compute environments. The data shows the Arc B390 competes with legacy NVIDIA mobile and OEM parts, while the Rubin GPU has no nearest rivals recorded in the database. Anyone seeking a lightweight integrated graphics solution should select the Arc B390. Anyone requiring a massive, high-bandwidth server accelerator should select the Rubin GPU. There is no overlap in their intended use cases.

Architecture Differences

The Intel Arc B390 is built on Intel's 3 nm process and uses the Panther Lake chip with the Xe3-LPG architecture. It belongs to the Arc Graphics-M (Panther Lake) generation. The NVIDIA Rubin GPU is also fabricated on a 3 nm process, but at TSMC, and uses the GR100 chip with the Rubin architecture. It belongs to the Server Rubin (Rxx) generation. Both products share a 3 nm manufacturing node, but they are produced by different foundries: Intel for the Arc B390 and TSMC for the Rubin GPU.

The Arc B390 integrates 1536 shading units, 48 texture mapping units, 24 ROPs, and 12 ray tracing cores. It reports no tensor core count in the database. The Rubin GPU contains 28672 shading units, 896 TMUs, 24 ROPs, and 896 tensor cores, with no ray tracing core count recorded. The shading unit difference is substantial: the Rubin GPU has roughly 18.7 times the shading units of the Arc B390, based on the numbers in the database (28672 divided by 1536 equals approximately 18.7). The texture unit count is also far higher on the Rubin GPU, at 896 versus 48.

Memory architecture separates these two products completely. The Arc B390 uses system shared memory, with the memory size, type, bus width, and bandwidth all listed as system dependent. The Rubin GPU uses 288 GB of HBM4 memory on a 16384-bit bus, delivering 22.1 TB/s of bandwidth. The memory clock for the Rubin GPU is listed as 2695 MHz with 10.8 Gbps effective data rate. The Arc B390's memory clock is listed as system shared, meaning it relies on the host system's memory.

The Rubin GPU's transistor count is recorded as 336,000 million transistors on a 1456 mm² die, with a transistor density of 230.8M per mm². The Arc B390's transistor count and die size are both unknown in the database. The Rubin GPU requires a suggested power supply of 2700 W and has a TDP of 2300 W, while the Arc B390 has a TDP of 80 W and uses no power connectors. The Arc B390 has no suggested PSU listed because it is an integrated part. The Rubin GPU uses a PCIe 6.0 x16 bus interface, while the Arc B390 uses an IGP bus interface.

API support also diverges. The Arc B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The Rubin GPU lists N/A for DirectX, OpenGL, and Vulkan, which is consistent with a server accelerator that has no display outputs. The Arc B390's display outputs are listed as portable device dependent, while the Rubin GPU has no outputs at all.

Head-to-Head Benchmarks

The database contains no head-to-head benchmark entries between the Intel Arc B390 and the NVIDIA Rubin GPU. The head-to-head benchmark list is empty, and both the win count for the Arc B390 and the win count for the Rubin GPU are zero. Direct comparison is therefore impossible from measured results.

The only recorded benchmark for the Arc B390 is the 3DMark Steel Nomad DX12 test, where it scores 1482. This score places the Arc B390 in the 9th percentile of all GPUs. Its nearest rivals in the database are all legacy NVIDIA parts: the GeForce GT 520MX with an average score of 1463 and a delta of 1.3 percent, the GeForce 800M with an average score of 1460 and a delta of 1.5 percent, the GeForce GT 625 OEM with an average score of 1446 and a delta of 2.5 percent, and the GeForce GT 710 with an average score of 1443 and a delta of 2.7 percent. The data confirms the Arc B390 leads each of these rivals by small margins, from 1.3 percent over the GT 520MX to 2.7 percent over the GT 710.

The Rubin GPU has no recorded benchmarks and an average benchmark score of zero. Its percentile versus all GPUs is listed at 50, which is a neutral midpoint value, not a measured performance result. There are no nearest rivals recorded for the Rubin GPU. The absence of benchmark data means the Rubin GPU's computational capabilities, such as 130.0 TFLOPS FP32 and 260.0 TFLOPS FP16, cannot be validated against the Arc B390 or any other product in the database.

The Arc B390's pixel rate is 60.00 GPixel/s, and its texture rate is 120.0 GTexel/s. The Rubin GPU's pixel rate is 54.41 GPixel/s, and its texture rate is 2,031.2 GTexel/s. Despite having fewer ROPs, the Arc B390 posts a higher pixel rate than the Rubin GPU (60.00 versus 54.41 GPixel/s), which is notable given the Rubin GPU's massive shading and texture resources. The texture rate difference heavily favors the Rubin GPU, which delivers 2,031.2 GTexel/s compared to 120.0 GTexel/s for the Arc B390.

FAQ

Q: Which GPU has a higher FP32 performance?

A: The NVIDIA Rubin GPU delivers 130.0 TFLOPS of FP32 compute, while the Intel Arc B390 delivers 7.680 TFLOPS. The Rubin GPU's FP32 figure is approximately 16.9 times higher, based on the recorded values.

Q: What memory does each GPU use?

A: The Intel Arc B390 uses system shared memory, with size, type, bus width, and bandwidth all dependent on the host system. The NVIDIA Rubin GPU uses 288 GB of HBM4 memory on a 16384-bit bus with 22.1 TB/s of bandwidth.

Q: Do both GPUs support the same APIs?

A: No. The Intel Arc B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The NVIDIA Rubin GPU lists N/A for DirectX, OpenGL, and Vulkan, and it has no display outputs.

Q: What is the power consumption of each GPU?

A: The Intel Arc B390 has a TDP of 80 W and uses no power connectors. The NVIDIA Rubin GPU has a TDP of 2300 W and a suggested power supply of 2700 W.

Q: How does the Arc B390 compare to its nearest rivals?

A: The Arc B390 scores 1482 in 3DMark Steel Nomad DX12. It leads the GeForce GT 520MX by 1.3 percent, the GeForce 800M by 1.5 percent, the GeForce GT 625 OEM by 2.5 percent, and the GeForce GT 710 by 2.7 percent.

Q: Are there any benchmark results for the Rubin GPU?

A: The database records no benchmark scores for the Rubin GPU. Its average benchmark score is 0, and its 50th percentile placement is not based on measured performance.

Where Each One Wins

The Intel Arc B390 wins in integrated graphics scenarios. It is an IGP with a 3 nm Intel process, an 80 W TDP, and no power connectors. Its portable device dependent display outputs make it suitable for compact systems where a discrete GPU is not viable. The Arc B390 also wins in API compatibility for consumer graphics workloads, supporting DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Its pixel rate of 60.00 GPixel/s exceeds the Rubin GPU's 54.41 GPixel/s, despite the Rubin GPU's far larger silicon. The Arc B390's nearest rivals are all legacy NVIDIA parts, and it beats each by a small margin in the recorded benchmark.

The NVIDIA Rubin GPU wins in server compute and memory bandwidth. Its 288 GB of HBM4 memory with 22.1 TB/s of bandwidth and 16384-bit bus is a server-class configuration. The Rubin GPU's 130.0 TFLOPS FP32 and 260.0 TFLOPS FP16 figures dwarf the Arc B390's 7.680 TFLOPS and 15.36 TFLOPS. The 896 tensor cores on the Rubin GPU provide dedicated AI acceleration hardware, while the Arc B390 has no tensor core count recorded. The Rubin GPU's 896 TMUs deliver a texture rate of 2,031.2 GTexel/s, versus 120.0 GTexel/s for the Arc B390. The Rubin GPU also uses a PCIe 6.0 x16 bus interface, whereas the Arc B390 is limited to an IGP bus interface.

The Rubin GPU wins on raw compute resources: 28672 shading units, 336,000 million transistors, and a 1456 mm² die. The Arc B390 has 1536 shading units and an unknown transistor count and die size. The Rubin GPU's 2300 W TDP and 2700 W suggested PSU reflect its server-class power envelope, while the Arc B390's 80 W TDP suits low-power mobile platforms. The Rubin GPU has no display outputs and no consumer API support, so it wins exclusively in headless compute environments. The Arc B390 wins in any scenario requiring display output or consumer graphics API support.