Intel Arc Graphics 128EU Mobile vs Intel Arc Pro B390 Comparison

The Verdict

The database places both parts at the 50th percentile among all GPUs, but that shared percentile masks two very different designs. The Intel Arc Graphics 128EU Mobile is a Meteor Lake integrated part built on a 10 nm process, and it consumes 28 W. The Intel Arc Pro B390 is a Panther Lake integrated part built on a 3 nm process, and it consumes 80 W. The B390 is the higher-performance part on paper: its FP32 throughput is 7.680 TFLOPS versus 4.608 TFLOPS for the 128EU, and its boost clock is 2500 MHz versus 2250 MHz. The B390 also adds 12 ray tracing cores, which the 128EU lacks entirely. For users who need raw compute and hardware ray tracing, the B390 is the clear choice. For users constrained by power and thermal limits, the 128EU delivers a moderate amount of performance at less than half the TDP. The data shows no benchmark results to separate them empirically, so the verdict rests on architectural and specification differences. The B390 wins on capability; the 128EU wins on efficiency.

Architecture Differences

The two parts come from different Intel process nodes and architectures. The Intel Arc Graphics 128EU Mobile uses the Xe-LPG architecture on a 10 nm process, while the Intel Arc Pro B390 uses the Xe3-LPG architecture on a 3 nm process. The 128EU is part of the Arc Graphics-M generation, associated with the Meteor Lake chip. The B390 is part of the Arc Graphics-WM generation, associated with the Panther Lake chip. The process node difference is substantial: 10 nm versus 3 nm. This accounts for the B390's higher boost clock (2500 MHz vs 2250 MHz) despite its higher TDP.

The shading unit count differs. The 128EU has 1024 shading units, while the B390 has 1536. However, the texture mapping unit (TMU) and raster operation pipeline (ROP) counts go the other way. The 128EU has 64 TMUs and 32 ROPs; the B390 has 48 TMUs and 24 ROPs. The B390's higher shading unit count drives its higher FP32 throughput, but its lower TMU and ROP counts mean its texture and pixel rates are actually lower. The 128EU delivers 144.0 GTexel/s and 72.00 GPixel/s; the B390 delivers 120.0 GTexel/s and 60.00 GPixel/s.

The B390 includes 12 ray tracing cores. The 128EU has no dedicated ray tracing hardware. This is a major architectural divergence: the B390 is built for workloads that need hardware-accelerated ray tracing, while the 128EU relies on software or lacks the feature entirely. Both parts support DirectX, but the B390 supports DirectX 12 Ultimate (12_2), while the 128EU supports DirectX 12 (12_1). OpenGL support is identical at 4.6, and Vulkan support is identical at 1.4.

Both parts use system-shared memory with system-dependent bandwidth, so memory size, type, and bus width are all system-shared. The B390 has no power connectors, and both are IGP slot-width parts. The B390's TDP is 80 W, which is nearly three times the 128EU's 28 W. The production status for both is Active. The 128EU's predecessor is HD Graphics-M, while the B390's predecessor is HD Graphics-WM.

Head-to-Head Benchmarks

The database contains no head-to-head benchmark results for these two parts. The winsA and winsB counters are both zero, and the headToHeadBenchmarks array is empty. This means the comparison must be made entirely from the specification sheet and architectural features.

The largest numerical advantage for the B390 is in FP32 compute. The B390 delivers 7.680 TFLOPS, which is 3.072 TFLOPS higher than the 128EU's 4.608 TFLOPS. In percentage terms, the B390 is 66.7% higher. The FP16 figures follow the same pattern: the B390 delivers 15.36 TFLOPS (2:1) versus 9.216 TFLOPS (2:1) for the 128EU, again a 66.7% advantage.

The boost clock also favors the B390. At 2500 MHz, it runs 11.1% higher than the 128EU's 2250 MHz. The shading unit count favors the B390 by 50% (1536 vs 1024). The ray tracing core count is a categorical difference: 12 cores on the B390, none on the 128EU.

However, the 128EU wins in texture and pixel throughput. Its texture rate of 144.0 GTexel/s beats the B390's 120.0 GTexel/s by 20%. Its pixel rate of 72.00 GPixel/s beats the B390's 60.00 GPixel/s by 20%. These wins come from the higher TMU and ROP counts, despite the lower shading unit count and lower clock.

The power draw is a major point of differentiation. The 128EU's 28 W TDP is 65% lower than the B390's 80 W TDP. For integrated parts that share power budgets with CPUs and other system components, this difference is significant. The 128EU can fit into a much tighter thermal envelope.

FAQ

Q: Which GPU has more shading units?

A: The Intel Arc Pro B390 has 1536 shading units, while the Intel Arc Graphics 128EU Mobile has 1024.

Q: Does either GPU support hardware ray tracing?

A: Yes, the Intel Arc Pro B390 has 12 ray tracing cores. The Intel Arc Graphics 128EU Mobile has no ray tracing cores listed.

Q: Which GPU has the higher boost clock?

A: The Intel Arc Pro B390 boosts to 2500 MHz. The Intel Arc Graphics 128EU Mobile boosts to 2250 MHz.

Q: What is the TDP difference between the two?

A: The Intel Arc Graphics 128EU Mobile has a 28 W TDP. The Intel Arc Pro B390 has an 80 W TDP.

Q: Which GPU has higher pixel throughput?

A: The Intel Arc Graphics 128EU Mobile has a pixel rate of 72.00 GPixel/s. The Intel Arc Pro B390 has a pixel rate of 60.00 GPixel/s.

Q: What DirectX versions do they support?

A: The Intel Arc Pro B390 supports DirectX 12 Ultimate (12_2). The Intel Arc Graphics 128EU Mobile supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4.

Where Each One Wins

The Intel Arc Pro B390 wins in compute-heavy workloads. Its FP32 throughput of 7.680 TFLOPS is 66.7% higher than the 128EU's 4.608 TFLOPS. This advantage extends to FP16, where it delivers 15.36 TFLOPS versus 9.216 TFLOPS. The B390's higher shading unit count (1536 vs 1024) and higher boost clock (2500 MHz vs 2250 MHz) support this compute advantage. For applications that rely on shader throughput, such as rendering, simulation, or compute acceleration, the B390 is the stronger part.

The B390 also wins in ray tracing workloads. Its 12 ray tracing cores provide dedicated hardware for this task, while the 128EU has none. The B390's DirectX 12 Ultimate support also enables features that the 128EU's DirectX 12 (12_1) cannot access. For games or applications that use ray-traced effects, the B390 is the only choice between the two.

The Intel Arc Graphics 128EU Mobile wins in texture and pixel throughput. Its 64 TMUs and 32 ROPs deliver 144.0 GTexel/s and 72.00 GPixel/s, respectively. The B390's 48 TMUs and 24 ROPs deliver 120.0 GTexel/s and 60.00 GPixel/s. For fill-rate-bound workloads, such as traditional rasterization at high resolutions, the 128EU has a 20% advantage in both metrics.

The 128EU also wins in power efficiency. Its 28 W TDP is 65% lower than the B390's 80 W TDP. For mobile or compact systems with limited cooling, the 128EU is the more practical part. The B390's higher TDP likely requires a larger thermal solution and a more substantial power delivery system.

The process node difference also matters. The B390's 3 nm process is newer than the 128EU's 10 nm process. This likely contributes to the B390's ability to reach a higher clock speed and pack more shading units into a similar or smaller die area, although the die size is listed as unknown for both parts.

Specification Differences

The two parts differ in nearly every major specification category. The process node is the most fundamental difference: the Intel Arc Graphics 128EU Mobile uses 10 nm, while the Intel Arc Pro B390 uses 3 nm. The architectures differ accordingly: Xe-LPG for the 128EU, Xe3-LPG for the B390. The chip names also differ: Meteor Lake for the 128EU, Panther Lake for the B390.

The shading unit count is 1024 for the 128EU and 1536 for the B390. The TMU count is 64 for the 128EU and 48 for the B390. The ROP count is 32 for the 128EU and 24 for the B390. The ray tracing core count is null for the 128EU and 12 for the B390.

The boost clock is 2250 MHz for the 128EU and 2500 MHz for the B390. The base clock is the same for both at 300 MHz. The pixel rate is 72.00 GPixel/s for the 128EU and 60.00 GPixel/s for the B390. The texture rate is 144.0 GTexel/s for the 128EU and 120.0 GTexel/s for the B390.

The FP32 throughput is 4.608 TFLOPS for the 128EU and 7.680 TFLOPS for the B390. The FP16 throughput is 9.216 TFLOPS (2:1) for the 128EU and 15.36 TFLOPS (2:1) for the B390.

The TDP is 28 W for the 128EU and 80 W for the B390. The power connector is null for the 128EU and "None" for the B390. The bus interface is "Ring Bus" for the 128EU and "IGP" for the B390. The DirectX version is 12 (12_1) for the 128EU and 12 Ultimate (12_2) for the B390. OpenGL and Vulkan versions are identical at 4.6 and 1.4, respectively.

The release dates differ: the 128EU released on 2023-12-13, while the B390 released on 2026-01-26. The predecessor names also differ: HD Graphics-M for the 128EU, HD Graphics-WM for the B390. The generation names are Arc Graphics-M (Meteor Lake) for the 128EU and Arc Graphics-WM (Panther Lake) for the B390. Both parts use system-shared memory with system-dependent bandwidth, and both are IGP slot-width parts with portable-device-dependent display outputs.