Intel Arc A310E vs Intel Arc B390 Comparison

The Intel Arc A310E and Intel Arc B390 represent two very different approaches to Intel’s graphics lineup. The A310E is a discrete, low-profile card built on the Alchemist architecture for embedded and small-form-factor systems. The B390 is an integrated graphics processor (IGP) embedded within Intel’s Panther Lake mobile platform, using the newer Xe3-LPG architecture. The database contains only one recorded benchmark for the B390, while the A310E has no benchmark entries. This limits direct numerical comparison, but the available data still reveals significant performance and architectural gaps.

Head-to-Head Benchmarks

The only benchmark record in the database is for the Intel Arc B390: a 3DMark Steel Nomad DX12 score of 1482. The Intel Arc A310E has no benchmark entries in the database, so no direct head-to-head score can be cited. However, the B390’s result can be positioned against its nearest rivals, which are all NVIDIA integrated or low-end discrete parts. The B390 scores 1.3% higher than the NVIDIA GeForce GT 520MX (which averages 1463), 1.5% higher than the NVIDIA GeForce 800M (1460), 2.5% higher than the NVIDIA GeForce GT 625 OEM (1446), and 2.7% higher than the NVIDIA GeForce GT 710 (1443). These are narrow margins, all under 3%, indicating that the B390’s performance sits at the very bottom of the current GPU percentile ranking, at the 9th percentile among all GPUs.

Because the A310E lacks any benchmark score, the database offers no recorded performance delta between the two Intel parts. The A310E’s percentile is listed at 50, but with an average benchmark score of zero, this percentile is not backed by measured results. In contrast, the B390’s 9th percentile is tied to its single Steel Nomad score. The data shows that the B390, despite having more shading units and higher theoretical throughput, ranks extremely low in actual benchmark performance. The A310E’s percentile is effectively unverified, so any claim of superiority must rely on architectural specifications rather than measured outcomes.

Where Each One Wins

The Intel Arc B390 wins in raw compute potential based on recorded specifications. Its FP32 throughput is 7.680 TFLOPS, which is 2.5 times the A310E’s 3.072 TFLOPS. The B390 also delivers 15.36 TFLOPS for FP16 (2:1), versus 6.144 TFLOPS for the A310E. Pixel rate favors the B390 at 60.00 GPixel/s compared to 32.00 GPixel/s for the A310E, and texture rate is 120.0 GTexel/s versus 64.00 GTexel/s. These numbers suggest that in workloads that can fully use the integrated GPU’s resources, the B390 should outperform the A310E, but the benchmark data does not confirm this because the A310E has no recorded score.

The Intel Arc A310E wins in memory architecture and system integration. It has dedicated 4 GB of GDDR6 memory with a 64-bit bus and 124.0 GB/s bandwidth. The B390 uses system shared memory, with bandwidth listed as system dependent, meaning its performance is tied to the host platform’s memory speed and capacity. For applications that rely on consistent, dedicated memory bandwidth, the A310E has a clear advantage. The A310E also has a fixed 2000 MHz base and boost clock, whereas the B390 has a 300 MHz base clock and a 2500 MHz boost clock. The low base clock may impact sustained workloads if power or thermal limits prevent reaching boost.

The A310E also wins in form factor flexibility. It is a single-slot, 168 mm long, 69 mm high, 20 mm wide card with four mini-DisplayPort 2.0 outputs. The B390 is an IGP with no physical dimensions and its display outputs are portable device dependent. For systems requiring multiple external displays or a discrete card in a PCIe slot, the A310E is the only option. The B390 is limited to the host device’s built-in display capabilities.

Architecture Differences

The two GPUs come from different generations and foundries. The A310E uses the DG2-128 chip on the Xe-HPG architecture, with a 6 nm process from TSMC. The B390 uses a Panther Lake chip on the Xe3-LPG architecture, with a 3 nm process from Intel. The A310E is part of the Alchemist (Arc 3) generation, while the B390 belongs to Arc Graphics-M (Panther Lake). The process node difference is substantial: 6 nm versus 3 nm, which typically implies higher transistor density and lower power draw per transistor for the newer part, though the B390’s transistor count and die size are unknown.

Compute resources differ significantly. The A310E has 768 shading units, 32 texture mapping units, 16 ROPs, and 6 ray tracing cores. The B390 has 1536 shading units, 48 TMUs, 24 ROPs, and 12 ray tracing cores. The B390 doubles the shading units and ray tracing cores, and increases TMUs by 50% and ROPs by 50%. Neither part lists tensor cores. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API coverage is identical.

Memory architecture is a major divergence. The A310E uses dedicated GDDR6, while the B390 relies on system shared memory. The A310E’s memory clock is 1937 MHz, yielding 15.5 Gbps effective, with a fixed bandwidth of 124.0 GB/s. The B390’s memory speed and bandwidth are system dependent, meaning they vary with the host CPU’s memory controller and installed RAM. This makes the B390’s memory performance unpredictable without specific platform data.

Power and physical characteristics also differ. The A310E has a 75 W TDP and a suggested power supply of 250 W, uses no power connectors, and is single-slot. The B390 has an 80 W TDP, which is higher than the A310E despite being an integrated part, and has no suggested PSU because it draws power from the host platform. The B390 is listed as an IGP, so it has no slot width, no bus interface beyond the integrated path, and no dimensions.

The Verdict

From the recorded data, the Intel Arc B390 presents a paradox: it has far higher theoretical compute specifications but a benchmark score that places it at the 9th percentile, barely above obsolete NVIDIA parts from a decade earlier. The single 3DMark Steel Nomad DX12 score of 1482 is only 1.3% to 2.7% above the nearest rivals, all of which are low-end GPUs. This suggests that the B390’s integrated design, with its 300 MHz base clock and system shared memory, severely limits real-world performance despite the 2500 MHz boost clock and 7.680 TFLOPS FP32 rating.

The Intel Arc A310E has no benchmark data, so the database cannot confirm its performance. Its 50th percentile ranking is unverified, and its average benchmark score is zero. However, its dedicated GDDR6 memory and fixed 2000 MHz clocks provide a more predictable performance profile. For users who need a discrete, single-slot card with four mini-DisplayPort 2.0 outputs and a dedicated 4 GB memory pool, the A310E is the only choice. For users who require the highest theoretical FP32 throughput, the B390 offers 7.680 TFLOPS, but its benchmark result indicates that this translate poorly into actual 3D performance.

The data shows no head-to-head benchmark wins for either part, as the head-to-head list is empty and the wins counters are both zero. Therefore, the verdict rests on specifications and the one available benchmark. The B390 is the better choice for compute-heavy tasks that can leverage its 1536 shading units and 12 ray tracing cores, provided the host system has fast shared memory. The A310E is the better choice for applications that demand stable memory bandwidth, multiple display outputs, and a dedicated GPU that does not depend on system RAM. The B390’s 9th percentile ranking is a strong caution against expecting competitive gaming performance.

FAQ

Q: Which GPU has a higher FP32 compute throughput?

A: The Intel Arc B390 has 7.680 TFLOPS FP32, while the Intel Arc A310E has 3.072 TFLOPS, making the B390 2.5 times higher.

Q: What is the only recorded benchmark score in the database for these two GPUs?

A: The Intel Arc B390 has a 3DMark Steel Nomad DX12 score of 1482. The Intel Arc A310E has no recorded benchmark scores.

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

A: The B390 scores 1.3% higher than the NVIDIA GeForce GT 520MX, 1.5% higher than the NVIDIA GeForce 800M, 2.5% higher than the NVIDIA GeForce GT 625 OEM, and 2.7% higher than the NVIDIA GeForce GT 710.

Q: What memory configuration does each GPU use?

A: The A310E uses 4 GB of dedicated GDDR6 memory with a 64-bit bus and 124.0 GB/s bandwidth. The B390 uses system shared memory with a system dependent bus width and bandwidth.

Q: What are the power requirements for each GPU?

A: The A310E has a 75 W TDP and a suggested PSU of 250 W. The B390 has an 80 W TDP and no suggested PSU, as it is an integrated part.

Q: What is the process node for each GPU?

A: The A310E uses a 6 nm process from TSMC, while the B390 uses a 3 nm process from Intel.

Specification Differences

| Specification | Intel Arc A310E | Intel Arc B390 |

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

| Architecture | Xe-HPG | Xe3-LPG |

| Generation | Alchemist (Arc 3) | Arc Graphics-M (Panther Lake) |

| Process node | 6 nm (TSMC) | 3 nm (Intel) |

| Chip | DG2-128 | Panther Lake |

| Transistors | 7,200 million | Unknown |

| Die size | 157 mm² | Unknown |

| Shading units | 768 | 1536 |

| TMUs | 32 | 48 |

| ROPs | 16 | 24 |

| Ray tracing cores | 6 | 12 |

| Base clock | 2000 MHz | 300 MHz |

| Boost clock | 2000 MHz | 2500 MHz |

| Memory size | 4 GB GDDR6 | System Shared |

| Memory bus | 64 bit | System Shared |

| Memory bandwidth | 124.0 GB/s | System Dependent |

| FP32 | 3.072 TFLOPS | 7.680 TFLOPS |

| FP16 | 6.144 TFLOPS (2:1) | 15.36 TFLOPS (2:1) |

| Pixel rate | 32.00 GPixel/s | 60.00 GPixel/s |

| Texture rate | 64.00 GTexel/s | 120.0 GTexel/s |

| TDP | 75 W | 80 W |

| Slot width | Single-slot | IGP |

| Power connectors | None | None |

| Suggested PSU | 250 W | None |

| Bus interface | PCIe 4.0 x8 | IGP |

| Display outputs | 4x mini-DisplayPort 2.0 | Portable Device Dependent |

| Dimensions | 168 mm x 69 mm x 20 mm | None |

| Production status | End-of-life | Active |

| Release date | 2024-03-31 | 2026-01-26 |

| DirectX | 12 Ultimate (12_2) | 12 Ultimate (12_2) |

| OpenGL | 4.6 | 4.6 |

| Vulkan | 1.4 | 1.4 |

| Percentile vs all GPUs | 50 | 9 |

| Average benchmark score | 0 | 1482 |