Intel Arc A380E x2 vs Intel Arc B390 Comparison

Head-to-Head Benchmarks

The benchmark data for this comparison is one-sided. The Intel Arc A380E x2 has no recorded benchmark scores in the database, giving it an average benchmark score of zero. The Intel Arc B390 has a single recorded result: a score of 1482 in the 3DMark Steel Nomad DX12 test. This asymmetry means every head-to-head comparison relies entirely on the B390's recorded performance and the A380E's absence of data.

The B390's 1482 score places it at the 9th percentile among all GPUs in the database. Its nearest rivals provide context for where this integrated part sits. The NVIDIA GeForce GT 520MX scores 1463, which puts the B390 1.3% ahead. The NVIDIA GeForce 800M scores 1460, placing the B390 1.5% ahead. The NVIDIA GeForce GT 625 OEM scores 1446, making the B390 2.5% faster. The NVIDIA GeForce GT 710 scores 1443, leaving the B390 2.7% ahead. These deltas are narrow, indicating the B390 performs in a cluster of low-end discrete and integrated graphics solutions from past generations.

The A380E x2 cannot be compared directly because it has no benchmark entries. The database records zero wins for the A380E and zero wins for the B390 in head-to-head tests, reflecting that no direct comparison benchmarks exist. What the data does show is that the B390, despite being an integrated graphics processor, delivers a measurable 3DMark result, while the A380E x2, a discrete dual-card configuration, lacks any recorded performance data.

Architecture Differences

The two Intel parts come from different architectural generations and are built on different manufacturing processes. The Intel Arc A380E x2 uses the DG2-128 chip based on Xe-HPG architecture, belonging to the Alchemist (Arc 3) generation. It is fabricated on a 6 nm process at TSMC, with 7,200 million transistors on a 157 mm² die, resulting in a transistor density of 45.9 million per square millimeter.

The Intel Arc B390 uses the Panther Lake chip based on Xe3-LPG architecture, belonging to the Arc Graphics-M (Panther Lake) generation. It is fabricated on a 3 nm process at Intel's own foundry. The transistor count and die size are recorded as unknown in the database, and no transistor density is available.

Clock behavior differs substantially. The A380E x2 runs at a fixed 2000 MHz for both base and boost clocks, with memory clocked at 1937 MHz (15.5 Gbps effective). The B390 has a base clock of 300 MHz and a boost clock of 2500 MHz. The B390's base clock is far lower, but its boost clock exceeds the A380E's fixed clock by 500 MHz.

The A380E x2 is a discrete card with 6 GB of GDDR6 memory on a 96-bit bus, delivering 186.0 GB/s of bandwidth. The B390 is an integrated graphics processor with system-shared memory, where bus width and bandwidth are system dependent. This fundamental difference means the A380E x2 has dedicated memory resources, while the B390 depends on the host system's memory configuration.

Compute resources also differ. The A380E x2 has 1024 shading units, 64 texture mapping units, and 32 raster operations pipelines. The B390 has 1536 shading units, 48 texture mapping units, and 24 raster operations pipelines. The B390 has more shading units by 512, but fewer texture units by 16 and fewer ROPs by 8. Ray tracing core counts favor the B390 with 12 cores versus 8 on the A380E x2.

The measured rates reflect these configurations. The A380E x2 achieves a pixel rate of 64.00 GPixel/s and a texture rate of 128.0 GTexel/s. The B390 achieves 60.00 GPixel/s and 120.0 GTexel/s. Despite having fewer ROPs and TMUs, the B390's pixel and texture rates are close to the A380E's, likely due to its higher boost clock.

Compute throughput favors the B390 significantly. The A380E x2 delivers 4.096 TFLOPS of FP32 performance and 8.192 TFLOPS of FP16 performance at a 2:1 ratio. The B390 delivers 7.680 TFLOPS of FP32 and 15.36 TFLOPS of FP16, also at a 2:1 ratio. The B390 is 87.5% ahead in FP32 throughput and 87.5% ahead in FP16 throughput.

Power characteristics differ notably. The A380E x2 is rated at 130 W TDP, requires a 300 W suggested power supply, and uses a single 6-pin power connector. The B390 is rated at 80 W TDP, has no power connectors, and has no suggested PSU rating, consistent with its integrated nature.

Physical characteristics are entirely different. The A380E x2 is a single-slot card measuring 265 mm in length, 127 mm in height, and 20 mm in width. It uses a PCIe 4.0 x8 bus interface and provides 8x mini-DisplayPort 2.0 outputs. The B390 has no recorded dimensions, uses an IGP bus interface, and its display outputs are portable device dependent.

Both parts support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The production status shows the A380E x2 is end-of-life, while the B390 is active. The A380E x2 was released in March 2024, with its predecessor listed as Xe Graphics and its successor as Battlemage. The B390 was released in January 2026, with no predecessor or successor recorded.

Where Each One Wins

The B390 wins on raw compute metrics. Its FP32 throughput of 7.680 TFLOPS is 87.5% higher than the A380E x2's 4.096 TFLOPS. Its FP16 throughput of 15.36 TFLOPS is 87.5% higher than the A380E x2's 8.192 TFLOPS. Its shading unit count of 1536 exceeds the A380E x2's 1024 by 50%. Its ray tracing core count of 12 exceeds the A380E x2's 8 by 50%. Its boost clock of 2500 MHz is 25% higher than the A380E x2's fixed 2000 MHz.

The A380E x2 wins on memory and bandwidth. Its dedicated 6 GB of GDDR6 memory with 186.0 GB/s bandwidth provides a fixed, dedicated pool that does not compete with system memory. The B390 uses system-shared memory, which is system dependent and cannot offer a fixed bandwidth figure.

The A380E x2 also wins on pixel and texture throughput. Its pixel rate of 64.00 GPixel/s exceeds the B390's 60.00 GPixel/s by 6.7%. Its texture rate of 128.0 GTexel/s exceeds the B390's 120.0 GTexel/s by 6.7%. These advantages come from its higher ROP count of 32 versus 24 and higher TMU count of 64 versus 48.

The A380E x2 wins on connectivity. It provides 8x mini-DisplayPort 2.0 outputs, which allows multi-display configurations. The B390's display outputs are portable device dependent, meaning its display capabilities depend entirely on the host device.

The A380E x2 wins on power delivery infrastructure. It has a defined 130 W TDP with a single 6-pin connector and a 300 W suggested PSU. The B390 has an 80 W TDP with no connectors, drawing power through the host system.

The B390 wins on process technology. Its 3 nm node at Intel is smaller than the A380E x2's 6 nm node at TSMC. The B390 also wins on power efficiency in terms of FLOPS per watt: 7.680 TFLOPS divided by 80 W gives 96 TFLOPS per kilowatt, while the A380E x2 gives 4.096 TFLOPS divided by 130 W, which is 31.5 TFLOPS per kilowatt. The B390 delivers over three times the FP32 throughput per watt.

The B390 wins on availability. Its production status is active, while the A380E x2 is end-of-life.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The Intel Arc B390 delivers 7.680 TFLOPS of FP32 performance, which is 87.5% higher than the Intel Arc A380E x2's 4.096 TFLOPS.

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

A: The B390 scores 1482 in 3DMark Steel Nomad DX12. It is 1.3% ahead of the NVIDIA GeForce GT 520MX (1463), 1.5% ahead of the NVIDIA GeForce 800M (1460), 2.5% ahead of the NVIDIA GeForce GT 625 OEM (1446), and 2.7% ahead of the NVIDIA GeForce GT 710 (1443).

Q: Does the A380E x2 have any recorded benchmark scores?

A: No. The database records no benchmarks for the A380E x2, giving it an average benchmark score of zero.

Q: What memory configurations do the two GPUs use?

A: The A380E x2 uses 6 GB of GDDR6 memory on a 96-bit bus with 186.0 GB/s bandwidth. The B390 uses system-shared memory, with bus width and bandwidth marked as system dependent.

Q: What are the power requirements for each GPU?

A: The A380E x2 has a 130 W TDP, uses a single 6-pin power connector, and requires a 300 W suggested power supply. The B390 has an 80 W TDP, uses no power connectors, and has no suggested PSU rating.

Q: Which GPU has more shading units and ray tracing cores?

A: The B390 has 1536 shading units and 12 ray tracing cores. The A380E x2 has 1024 shading units and 8 ray tracing cores.

The Verdict

The recorded data paints a clear picture for compute-heavy workloads. The Intel Arc B390 delivers 7.680 TFLOPS of FP32 performance, which is 87.5% ahead of the A380E x2. It also has 50% more shading units and 50% more ray tracing cores. Its 3DMark Steel Nomad DX12 score of 1482 places it at the 9th percentile of all GPUs, narrowly ahead of several NVIDIA parts in its vicinity.

The A380E x2 retains advantages in fixed memory resources and display connectivity. Its dedicated 6 GB GDDR6 pool with 186.0 GB/s bandwidth removes dependence on system memory. Its 8x mini-DisplayPort 2.0 outputs enable multi-display setups that the B390 cannot guarantee, since its outputs are portable device dependent.

The B390 is the more power-efficient part. It produces 7.680 TFLOPS within an 80 W TDP, while the A380E x2 produces 4.096 TFLOPS within 130 W. The B390's 3 nm process also represents a newer manufacturing generation than the A380E's 6 nm node.

The A380E x2 is end-of-life, while the B390 is active. For users selecting a new part, the B390 offers higher compute throughput, more shading units, more ray tracing cores, a newer architecture, active production status, and a lower TDP. The A380E x2 offers dedicated memory bandwidth and fixed multi-display outputs, but its end-of-life status and lack of recorded benchmark data limit its appeal in direct comparison.

The data indicates the B390 is the stronger overall choice for compute performance and efficiency. The A380E x2's strengths lie in its dedicated memory subsystem and display output flexibility, which matter for specific multi-display or memory-sensitive scenarios.

Specification Differences

| Specification | Intel Arc A380E x2 | Intel Arc B390 |

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

| Chip | DG2-128 | Panther Lake |

| Architecture | Xe-HPG | Xe3-LPG |

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

| Process Node | 6 nm | 3 nm |

| Foundry | TSMC | Intel |

| Transistors | 7,200 million | unknown |

| Die Size | 157 mm² | unknown |

| Base Clock | 2000 MHz | 300 MHz |

| Boost Clock | 2000 MHz | 2500 MHz |

| Memory Size | 6 GB | System Shared |

| Memory Type | GDDR6 | System Shared |

| Bus Width | 96 bit | System Shared |

| Bandwidth | 186.0 GB/s | System Dependent |

| Shading Units | 1024 | 1536 |

| TMUs | 64 | 48 |

| ROPs | 32 | 24 |

| RT Cores | 8 | 12 |

| Pixel Rate | 64.00 GPixel/s | 60.00 GPixel/s |

| Texture Rate | 128.0 GTexel/s | 120.0 GTexel/s |

| FP32 | 4.096 TFLOPS | 7.680 TFLOPS |

| FP16 | 8.192 TFLOPS (2:1) | 15.36 TFLOPS (2:1) |

| TDP | 130 W | 80 W |

| Slot Width | Single-slot | IGP |

| Power Connectors | 1x 6-pin | None |

| Suggested PSU | 300 W | null |

| Bus Interface | PCIe 4.0 x8 | IGP |

| Display Outputs | 8x mini-DisplayPort 2.0 | Portable Device Dependent |

| Production Status | End-of-life | Active |

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

| Predecessor | Xe Graphics | null |

| Successor | Battlemage | null |