Intel Arc A380E x2 vs Intel Arc A380M Comparison

The Verdict

The Intel Arc A380E x2 and Intel Arc A380M are built on the same DG2-128 silicon, the same Xe-HPG architecture, and the same Alchemist generation. The database records identical shader, texture, and pixel throughput for both cards, so raw compute capability is not the differentiator. The decisive factors are the power envelope and the physical form factor.

The Arc A380E x2 is a single-slot, 130 W desktop card with a 6-pin power connector and eight mini-DisplayPort 2.0 outputs. It requires a 300 W suggested power supply and uses a PCIe 4.0 x8 interface. Its production status is end-of-life. This card is for a system builder who needs a known desktop footprint, a fixed slot width, and a conventional power delivery path. The 130 W TDP is roughly 3.7 times the 35 W TDP of the mobile part, which means the desktop card can sustain its 2000 MHz boost clock without the thermal constraints of a compact module.

The Arc A380M is an MXM Module with a 35 W TDP, no power connector listed, no suggested PSU figure, and display outputs that depend on the portable device. It uses an MXM-A (3.1) bus interface. Its production status is active, and its release date is earlier in the database. This card is for an embedded or portable system where the host platform supplies power and display routing. The 35 W envelope is far more restrictive, and the base clock reflects that: 1550 MHz versus 2000 MHz on the desktop part.

Benchmark results are not yet recorded for either card, so the verdict rests on the specification differences. The Arc A380E x2 is the choice when the system has a 300 W PSU budget and a PCIe slot. The Arc A380M is the choice when the platform is an MXM socket and power is at a premium. The data does not support a compute performance advantage for either card, because the FP32 and FP16 figures are identical at 4.096 TFLOPS and 8.192 TFLOPS respectively. The only clock difference is the base clock; the boost clock is 2000 MHz on both.

Architecture Differences

Both cards share the same chip, DG2-128, and the same Xe-HPG architecture. The foundry is TSMC, and the process node is 6 nm. Transistor count is identical at 7,200 million, and die size is identical at 157 mm². Transistor density is 45.9M per mm² on both. The generation field differs slightly in labeling: the desktop card is listed as Alchemist (Arc 3), while the mobile card is listed as Alchemist (Arc 3 Mobile). This is a naming distinction, not a silicon distinction.

The shading array is the same: 1024 shading units, 64 TMUs, 32 ROPs, and 8 ray tracing cores. The memory subsystem is identical: 6 GB of GDDR6 on a 96-bit bus, with 186.0 GB/s of bandwidth and a memory clock of 1937 MHz (15.5 Gbps effective). API support matches on both: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The architectural differences are in clock behavior and integration. The desktop Arc A380E x2 has a base clock of 2000 MHz, which equals its boost clock. The mobile Arc A380M has a base clock of 1550 MHz and a boost clock of 2000 MHz. This means the desktop part runs at full speed at all times, while the mobile part must boost to reach the same peak. The pixel rate and texture rate are listed as identical (64.00 GPixel/s and 128.0 GTexel/s), which reflects the boost clock state, not the sustained base-clock state.

The physical integration differs substantially. The desktop card is single-slot, 265 mm long, 127 mm tall, and 20 mm wide. The mobile module has no recorded dimensions. The desktop card has eight mini-DisplayPort 2.0 outputs; the mobile card has outputs that depend on the portable device. The desktop card uses PCIe 4.0 x8; the mobile card uses MXM-A (3.1). The desktop card has a 1x 6-pin power connector; the mobile card has none listed. The desktop card has a suggested PSU of 300 W; the mobile card has no suggested PSU.

The production status differs: the desktop card is end-of-life, while the mobile card is active. The release dates differ: the desktop card released on 2024-03-31, and the mobile card on 2023-01-23. The desktop card lists its predecessor as Xe Graphics and its successor as Battlemage; the mobile card lists neither.

FAQ

Q: Do the two cards have the same compute performance?

A: The recorded FP32 throughput is 4.096 TFLOPS on both, and FP16 is 8.192 TFLOPS (2:1) on both. Pixel rate is 64.00 GPixel/s on both, and texture rate is 128.0 GTexel/s on both. The identical boost clock of 2000 MHz means peak compute is the same.

Q: Why does the desktop card have a higher TDP?

A: The Arc A380E x2 is rated at 130 W, while the Arc A380M is rated at 35 W. The desktop card also lists a suggested PSU of 300 W and a 6-pin power connector. The mobile card has no power connector or PSU figure, relying on the host platform.

Q: What is the memory configuration?

A: Both cards use 6 GB of GDDR6 on a 96-bit bus with 186.0 GB/s bandwidth. The memory clock is 1937 MHz, which is 15.5 Gbps effective on both.

Q: Which card supports which bus interface?

A: The Arc A380E x2 uses PCIe 4.0 x8. The Arc A380M uses MXM-A (3.1). These are not interchangeable.

Q: Are the display outputs the same?

A: No. The desktop card has 8x mini-DisplayPort 2.0 outputs. The mobile card has display outputs that depend on the portable device.

Q: What is the production status of each card?

A: The Arc A380E x2 is end-of-life, and the Arc A380M is active. The desktop card released on 2024-03-31; the mobile card released on 2023-01-23.

Q: Is the die size or transistor count different?

A: No. Both use the DG2-128 chip with 7,200 million transistors on a 157 mm² die at 6 nm from TSMC.

Specification Differences

The two cards differ in the following recorded fields:

• Base clock: 2000 MHz on the Arc A380E x2, 1550 MHz on the Arc A380M. Boost clock is 2000 MHz on both.
• TDP: 130 W on the desktop card, 35 W on the mobile card.
• Slot width: single-slot on the desktop card, MXM Module on the mobile card.
• Power connectors: 1x 6-pin on the desktop card, none listed on the mobile card.
• Suggested PSU: 300 W on the desktop card, none listed on the mobile card.
• Bus interface: PCIe 4.0 x8 on the desktop card, MXM-A (3.1) on the mobile card.
• Display outputs: 8x mini-DisplayPort 2.0 on the desktop card, portable device dependent on the mobile card.
• Dimensions: 265 mm length, 127 mm height, 20 mm width on the desktop card; none recorded on the mobile card.
• Production status: end-of-life on the desktop card, active on the mobile card.
• Release date: 2024-03-31 on the desktop card, 2023-01-23 on the mobile card.
• Predecessor and successor: Xe Graphics and Battlemage on the desktop card, none on the mobile card.

Identical fields include: chip (DG2-128), architecture (Xe-HPG), generation (Alchemist, with a mobile suffix on the M), process node (6 nm), foundry (TSMC), transistors (7,200 million), die size (157 mm²), transistor density (45.9M / mm²), boost clock (2000 MHz), memory clock (1937 MHz, 15.5 Gbps effective), memory size (6 GB), memory type (GDDR6), bus width (96 bit), bandwidth (186.0 GB/s), shading units (1024), TMUs (64), ROPs (32), RT cores (8), pixel rate (64.00 GPixel/s), texture rate (128.0 GTexel/s), FP32 (4.096 TFLOPS), FP16 (8.192 TFLOPS), DirectX (12 Ultimate 12_2), OpenGL (4.6), and Vulkan (1.4).

Head-to-Head Benchmarks

The database currently records no head-to-head benchmark entries for this pair. Both cards have zero wins in the head-to-head table, and both have an average benchmark score of zero. The percentile versus all GPUs is 50 for both cards, which places them at the median of the recorded GPU population, but this is a ranking position, not a measured performance delta.

Without recorded benchmark scores, the analysis must use the specification data. The most direct comparison is the base clock. The Arc A380E x2 runs at 2000 MHz base, which is 450 MHz higher than the Arc A380M base of 1550 MHz. That is a 29% higher base clock. In steady-state workloads where the mobile card cannot sustain its boost clock due to the 35 W envelope, the desktop card would hold a clear advantage. The desktop card's 130 W TDP provides the headroom to maintain the boost clock indefinitely, assuming the 300 W suggested PSU is used.

The memory subsystem is identical, so bandwidth-bound workloads show no difference. FP32 and FP16 throughput are identical, so compute-bound workloads show no difference at the same clock state. The pixel rate and texture rate are identical at the boost clock, which means rasterization throughput peaks at the same level on both. The difference is sustained operation: the desktop card has no base-to-boost gap, while the mobile card has a 450 MHz gap.

The physical and platform differences are also part of the comparison. The desktop card offers eight mini-DisplayPort 2.0 outputs, which is a very high output count for a single-slot card. The mobile card's outputs depend on the host device, so the achievable display configuration is not defined by the card itself. The bus interface difference, PCIe 4.0 x8 versus MXM-A (3.1), means the two cards cannot be swapped into the same system.

Where Each One Wins

The Arc A380E x2 wins in scenarios that demand sustained full-clock execution. Its 2000 MHz base clock equals its boost clock, so there is no clock ramp needed. The 130 W TDP and 300 W suggested PSU provide a clear power path. The single-slot form factor, with recorded dimensions of 265 mm by 127 mm by 20 mm, fits standard desktop chassis layouts. The 8x mini-DisplayPort 2.0 outputs make it suitable for multi-display desktop installations. The PCIe 4.0 x8 interface is a standard desktop bus. Its end-of-life status means it is a legacy option, but the specification list is complete.

The Arc A380M wins in scenarios where power is constrained. Its 35 W TDP is about 27% of the desktop card's 130 W TDP. It has no power connector and no suggested PSU, which means it draws from the host platform. The MXM-A (3.1) form factor suits portable or embedded devices where a full-height desktop card cannot fit. Its active production status means it remains available in the database. The boost clock of 2000 MHz matches the desktop card, so peak performance is available when the host platform can supply the power and cooling headroom.

Neither card wins on raw compute, because the recorded FP32, FP16, pixel rate, and texture rate are identical. Neither card wins on memory, because the 6 GB GDDR6, 96-bit bus, and 186.0 GB/s bandwidth are identical. The selection is purely about platform fit and power delivery. The desktop card is the right part for a fixed desktop system with a 300 W PSU budget. The mobile card is the right part for an MXM-based portable system with a 35 W power budget. The data does not indicate a performance hierarchy between them; it indicates a physical and electrical separation.