AMD Instinct MI350P vs Intel Arc A380E Comparison

Where Each One Wins

The recorded data for the AMD Instinct MI350P and the Intel Arc A380E shows two processors built for completely different workloads, and the benchmark split reflects that. The MI350P is an accelerator with no display outputs, no graphics API support, and a massive memory pool, while the A380E is a conventional graphics card with full DirectX 12 Ultimate, Vulkan 1.4, and four DisplayPort 2.0 outputs. The win count is even at zero each in the head-to-head benchmarks, because the database has no overlapping test results for these two parts. That absence is itself informative: the MI350P targets compute-heavy environments, whereas the A380E targets rendering and display tasks.

For raw compute throughput, the MI350P dominates. It delivers 36.04 TFLOPS of FP32 and 36.04 TFLOPS of FP16 in a 1:1 ratio, which is 8.8 times higher FP32 than the A380E's 4.096 TFLOPS and 4.4 times higher FP16 than the A380E's 8.192 TFLOPS (which runs at a 2:1 ratio). Texture rate tells a similar story: 1,126.4 GTexel/s versus 128.0 GTexel/s, a 8.8 times advantage. Pixel rate, however, is a different matter. The MI350P records 0 MPixel/s, while the A380E manages 64.00 GPixel/s. The MI350P has no raster output units, so it cannot generate pixels at all. The A380E has 32 ROPs and can drive displays directly.

Memory capacity and bandwidth are also polar opposites. The MI350P has 144 GB of HBM3e on an 8192-bit bus, yielding 8.19 TB/s of bandwidth. The A380E has 6 GB of GDDR6 on a 96-bit bus, yielding 186.0 GB/s. That is a 24 times capacity difference and a 44 times bandwidth difference. For large dataset processing, large model inference, or high-bandwidth scientific workloads, the MI350P is the only viable choice. For desktop output, video decode, or light graphics acceleration, the A380E is the only one that can connect to a monitor.

Architecture Differences

The two chips come from different architectural lineages. The MI350P uses CDNA 4.0, AMD's compute-focused architecture, built on a 3 nm TSMC process. The A380E uses Xe-HPG, Intel's gaming and graphics architecture, built on a 6 nm TSMC process. The MI350P packs 73,000 million transistors into a 1190 mm² die, giving a transistor density of 61.3M per mm². The A380E has 7,200 million transistors on a 157 mm² die, for 45.9M per mm². The MI350P is a massive, multi-chip-module-class design; the A380E is a small, efficient discrete GPU.

The MI350P is built around the "MI350 128CU" chip, with 8192 shading units and 512 texture mapping units. It has zero ROPs, which explains its 0 MPixel/s pixel rate. It also has no ray tracing cores listed, no tensor cores listed, and no graphics API support (DirectX, OpenGL, and Vulkan are all marked N/A). This is a pure compute accelerator. The A380E uses the "DG2-128" chip, with 1024 shading units, 64 TMUs, 32 ROPs, and 8 ray tracing cores. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, making it a fully featured graphics card.

Clock behavior differs sharply. The MI350P has a base clock of 1000 MHz and a boost clock of 2200 MHz, a 120% boost range. The A380E has a flat 2000 MHz base and boost, meaning it runs at a constant frequency. Memory clocks also differ: the MI350P's HBM3e runs at 2000 MHz with 8 Gbps effective data rate, while the A380E's GDDR6 runs at 1937 MHz with 15.5 Gbps effective. The A380E's higher effective memory speed per pin is offset by its narrow 96-bit bus versus the MI350P's 8192-bit bus.

The bus interface reflects their roles. The MI350P uses PCIe 5.0 x16, while the A380E uses PCIe 4.0 x8. The MI350P requires a 600 W TDP, a dual-slot cooler, a 1x 16-pin power connector, and a 1000 W suggested PSU. The A380E draws 75 W, is single-slot, has no power connectors, and needs only a 250 W suggested PSU. Physically, the MI350P is 267 mm long, 111 mm tall, and 40 mm wide. The A380E is 254 mm long, 127 mm tall, and 20 mm wide. The MI350P is thicker and denser; the A380E is taller but slimmer.

FAQ

Q: Which processor has more memory?

A: The AMD Instinct MI350P has 144 GB of HBM3e memory, while the Intel Arc A380E has 6 GB of GDDR6. The MI350P offers 24 times the capacity.

Q: Can either card output to a display?

A: No. The MI350P has no display outputs at all. The A380E has 4x DisplayPort 2.0 outputs and is the only one of the two that can drive a monitor.

Q: What is the FP32 performance difference?

A: The MI350P delivers 36.04 TFLOPS of FP32, which is 8.8 times the A380E's 4.096 TFLOPS. The MI350P also has 8.8 times the texture rate at 1,126.4 GTexel/s versus 128.0 GTexel/s.

Q: Does the MI350P support ray tracing?

A: The database lists no ray tracing cores for the MI350P, and its DirectX, OpenGL, and Vulkan APIs are all marked N/A. The A380E has 8 ray tracing cores and supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4.

Q: Which card is more power efficient?

A: The A380E has a 75 W TDP and no power connectors, while the MI350P has a 600 W TDP and requires a 1x 16-pin connector with a 1000 W suggested PSU. The A380E is far lighter on power draw.

Q: What are the release dates?

A: The MI350P was released on May 6, 2026. The A380E was released on March 31, 2024, and its production status is marked as end-of-life.

Specification Differences

| Specification | AMD Instinct MI350P | Intel Arc A380E |

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

| Architecture | CDNA 4.0 | Xe-HPG |

| Generation | Instinct (MIx) | Alchemist (Arc 3) |

| Process node | 3 nm | 6 nm |

| Foundry | TSMC | TSMC |

| Transistors | 73,000 million | 7,200 million |

| Die size | 1190 mm² | 157 mm² |

| Transistor density | 61.3M / mm² | 45.9M / mm² |

| Base clock | 1000 MHz | 2000 MHz |

| Boost clock | 2200 MHz | 2000 MHz |

| Memory clock | 2000 MHz, 8 Gbps effective | 1937 MHz, 15.5 Gbps effective |

| Memory size | 144 GB | 6 GB |

| Memory type | HBM3e | GDDR6 |

| Memory bus width | 8192 bit | 96 bit |

| Memory bandwidth | 8.19 TB/s | 186.0 GB/s |

| Shading units | 8192 | 1024 |

| Texture mapping units | 512 | 64 |

| Raster output units | 0 | 32 |

| Ray tracing cores | N/A | 8 |

| Pixel rate | 0 MPixel/s | 64.00 GPixel/s |

| Texture rate | 1,126.4 GTexel/s | 128.0 GTexel/s |

| FP32 performance | 36.04 TFLOPS | 4.096 TFLOPS |

| FP16 performance | 36.04 TFLOPS (1:1) | 8.192 TFLOPS (2:1) |

| TDP | 600 W | 75 W |

| Slot width | Dual-slot | Single-slot |

| Power connectors | 1x 16-pin | None |

| Suggested PSU | 1000 W | 250 W |

| Bus interface | PCIe 5.0 x16 | PCIe 4.0 x8 |

| Display outputs | No outputs | 4x DisplayPort 2.0 |

| DirectX support | N/A | 12 Ultimate (12_2) |

| OpenGL support | N/A | 4.6 |

| Vulkan support | N/A | 1.4 |

| Length | 267 mm (10.5 inches) | 254 mm (10 inches) |

| Height | 111 mm (4.4 inches) | 127 mm (5 inches) |

| Width | 40 mm (1.6 inches) | 20 mm (0.8 inches) |

| Production status | Not specified | End-of-life |

| Release date | 2026-05-06 | 2024-03-31 |

| Predecessor | Radeon Instinct | Xe Graphics |

| Successor | Not specified | Battlemage |

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark entries for the MI350P versus the A380E, and both parts sit at the 50th percentile among all GPUs with an average benchmark score of zero. This means the comparison relies entirely on their recorded specifications rather than measured performance tests. The differences in those specifications are extreme enough to make the outcome of any hypothetical test predictable.

The largest win for the MI350P is memory bandwidth. At 8.19 TB/s, it offers 44 times the A380E's 186.0 GB/s. This is not a marginal gap; it is a different class of memory subsystem. The 8192-bit bus versus 96-bit bus explains most of that gap, with the MI350P's HBM3e running at a lower effective clock but with vastly more parallel lanes. Any workload that streams large data sets will be bottlenecked by the A380E's bandwidth long before the MI350P approaches its limit.

Compute throughput follows the same pattern. The MI350P's 36.04 TFLOPS FP32 is 8.8 times the A380E's 4.096 TFLOPS. In FP16, the MI350P again delivers 36.04 TFLOPS, while the A380E reaches 8.192 TFLOPS. The MI350P's 1:1 FP16 ratio means it does not halve throughput for half-precision work, whereas the A380E's 2:1 ratio means it doubles FP16 output relative to FP32. Even so, the MI350P remains 4.4 times ahead in FP16. Texture rate mirrors FP32: 1,126.4 GTexel/s versus 128.0 GTexel/s, an 8.8 times advantage for the MI350P.

The A380E wins in pixel throughput and display capability. Its 64.00 GPixel/s is effectively infinite compared to the MI350P's 0 MPixel/s, because the MI350P has no ROPs and cannot rasterize. The A380E's 32 ROPs, 8 ray tracing cores, and full graphics API stack make it functional for rendering tasks. The MI350P cannot execute any DirectX, OpenGL, or Vulkan workloads, so any graphics-oriented benchmark would be a non-starter for it.

Power and physical footprint also favor the A380E. Its 75 W TDP is 8 times lower than the MI350P's 600 W. It requires no external power connectors and only a 250 W suggested PSU, versus the MI350P's 1000 W suggestion. The A380E is single-slot and 20 mm wide, while the MI350P is dual-slot and 40 mm wide. For systems with limited space or power budgets, the A380E is the practical option.

The Verdict

The data points to a clear separation of roles. The AMD Instinct MI350P is a compute accelerator for server or workstation environments where massive memory capacity, extreme bandwidth, and high FP32/FP16 throughput matter. It has 144 GB of HBM3e, 8.19 TB/s of bandwidth, 36.04 TFLOPS of FP32, and 36.04 TFLOPS of FP16. It carries no display outputs and no graphics API support, so it cannot replace a conventional GPU in a desktop setup. Its 600 W TDP, dual-slot cooler, 16-pin connector, and 1000 W suggested PSU indicate it belongs in a dedicated compute chassis, not a typical consumer build.

The Intel Arc A380E is a conventional graphics card for systems that need display output, graphics acceleration, or light compute. It has 4x DisplayPort 2.0 outputs, DirectX 12 Ultimate, OpenGL 4.6, Vulkan 1.4, and 8 ray tracing cores. Its 6 GB of GDDR6 and 186.0 GB/s of bandwidth are modest, and its 4.096 TFLOPS FP32 is low by modern standards, but its 75 W TDP, no power connectors, single-slot design, and 250 W suggested PSU make it easy to integrate into small or low-power systems. Its production status is end-of-life, with Battlemage listed as its successor.

The choice between the two depends entirely on the workload. For machine learning training, large-scale simulations, or any task that benefits from 144 GB of memory and 8.19 TB/s of bandwidth, the MI350P is the only option of the pair. For desktop use, media playback, or graphics rendering with a display attached, the A380E is the only option of the pair. Neither can substitute for the other. The MI350P's 24 times memory capacity and 44 times bandwidth come with an 8 times higher power draw and no visual output. The A380E's pixel rate, display support, and low power draw come with a fraction of the compute throughput. The database records no overlapping benchmarks, which aligns with these parts serving distinct markets. Buyers should select based on whether the task requires compute density or graphics output, not on a direct performance comparison.