AMD Radeon 860M vs Intel Arc B580 Comparison

FAQ

Q: How do the two GPUs compare in raw compute benchmarks?

A: The Intel Arc B580 dominates in both recorded compute tests. In Geekbench OpenCL, the Arc B580 scores 92,821 versus 22,759 for the AMD Radeon 860M, a 75.5% advantage. In Geekbench Vulkan, the Arc B580 scores 109,672 against 30,043, a 72.6% lead.

Q: Which GPU has the higher overall benchmark average?

A: The AMD Radeon 860M has a higher average benchmark score of 26,401 compared to the Intel Arc B580's 23,021. This is despite the Arc B580 winning both head-to-head tests, because the averages include different test suites for each card.

Q: What is the performance percentile ranking for each GPU?

A: The AMD Radeon 860M sits at the 72nd percentile among all GPUs, while the Intel Arc B580 ranks at the 68th percentile. The 860M's percentile is higher despite the Arc B580's raw compute wins.

Q: Who are the closest rivals to each GPU in the database?

A: The Radeon 860M is closest to the NVIDIA GeForce MX550 (0.1% lower), NVIDIA GeForce RTX 5060 (0.3% higher), AMD Radeon RX 5700 XT 50th Anniversary (0.6% lower), and NVIDIA RTX A4000 (1.1% lower). The Arc B580 sits near the AMD Radeon RX 580 2048SP (0.2% lower), NVIDIA GeForce RTX 2080 (0.6% higher), NVIDIA GeForce RTX 3080 (0.7% lower), and NVIDIA P106-100 (1.0% lower).

Q: What are the power requirements for each card?

A: The Radeon 860M has a 15 W TDP and requires no power connectors, as it is an integrated graphics processor. The Arc B580 has a 190 W TDP, uses a single 8-pin power connector, and the database suggests a 450 W power supply.

Q: Which GPU has dedicated video memory?

A: The Intel Arc B580 has 12 GB of GDDR6 memory on a 192-bit bus with 456.0 GB/s bandwidth. The AMD Radeon 860M uses system-shared memory with a system-dependent bandwidth, meaning it borrows from the host system's RAM.

Architecture Differences

The AMD Radeon 860M and Intel Arc B580 represent fundamentally different design philosophies. The 860M is an integrated GPU built on the Krackan Point chip using RDNA 3.5 architecture, fabricated on TSMC's 4 nm process. It belongs to the Navi III IGP generation for Strix Point Mobile platforms. The Arc B580 is a discrete graphics card using the BMG-G21 chip with Xe2-HPG architecture, belonging to Intel's Battlemage generation at the Arc 5 tier, built on a 5 nm TSMC process.

The transistor counts tell a story of scale. Intel's BMG-G21 packs 19,600 million transistors on a 272 mm² die, yielding a density of 72.1 million transistors per square millimeter. The 860M's transistor count and die size are listed as unknown in the database, but its status as an integrated part of a mobile APU implies a much smaller footprint shared with CPU components.

Memory architecture separates these two drastically. The Arc B580 uses 12 GB of dedicated GDDR6 memory on a 192-bit interface, delivering 456.0 GB/s of bandwidth. The 860M relies entirely on system-shared memory, with bandwidth described as system dependent. This means the integrated GPU's performance scales with the host system's RAM speed and configuration, while the Arc B580 has fixed, dedicated bandwidth.

Compute resources differ by a factor of five in shading units. The Arc B580 has 2,560 shading units, 160 texture mapping units, and 80 render output units. The 860M has 512 shading units, 32 TMUs, and 16 ROPs. Ray tracing cores follow a similar pattern: 20 on the Arc B580 versus 8 on the 860M. Neither GPU lists tensor cores in the database.

Clock behavior highlights the integrated versus discrete distinction. The 860M has a base clock of 600 MHz and a boost clock of 3000 MHz, allowing it to ramp aggressively under load while idling low to save power. The Arc B580 runs at a fixed 2670 MHz for both base and boost, with memory clocked at 2375 MHz (19 Gbps effective). The 860M's TDP of 15 W versus the Arc B580's 190 W reflects the enormous power envelope difference.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The Arc B580 uses a PCIe 4.0 x8 interface and is a dual-slot card measuring 272 mm long, 115 mm tall, and 45 mm wide. The 860M is an IGP with no slot width, no power connectors, and its display outputs are portable device dependent.

Head-to-Head Benchmarks

The database records exactly two head-to-head benchmark comparisons between these GPUs, and the Intel Arc B580 wins both decisively.

In Geekbench OpenCL, the Arc B580 posts a score of 92,821 while the 860M manages 22,759. The delta percentage of -75.5% indicates the 860M trails by three-quarters of the Arc B580's performance. This test exercises general-purpose compute workloads using the OpenCL API, and the discrete card's dedicated memory and far higher shading unit count produce a massive gap.

Geekbench Vulkan shows a similar pattern. The Arc B580 scores 109,672 versus the 860M's 30,043, a 72.6% deficit for the AMD part. Vulkan is a low-overhead graphics API used heavily in modern games, and the Arc B580's 2,560 shading units versus 512 gives it a raw throughput advantage that the API efficiently exploits.

The delta percentages are worth interpreting carefully. A -75.5% delta means the 860M's score is 75.5% lower than the Arc B580's, not that it is 75.5% as fast. In practical terms, the Arc B580 delivers roughly four times the compute performance in these tests. The 860M's Vulkan score of 30,043 is higher than its OpenCL score of 22,759, suggesting the RDNA 3.5 architecture handles Vulkan more efficiently relative to OpenCL. The Arc B580's Vulkan score is also higher than its OpenCL score, but the proportional gap is smaller.

Despite losing both head-to-head tests, the 860M maintains a higher average benchmark score in the overall database. This occurs because the two cards run different benchmark suites. The Arc B580's average of 23,021 is dragged down by additional tests including Passmark DirectX 9 (183), DirectX 10 (76), DirectX 11 (128), DirectX 12 (76), G2D (709), G3D (15,748), and GPU compute (7,729), plus a 3DMark Steel Nomad DX12 score of 3,068. The 860M's average of 26,401 comes from just the two Geekbench tests, which happen to score well relative to the broader database.

Specification Differences

| Specification | AMD Radeon 860M | Intel Arc B580 |

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

| Architecture | RDNA 3.5 | Xe2-HPG |

| Generation | Navi III IGP (Strix Point Mobile) | Battlemage (Arc 5) |

| Process Node | 4 nm | 5 nm |

| Chip | Krackan Point | BMG-G21 |

| Transistors | Unknown | 19,600 million |

| Die Size | Unknown | 272 mm² |

| Transistor Density | Not listed | 72.1M / mm² |

| Base Clock | 600 MHz | 2670 MHz |

| Boost Clock | 3000 MHz | 2670 MHz |

| Memory Size | System Shared | 12 GB GDDR6 |

| Memory Bus | System Shared | 192 bit |

| Memory Bandwidth | System Dependent | 456.0 GB/s |

| Shading Units | 512 | 2,560 |

| TMUs | 32 | 160 |

| ROPs | 16 | 80 |

| Ray Tracing Cores | 8 | 20 |

| Pixel Rate | 48.00 GPixel/s | 213.6 GPixel/s |

| Texture Rate | 96.00 GTexel/s | 427.2 GTexel/s |

| FP32 Performance | 3.072 TFLOPS | 13.67 TFLOPS |

| FP16 Performance | 3.072 TFLOPS (1:1) | 27.34 TFLOPS (2:1) |

| TDP | 15 W | 190 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 8-pin |

| Suggested PSU | Not listed | 450 W |

| Dimensions | Not listed | 272 mm x 115 mm x 45 mm |

| Display Outputs | Portable Device Dependent | 1x HDMI 2.1a, 3x DisplayPort 2.1 |

| Release Date | 2025-02-28 | 2024-12-12 |

| Predecessor | Navi II IGP | Alchemist |

| Launch MSRP | Not available | 249 USD |

Where Each One Wins

The AMD Radeon 860M wins in the efficiency and portability domain. Its 15 W TDP means it can operate in thin-and-light laptops without active cooling concerns, and its integrated nature eliminates the need for a separate power connector or power supply upgrade. The 3000 MHz boost clock shows it can ramp up when thermal headroom allows, and its 72nd percentile ranking among all GPUs is respectable for an integrated solution. The 860M's FP32 performance of 3.072 TFLOPS, while far below the Arc B580, is delivered within a tiny power envelope.

The Radeon 860M also wins on integration simplicity. It uses system-shared memory, which removes the cost and complexity of dedicated VRAM. For a portable device, the display outputs being device dependent is a feature rather than a limitation, as the host laptop handles connectivity. Its 4 nm process node is more advanced than the Arc B580's 5 nm, which contributes to its efficiency.

The Intel Arc B580 wins decisively in raw performance across every measured compute test. Its FP32 throughput of 13.67 TFLOPS is over four times the 860M's 3.072 TFLOPS, and its FP16 performance of 27.34 TFLOPS leverages a 2:1 ratio versus the 860M's 1:1 FP16 capability. The 456.0 GB/s memory bandwidth is a massive advantage for texture-heavy workloads and large datasets, and the 12 GB frame buffer allows for high-resolution textures that would overflow shared memory.

The Arc B580 also wins on display connectivity. Its 1x HDMI 2.1a and 3x DisplayPort 2.1 outputs support multi-monitor setups and high refresh rate displays, while the 860M's outputs depend entirely on the host portable device. The Arc B580's 20 ray tracing cores versus 8 give it a significant advantage in ray-traced workloads, and its 213.6 GPixel/s pixel rate is 4.45 times the 860M's 48.00 GPixel/s.

The Verdict

The data presents a clear split between two very different products. The AMD Radeon 860M is an integrated GPU designed for mobile efficiency, and it succeeds at that role. Its 15 W TDP, system-shared memory, and 4 nm process make it suitable for ultraportable laptops where battery life and thermals matter more than peak frame rates. The 72nd percentile ranking shows it outperforms many older discrete GPUs while consuming a fraction of their power. However, its benchmark scores in OpenCL and Vulkan are less than a third of the Arc B580's, and its shared memory architecture will bottleneck any memory-intensive workload.

The Intel Arc B580 is a full-sized discrete graphics card with the power budget to match. Its 190 W TDP, dual-slot cooler, and 8-pin power connector signal that it belongs in a desktop system with a 450 W or better power supply. The 12 GB of GDDR6 memory on a 192-bit bus with 456.0 GB/s bandwidth provides the data throughput that modern games and compute applications demand. The 2,560 shading units and 20 ray tracing cores give it a generational advantage over the integrated part. Its 68th percentile ranking is slightly lower than the 860M's, but this reflects the broader test suite it runs, not a deficiency in raw capability.

The verdict depends entirely on the use case. For a portable device where battery life, thermals, and physical space are constraints, the Radeon 860M is the only viable option of the two. Its performance is adequate for light gaming and general compute tasks, and it does so within a 15 W envelope that no discrete card can match. For a desktop system where performance is the priority, the Arc B580 is the clear choice. Its compute scores are 3.5 to 4 times higher than the 860M, its memory bandwidth is fixed and generous, and its display outputs support modern multi-monitor setups. The launch MSRP of 249 USD positions it as a mid-range discrete option, though the database records no equivalent pricing for the 860M.

The head-to-head data shows no ambiguity: the Arc B580 wins both recorded comparisons by margins exceeding 72%. The 860M's higher average benchmark score is a statistical artifact of different test suites, not evidence of superior performance. Users who need integrated graphics should choose the 860M for its efficiency. Users who need discrete performance should choose the Arc B580 for its compute capability. There is no overlap in their intended roles, and the benchmark data reflects that separation.