AMD Radeon Pro Vega 20 vs Intel Arc A350M Comparison

Head-to-Head Benchmarks

The recorded benchmark data shows the AMD Radeon Pro Vega 20 taking a decisive win in both shared tests. In Geekbench OpenCL, the AMD part scores 26679 against the Intel Arc A350M's 24546, a lead of 8.7%. The Vulkan result is closer but still favors AMD: 26410 versus 24747, a 6.7% advantage. Across the two head-to-head tests, the Radeon Pro Vega 20 wins 2, the Arc A350M wins 0.

The average benchmark score tells a similar story. The Radeon Pro Vega 20 sits at 27839, while the Arc A350M averages 24647. That is a gap of roughly 3,200 points, or about 12.9% in favor of the AMD part. Both cards sit in the same general percentile band, 73rd for AMD and 70th for Intel, but the margin between them is consistent across every metric the database records.

The OpenCL result is the more telling one for compute workloads. An 8.7% win means the Vega 20 sustains a meaningful performance advantage when the workload scales across the GPU's compute units. The Vulkan gap narrows to 6.7%, which suggests the Intel part closes some of the distance under a lower-level graphics API, but it still does not overtake the AMD card in any recorded test.

Looking at the nearest rival data places both cards in context. The Radeon Pro Vega 20's average score of 27839 puts it within 0.2% of the AMD Radeon RX 7800M (27883) and 0.5% of the AMD Radeon Pro W5500X (27973). It is also 1% ahead of the NVIDIA GeForce RTX 3090 (27565) in the database's average score, which is notable given the very different market positions of those two products. The Arc A350M's 24647 average lands it 0.4% behind the AMD Radeon RX 590 (24744) and 0.5% behind the NVIDIA RTX A5000 Mobile (24763). It runs 0.8% ahead of the AMD Radeon RX 6600 XT (24442) and 1.5% ahead of the NVIDIA GeForce GTX 1630 (24277).

In short, the data points one direction. The AMD Radeon Pro Vega 20 is the faster card in every recorded comparison, with its largest verified win coming in OpenCL.

Where Each One Wins

The AMD Radeon Pro Vega 20 wins both recorded benchmark disciplines, but the shape of its advantage differs by test. In OpenCL, the 8.7% margin reflects a strong compute architecture that handles general-purpose workloads well. This is the kind of result that matters for tasks like rendering, simulation, and other compute-heavy applications that lean on OpenCL. The Arc A350M's OpenCL score of 24546 is respectable, but it trails the AMD card by a clear margin.

In Vulkan, the AMD card wins again, but the margin shrinks to 6.7%. Vulkan tends to reward modern driver efficiency and low-level hardware utilization, so the smaller gap suggests the Intel Xe-HPG architecture is comparatively stronger in that API than in OpenCL. Still, the Radeon Pro Vega 20 holds the lead, so for Vulkan-based games or applications, the AMD part remains the safer pick from the recorded data.

The average benchmark score reinforces the overall picture. The Vega 20's 27839 average is 12.9% higher than the Arc's 24647, meaning the AMD card wins not just in individual tests but across the aggregate of all recorded benchmarks. The Arc A350M does not win any single recorded test, so there is no workload category in the database where it takes the lead.

Where the Intel card might appeal is efficiency rather than raw speed. The Arc A350M draws 25 W against the Vega 20's 100 W, a fourfold difference in power. For a portable device, that lower power draw could translate into quieter operation, less heat, and longer battery life, even if the performance ceiling is lower. The data does not record battery or thermal results, so that remains a qualitative observation based on the power figures in the database.

Architecture Differences

The two GPUs come from different architectural lineages. The AMD Radeon Pro Vega 20 uses the Vega 12 chip built on GCN 5.0 architecture, fabricated on a 14 nm process at GlobalFoundries. The Intel Arc A350M uses the DG2-128 chip with Xe-HPG architecture, built on a 6 nm process at TSMC. That process gap is significant: 6 nm is a much denser, more modern node than 14 nm, which helps explain how Intel fits 7,200 million transistors into a 157 mm² die, a transistor density of 45.9M per mm². The AMD chip's transistor count and die size are not recorded in the database.

The compute configurations differ substantially. The Vega 20 packs 1280 shading units, 80 texture mapping units, and 32 ROPs. The Arc A350M has 768 shading units, 48 TMUs, and 24 ROPs. Despite fewer units, the Intel card posts higher raw throughput numbers: its pixel rate is 52.80 GPixel/s versus 41.06 GPixel/s for AMD, and its texture rate is 105.6 GTexel/s versus 102.6 GTexel/s. In FP32 compute, the Arc A350M reaches 3.379 TFLOPS, narrowly ahead of the Vega 20's 3.284 TFLOPS. The FP16 numbers follow the same pattern, with Intel at 6.758 TFLOPS and AMD at 6.569 TFLOPS, both at a 2:1 ratio.

Clock speeds help explain those throughput results. The Arc A350M runs a base clock of 1150 MHz and a boost of 2200 MHz, while the Vega 20 runs 815 MHz base and 1283 MHz boost. The Intel part's much higher clocks compensate for its smaller compute unit count, and in raw throughput terms it actually comes out ahead, even though the AMD card wins the recorded benchmarks. That suggests the AMD architecture extracts more real-world performance per theoretical FLOP, or that driver and API efficiency favor the older GCN design in these specific tests.

Memory is another major split. The Vega 20 uses 4 GB of HBM2 on a 1024 bit bus, delivering 189.4 GB/s of bandwidth. The Arc A350M uses 4 GB of GDDR6 on a 64 bit bus, delivering 112.0 GB/s. That is a 69% bandwidth advantage for the AMD card, which is likely a significant factor in its benchmark wins, especially in compute workloads that stress memory throughput. The memory clocks reflect this: AMD runs at 740 MHz with 1480 Mbps effective, while Intel runs at 1750 MHz with 14 Gbps effective. The narrower bus on the Intel card limits its total bandwidth despite the faster memory clock.

Feature support also differs. The Arc A350M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the Vega 20 supports DirectX 12 (12_1) and Vulkan 1.3. Both cards support OpenGL 4.6. The Intel part also includes 6 ray tracing cores, while the AMD card has none recorded. The Arc A350M connects via PCIe 4.0 x8, while the Vega 20 uses PCIe 3.0 x16. Both are integrated-class parts with portable device dependent display outputs.

FAQ

Q: Which GPU is faster in the recorded benchmarks?

A: The AMD Radeon Pro Vega 20 wins both shared tests. It scores 26679 in OpenCL versus 24546 for the Arc A350M, and 26410 in Vulkan versus 24747. Its average benchmark score is 27839 against 24647.

Q: How do the two cards compare in raw compute throughput?

A: The Intel Arc A350M has slightly higher peak FP32 performance at 3.379 TFLOPS versus 3.284 TFLOPS for AMD. Its pixel rate is also higher at 52.80 GPixel/s versus 41.06 GPixel/s, and its texture rate is 105.6 GTexel/s versus 102.6 GTexel/s.

Q: Why does the AMD card win benchmarks if the Intel card has higher clocks and throughput?

A: The Vega 20 has substantially more memory bandwidth at 189.4 GB/s versus 112.0 GB/s for the Arc A350M. The Intel card's higher clock speeds, 2200 MHz boost versus 1283 MHz, do not translate into benchmark wins in the recorded tests.

Q: What are the power requirements of each card?

A: The AMD Radeon Pro Vega 20 has a TDP of 100 W, while the Intel Arc A350M draws only 25 W.

Q: Which card has better API support?

A: The Intel Arc A350M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the AMD card supports DirectX 12 (12_1) and Vulkan 1.3. Both support OpenGL 4.6.

Q: How does each card compare to its nearest rivals in the database?

A: The Vega 20's average score of 27839 is within 0.2% of the AMD Radeon RX 7800M and 1% ahead of the NVIDIA GeForce RTX 3090. The Arc A350M's 24647 average is 0.4% behind the AMD Radeon RX 590 and 0.8% ahead of the AMD Radeon RX 6600 XT.

Specification Differences

| Specification | AMD Radeon Pro Vega 20 | Intel Arc A350M |

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

| Chip | Vega 12 | DG2-128 |

| Architecture | GCN 5.0 | Xe-HPG |

| Process node | 14 nm | 6 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | Not recorded | 7,200 million |

| Die size | Not recorded | 157 mm² |

| Base clock | 815 MHz | 1150 MHz |

| Boost clock | 1283 MHz | 2200 MHz |

| Memory type | HBM2 | GDDR6 |

| Memory bus width | 1024 bit | 64 bit |

| Memory bandwidth | 189.4 GB/s | 112.0 GB/s |

| Shading units | 1280 | 768 |

| TMUs | 80 | 48 |

| ROPs | 32 | 24 |

| Ray tracing cores | Not recorded | 6 |

| Pixel rate | 41.06 GPixel/s | 52.80 GPixel/s |

| Texture rate | 102.6 GTexel/s | 105.6 GTexel/s |

| FP32 | 3.284 TFLOPS | 3.379 TFLOPS |

| FP16 | 6.569 TFLOPS (2:1) | 6.758 TFLOPS (2:1) |

| TDP | 100 W | 25 W |

| Bus interface | PCIe 3.0 x16 | PCIe 4.0 x8 |

| DirectX | 12 (12_1) | 12 Ultimate (12_2) |

| Vulkan | 1.3 | 1.4 |

| Release date | 2018-11-13 | 2022-03-29 |

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

The Verdict

The data makes the performance hierarchy clear. The AMD Radeon Pro Vega 20 is the faster GPU in every recorded benchmark, winning OpenCL by 8.7% and Vulkan by 6.7%, with an average score 12.9% higher than the Intel Arc A350M. Anyone prioritizing raw benchmark performance should choose the AMD part without hesitation.

The Arc A350M's case rests on efficiency, not speed. At 25 W versus 100 W, it draws a quarter of the power while delivering roughly 88.5% of the average benchmark score. For thin-and-light portables where heat and battery life matter more than peak performance, that tradeoff can be the right one. The Intel card also brings modern features the AMD part lacks, including DirectX 12 Ultimate support, Vulkan 1.4, and 6 ray tracing cores.

The Vega 20 counters with a massive memory bandwidth advantage, 189.4 GB/s on a 1024 bit HBM2 bus versus 112.0 GB/s on a 64 bit GDDR6 bus. That bandwidth is likely a major reason it wins the recorded compute benchmarks despite lower clock speeds and lower peak FP32 throughput.

For compute workloads and Vulkan-based graphics, the Radeon Pro Vega 20 is the verified winner. For power-constrained portable systems where lower draw matters and ray tracing support is desired, the Arc A350M is the sensible alternative. Both cards are end-of-life products, so the choice comes down to which tradeoff fits the use case: AMD for proven benchmark performance, Intel for efficiency and feature set.