AMD Radeon 890M vs AMD Radeon Vega 8 Comparison

AMD Radeon Vega 8 and AMD Radeon 890M are both integrated graphics processors from AMD, but they sit at opposite ends of the company’s IGP spectrum. The Vega 8, built on the Raven Ridge generation, is an end-of-life part from 2018, while the 890M is an active, current-generation Navi III IGP from 2024. Benchmark data shows they are remarkably close in overall average score—the Vega 8 averages 9221, and the 890M averages 9210—yet this aggregate masks a profound performance divergence in individual workloads. The two share the same 45th percentile ranking among all GPUs, and their nearest rival lists are nearly identical, with the NVIDIA GeForce GTX 960, GTX 465, and GTX 850M all within a single percentage point of both. This head-to-head analysis dissects where each part actually wins and why their architectures produce such different results despite similar averages.

Head-to-Head Benchmarks

The direct comparison between the two IGPs is stark, with the 890M dominating both shared benchmark tests. In Geekbench OpenCL, the 890M scores 37,254, while the Vega 8 scores 8,822—a delta of -76.3% from the perspective of the Vega 8, meaning the 890M is roughly 4.2 times faster. The Geekbench Vulkan test shows an even wider gap: the 890M scores 40,808 versus the Vega 8’s 8,134, a delta of -80.1%. These are not marginal improvements; they represent a generational leap in raw compute and graphics throughput.

The two tests available in the head-to-head set are both compute-oriented, which favors the 890M’s architectural advantages. The Vega 8’s best showing is its Geekbench Metal score of 10,706, which is not part of the direct comparison but indicates its performance ceiling in API-specific workloads. The 890M’s additional benchmarks—Passmark G3D at 8,076, Passmark GPU Compute at 4,157, and 3DMark Steel Nomad DX12 at 590—provide context for its broader capability, though none of these appear in the head-to-head table.

What is notable is that the average benchmark scores are nearly identical—9221 for the Vega 8 and 9210 for the 890M—despite the 890M winning both head-to-head tests. This is because the Vega 8’s three Geekbench scores (Metal, OpenCL, Vulkan) are all clustered between 8,134 and 10,706, producing a high average, while the 890M’s ten benchmarks include several low Passmark DX scores (33, 37, 73, 97) that drag its average down. The data suggests the 890M is a specialist in compute-heavy OpenCL and Vulkan workloads, while the Vega 8’s consistency across different APIs keeps its average competitive.

Where Each One Wins

The 890M wins decisively in OpenCL and Vulkan compute, with margins exceeding 76% in both cases. This makes it the clear choice for applications that leverage these APIs, such as GPU-accelerated rendering, scientific computing, or any workload that offloads parallel tasks to the GPU. Its Vulkan score of 40,808 is particularly strong, indicating excellent performance in modern cross-platform graphics and compute APIs. The 890M also holds advantages in pixel rate (92.80 GPixel/s versus 8.80 GPixel/s) and texture rate (185.6 GTexel/s versus 35.20 GTexel/s), which are not directly benchmarked but align with its superior hardware specifications.

The Vega 8’s wins are narrower and less obvious from the head-to-head data. Its Geekbench Metal score of 10,706 is its highest, and while there is no direct Metal comparison with the 890M, this suggests the Vega 8 is competitive in Apple’s Metal API, which is not listed among the 890M’s benchmark tests. The Vega 8 also has a higher average benchmark score (9221 versus 9210), meaning that when all available benchmarks are averaged, it edges out the 890M by 0.1%. This is a marginal difference, but it indicates that the Vega 8’s performance is more evenly distributed across different test types, whereas the 890M’s scores are highly variable—ranging from 33 in Passmark DX10 to 40,808 in Vulkan.

For gaming, the 890M’s higher pixel and texture rates suggest it would handle modern titles better, but the PASSMark DX9 (97), DX11 (73), and DX12 (37) scores are all low, indicating that legacy DirectX workloads are not its strength. The Vega 8, with a DirectX 12 (12_1) API support level, may be more consistent in older games, though its raw compute is far lower.

Architecture Differences

The two IGPs are separated by six years of architectural evolution. The Vega 8 uses GCN 5.0 architecture, manufactured on a 14 nm process at GlobalFoundries, with a die size of 210 mm² and 4,940 million transistors. The 890M uses RDNA 3.5, built on a 4 nm process at TSMC, with a die size of 233 mm² and 34,000 million transistors. The transistor density difference is dramatic: 23.5 million transistors per mm² for the Vega 8 versus 145.9 million per mm² for the 890M—a 6.2-fold increase in density.

The compute unit layout differs fundamentally. The Vega 8 has 512 shading units, 32 TMUs, and 8 ROPs, while the 890M doubles the shading units to 1,024, doubles TMUs to 64, and quadruples ROPs to 32. The 890M also introduces 16 ray tracing cores, a feature entirely absent from the Vega 8. Clock speeds reflect the process node advantage: the Vega 8 boosts to 1,100 MHz from a 300 MHz base, while the 890M boosts to 2,900 MHz from a 400 MHz base. This 2.6x boost clock advantage, combined with doubled shading units, explains the compute performance gap.

Memory architecture is similar—both use system-shared memory with system-dependent bandwidth—but the 890M interfaces via PCIe 4.0 x8, whereas the Vega 8 uses an IGP bus interface. The 890M’s FP32 throughput is 5.939 TFLOPS, more than 5 times the Vega 8’s 1,126.4 GFLOPS. FP16 performance also differs: the Vega 8 achieves 2.253 TFLOPS with a 2:1 ratio, while the 890M achieves 5.939 TFLOPS with a 1:1 ratio, meaning it handles FP16 at full rate without penalty. Power efficiency is notable—the 890M has a TDP of 15 W despite vastly higher performance, versus 25 W for the Vega 8.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon Vega 8 has a slightly higher average benchmark score of 9,221, compared to the AMD Radeon 890M’s 9,210, a difference of 0.1%.

Q: How much faster is the 890M in Geekbench OpenCL?

A: The 890M scores 37,254 in Geekbench OpenCL, while the Vega 8 scores 8,822, making the 890M approximately 76.3% faster.

Q: Does the 890M support ray tracing?

A: Yes, the 890M includes 16 ray tracing cores, whereas the Vega 8 has no RT cores listed.

Q: What is the process node difference?

A: The Vega 8 uses a 14 nm process at GlobalFoundries, while the 890M uses a 4 nm process at TSMC.

Q: Which GPU has a higher boost clock?

A: The 890M has a boost clock of 2,900 MHz, more than double the Vega 8’s 1,100 MHz boost.

Q: Are both GPUs integrated?

A: Yes, both are IGPs with system-shared memory, but the 890M uses a PCIe 4.0 x8 bus interface, while the Vega 8 uses a standard IGP bus interface.

Specification Differences

| Specification | AMD Radeon Vega 8 | AMD Radeon 890M |

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

| Architecture | GCN 5.0 | RDNA 3.5 |

| Process Node | 14 nm | 4 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 4,940 million | 34,000 million |

| Die Size | 210 mm² | 233 mm² |

| Transistor Density | 23.5M / mm² | 145.9M / mm² |

| Base Clock | 300 MHz | 400 MHz |

| Boost Clock | 1100 MHz | 2900 MHz |

| Shading Units | 512 | 1024 |

| TMUs | 32 | 64 |

| ROPs | 8 | 32 |

| RT Cores | None | 16 |

| Pixel Rate | 8.800 GPixel/s | 92.80 GPixel/s |

| Texture Rate | 35.20 GTexel/s | 185.6 GTexel/s |

| FP32 | 1,126.4 GFLOPS | 5.939 TFLOPS |

| FP16 | 2.253 TFLOPS (2:1) | 5.939 TFLOPS (1:1) |

| TDP | 25 W | 15 W |

| Bus Interface | IGP | PCIe 4.0 x8 |

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

| Vulkan Support | 1.3 | 1.4 |

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

| Release Date | 2018-02-11 | 2024-07-14 |

| Predecessor | GCN 3.0 IGP | Navi II IGP |

| Successor | Vega II IGP | None |

The Verdict

The benchmark data makes one thing clear: the 890M is the superior compute processor. Its Geekbench OpenCL and Vulkan scores are over 4 times higher than the Vega 8’s, and it does so while drawing 10 W less power (15 W versus 25 W). The 890M’s ray tracing cores, doubled shading units, and 2.9 GHz boost clock provide a hardware foundation that the Vega 8 cannot match, even with its higher average score. The 890M is the choice for users running Vulkan-based games, OpenCL compute tasks, or any workload that exploits its 5.939 TFLOPS FP32 throughput.

The Vega 8’s edge is narrower but real in specific contexts. Its Geekbench Metal score of 10,706 is the highest among its listed benchmarks, and its consistent scores across Metal, OpenCL, and Vulkan (all within 2,572 points of each other) suggest greater API versatility. For users locked into legacy DirectX 9, 10, or 11 applications, the Vega 8’s 12_1 support may be more stable than the 890M’s 12 Ultimate, though the 890M’s Passmark DX scores are quite low. The Vega 8 is also end-of-life, meaning it will only be found in older systems, while the 890M is an active product for current mobile platforms.

In practical terms, the 890M is the only rational choice for new hardware. Its 76.3% lead in OpenCL and 80.1% lead in Vulkan are decisive, and its lower TDP makes it more suitable for thin-and-light laptops. The Vega 8’s average score advantage of 0.1% is statistically negligible and driven by a lack of low-end benchmarks in its test set, not by actual performance superiority. The 890M’s 16 RT cores and PCIe 4.0 x8 interface are future-proofing features the Vega 8 simply lacks. Anyone choosing between these two should select the 890M, unless they specifically require the Vega 8’s Metal performance profile—and even then, the compute gap is likely to outweigh that single API advantage.