AMD Radeon HD 8690M vs AMD Radeon Vega 8 Mobile Comparison

Where Each One Wins

The benchmark data records only one direct head-to-head comparison between the AMD Radeon Vega 8 Mobile and the AMD Radeon HD 8690M, and that test is Geekbench OpenCL. The Vega 8 Mobile wins that test decisively, with a score of 7435 against the HD 8690M's 6137, a 21.2% advantage. This makes the use-case split straightforward: the Vega 8 Mobile is the stronger compute performer in the single measured workload, while the HD 8690M has no recorded benchmark victory in this comparison.

Looking beyond the direct comparison, the database places the Vega 8 Mobile in the 39th percentile of all GPUs, while the HD 8690M sits at the 36th percentile. The average benchmark score for the Vega 8 Mobile is 7203, compared to 6137 for the HD 8690M. These percentile rankings reinforce the same conclusion: the Vega 8 Mobile holds a consistent edge in overall recorded performance, even if the margin varies depending on the specific test.

The Vega 8 Mobile's nearest rivals in the database include the NVIDIA GeForce GTX 750 (average score 7222, only 0.3% lower), the NVIDIA GeForce GTX 560 SE (7171, 0.4% higher for the Vega), the Intel Iris Pro Graphics P580 (7170, 0.5% higher for the Vega), and the NVIDIA GeForce GTX 970 (7157, 0.6% higher for the Vega). These close margins indicate the Vega 8 Mobile operates in a competitive performance band where small differences separate adjacent GPUs.

For the HD 8690M, the nearest rivals are the Intel Iris Pro Graphics 6200 (6117, 0.3% higher for the HD), the NVIDIA RTX A400 (6078, 1% higher for the HD), the NVIDIA GeForce MX230 (6077, 1% higher for the HD), and the AMD Radeon HD 6950 (6210, 1.2% lower for the HD). The HD 8690M's position among these rivals shows it is slightly ahead of some modern entry-level parts but behind the older HD 6950 in the database's measurements.

Architecture Differences

The two GPUs come from different generations of AMD graphics architecture. The Vega 8 Mobile uses the GCN 5.0 architecture with the Raven-M chip, part of the Vega IGP generation for Raven Ridge-M processors. The HD 8690M uses the older GCN 1.0 architecture with the Sun chip, belonging to the Solar System generation for HD 8600M series.

The manufacturing process differs substantially. The Vega 8 Mobile is built on a 14 nm process at GlobalFoundries, while the HD 8690M uses a 28 nm process at TSMC. The Vega 8 Mobile integrates 4,940 million transistors on a 210 mm² die, giving a transistor density of 23.5 million per square millimeter. The HD 8690M contains 690 million transistors on a 56 mm² die, with a density of 12.3 million per square millimeter. The newer process node allows the Vega 8 Mobile to pack over seven times as many transistors into a die roughly four times larger.

The memory configurations are fundamentally different. The Vega 8 Mobile uses system-shared memory, with its size, type, bus width, and bandwidth all listed as system dependent. The HD 8690M has a dedicated 2 GB of GDDR5 memory on a 64-bit bus, delivering 32.00 GB/s of bandwidth. The HD 8690M's memory runs at 1000 MHz (4 Gbps effective), while the Vega 8 Mobile's memory clock is not a fixed specification.

The shading resources also differ. The Vega 8 Mobile has 512 shading units, 32 texture mapping units, and 8 raster output units. The HD 8690M has 320 shading units, 20 texture mapping units, and 8 raster output units. The Vega 8 Mobile therefore has 60% more shading units and 60% more TMUs, while both have the same ROP count.

The Vega 8 Mobile's clock speeds are lower in absolute terms: a base of 300 MHz and a boost of 1101 MHz. The HD 8690M runs at a base of 925 MHz and a boost of 975 MHz. Despite the lower base clock, the Vega 8 Mobile's much larger shader array produces higher throughput figures: 1,127.4 GFLOPS FP32 versus 624.0 GFLOPS for the HD 8690M. The Vega 8 Mobile also records 2.255 TFLOPS FP16 (2:1 ratio), while the HD 8690M has no recorded FP16 capability.

The Vega 8 Mobile is an integrated graphics processor (IGP) with no power connectors and a 25 W TDP. The HD 8690M is a discrete mobile GPU using a PCIe 3.0 x8 bus interface, with no TDP recorded in the database. The Vega 8 Mobile supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3, while the HD 8690M supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170.

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL, and the results are unambiguous. The AMD Radeon Vega 8 Mobile scores 7435, while the AMD Radeon HD 8690M scores 6137. The delta is 21.2% in favor of the Vega 8 Mobile. This is a substantial margin, indicating that in OpenCL compute workloads, the newer integrated part outperforms the older discrete part by roughly one-fifth.

To contextualize this 21.2% gap, consider the nearest rival data. The Vega 8 Mobile's closest rival, the NVIDIA GeForce GTX 750, has an average score of 7222, which is only 0.3% lower than the Vega 8 Mobile's average of 7203. The HD 8690M's nearest rival, the Intel Iris Pro Graphics 6200, scores 6117, just 0.3% below the HD 8690M's 6137. The difference between the two AMD parts (21.2%) is far larger than the differences between either part and its immediate neighbors in the database, which are all within 1.2%. This suggests the performance gap between the Vega 8 Mobile and HD 8690M is more significant than typical generation-to-generation differences seen among adjacent GPUs.

The Vega 8 Mobile also holds a clear advantage in raw throughput metrics derived from the database. Its pixel rate is 8.808 GPixel/s versus 7.800 GPixel/s for the HD 8690M, a 12.9% advantage. The texture rate is 35.23 GTexel/s versus 19.50 GTexel/s, an 80.7% advantage. The FP32 compute is 1,127.4 GFLOPS versus 624.0 GFLOPS, an 80.7% advantage. These ratios show that the Vega 8 Mobile's advantage is not uniform across all operations: it is most pronounced in texture and compute throughput, less so in pixel throughput.

The HD 8690M does have one clear specification advantage: its dedicated GDDR5 memory with a fixed 32.00 GB/s bandwidth. The Vega 8 Mobile's memory bandwidth is system dependent, meaning its actual performance can vary based on the host system's memory configuration. In the recorded OpenCL test, however, the Vega 8 Mobile still managed a 21.2% win despite this potential limitation.

FAQ

Q: Which GPU has the higher average benchmark score in the database?

A: The AMD Radeon Vega 8 Mobile has an average benchmark score of 7203, while the AMD Radeon HD 8690M has an average score of 6137. The Vega 8 Mobile also ranks in the 39th percentile of all GPUs, compared to the 36th percentile for the HD 8690M.

Q: How much faster is the Vega 8 Mobile in the only head-to-head benchmark?

A: In the Geekbench OpenCL test, the Vega 8 Mobile scores 7435 against the HD 8690M's 6137, a 21.2% advantage. This is the only direct head-to-head benchmark recorded in the database.

Q: Do the two GPUs use the same architecture?

A: No. The Vega 8 Mobile uses GCN 5.0 architecture on a 14 nm process at GlobalFoundries, while the HD 8690M uses GCN 1.0 architecture on a 28 nm process at TSMC.

Q: What are the memory differences between the two?

A: The HD 8690M has 2 GB of dedicated GDDR5 memory on a 64-bit bus with 32.00 GB/s bandwidth. The Vega 8 Mobile uses system-shared memory, with size, type, bus width, and bandwidth all dependent on the host system.

Q: Which GPU has more shading units?

A: The Vega 8 Mobile has 512 shading units, 32 TMUs, and 8 ROPs. The HD 8690M has 320 shading units, 20 TMUs, and 8 ROPs. Both have the same ROP count, but the Vega 8 Mobile has 60% more shading units and TMUs.

Q: What API support does each GPU offer?

A: The Vega 8 Mobile supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The HD 8690M supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. Both support DirectX 12 and OpenGL 4.6, but the Vega 8 Mobile has a higher DirectX feature level and a newer Vulkan version.

The Verdict

The data clearly favors the AMD Radeon Vega 8 Mobile in this comparison. It wins the only recorded head-to-head benchmark by 21.2%, has a higher average benchmark score (7203 versus 6137), and ranks in a higher percentile of all GPUs (39th versus 36th). Its architecture is significantly newer, built on a 14 nm process with GCN 5.0 versus the HD 8690M's 28 nm GCN 1.0. The Vega 8 Mobile also offers nearly double the FP32 compute throughput and over 80% higher texture rate.

The HD 8690M's advantages are limited to its memory configuration. It has dedicated 2 GB of GDDR5 with a fixed 32.00 GB/s bandwidth, whereas the Vega 8 Mobile depends on system-shared memory. For users or systems where dedicated VRAM is a strict requirement, the HD 8690M has that specification on paper. However, the recorded benchmark data shows that the Vega 8 Mobile still outperforms the HD 8690M in OpenCL compute despite this memory disadvantage.

For a buyer choosing between these two end-of-life parts, the Vega 8 Mobile is the stronger performer in every measured benchmark. Its higher shading unit count (512 versus 320), newer architecture, and superior throughput figures make it the better choice for compute-oriented workloads. The HD 8690M remains a functional option for systems that require a discrete GPU with its own memory, but the performance data does not support choosing it over the Vega 8 Mobile on performance grounds.

The nearest rival data reinforces this verdict. The Vega 8 Mobile sits within 0.6% of GPUs like the NVIDIA GeForce GTX 970 in average score, while the HD 8690M sits within 1.2% of parts like the AMD Radeon HD 6950 and the Intel Iris Pro Graphics 6200. The Vega 8 Mobile competes in a higher performance tier overall.

Specification Differences

| Specification | AMD Radeon Vega 8 Mobile | AMD Radeon HD 8690M |

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

| Architecture | GCN 5.0 | GCN 1.0 |

| Chip | Raven-M | Sun |

| Generation | Vega IGP (Raven Ridge-M) | Solar System (HD 8600M) |

| Process Node | 14 nm | 28 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 4,940 million | 690 million |

| Die Size | 210 mm² | 56 mm² |

| Transistor Density | 23.5M / mm² | 12.3M / mm² |

| Base Clock | 300 MHz | 925 MHz |

| Boost Clock | 1101 MHz | 975 MHz |

| Memory Size | System Shared | 2 GB |

| Memory Type | System Shared | GDDR5 |

| Memory Bus | System Shared | 64 bit |

| Memory Bandwidth | System Dependent | 32.00 GB/s |

| Shading Units | 512 | 320 |

| TMUs | 32 | 20 |

| ROPs | 8 | 8 |

| Pixel Rate | 8.808 GPixel/s | 7.800 GPixel/s |

| Texture Rate | 35.23 GTexel/s | 19.50 GTexel/s |

| FP32 Performance | 1,127.4 GFLOPS | 624.0 GFLOPS |

| FP16 Performance | 2.255 TFLOPS (2:1) | Not recorded |

| TDP | 25 W | Not recorded |

| Bus Interface | IGP | PCIe 3.0 x8 |

| DirectX Support | 12 (12_1) | 12 (11_1) |

| Vulkan Support | 1.3 | 1.2.170 |

| Release Date | 2019-01-07 | 2013-02-28 |

| Predecessor | GCN 3.0 IGP | London |

| Successor | Navi II IGP | Gem System |