AMD Radeon HD 8690M vs AMD Radeon HD 8730M Comparison

FAQ

Q: Which GPU has the higher benchmark score in the Geekbench OpenCL test?

A: The AMD Radeon HD 8690M scores 6137, which is 3.1% higher than the AMD Radeon HD 8730M's score of 5955.

Q: How do these two GPUs compare in terms of overall percentile ranking?

A: The HD 8690M sits in the 36th percentile of all GPUs, while the HD 8730M sits in the 34th percentile. This places both in the lower-middle range of the performance spectrum.

Q: What are the closest rivals to the HD 8690M in benchmark performance?

A: The nearest rivals include the Intel Iris Pro Graphics 6200 (score 6117, 0.3% behind), NVIDIA RTX A400 (score 6078, 1% behind), NVIDIA GeForce MX230 (score 6077, 1% behind), and AMD Radeon HD 6950 (score 6210, 1.2% ahead).

Q: Which GPU has the higher base and boost clock speeds?

A: The HD 8690M runs at a base clock of 925 MHz and boost clock of 975 MHz. The HD 8730M runs significantly lower at 650 MHz base and 700 MHz boost.

Q: What memory configurations do these GPUs use, and how do they differ?

A: The HD 8690M uses 2 GB of GDDR5 memory on a 64-bit bus with 32.00 GB/s bandwidth. The HD 8730M uses 2 GB of DDR3 memory on a 128-bit bus with 28.80 GB/s bandwidth.

Q: Do both GPUs share the same architectural foundation?

A: Yes, both are built on AMD's GCN 1.0 architecture using a 28 nm process at TSMC, and both support DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170.

Architecture Differences

Both GPUs belong to the same Solar System generation but occupy different segments within it. The HD 8690M uses the Sun chip, while the HD 8730M uses the Mars chip. Despite the differing chip names, both are fabricated on the same 28 nm process at TSMC and share the GCN 1.0 architecture.

The transistor counts reveal a significant physical difference. The HD 8730M's Mars chip packs 950 million transistors on a 77 mm² die. The HD 8690M's Sun chip contains 690 million transistors on a smaller 56 mm² die. Interestingly, both achieve the same transistor density of 12.3M / mm², indicating the same manufacturing process efficiency.

Compute resources diverge notably. The HD 8730M fields 384 shading units and 24 texture mapping units, while the HD 8690M has 320 shading units and 20 TMUs. Both GPUs have 8 ROPs. This gives the HD 8730M a raw shader-count advantage of 20%, yet the actual benchmark outcome does not favor it.

Clock speeds reverse this advantage. The HD 8690M runs at 925 MHz base and 975 MHz boost, compared to the HD 8730M's 650 MHz base and 700 MHz boost. This clock advantage is substantial — the HD 8690M's boost clock is 275 MHz higher — and it more than compensates for the HD 8730M's extra shaders in practice.

Memory architecture takes a different route on each card. The HD 8690M uses 2 GB of GDDR5 at 1000 MHz (4 Gbps effective) on a 64-bit bus. The HD 8730M uses 2 GB of DDR3 at 900 MHz (1800 Mbps effective) on a 128-bit bus. The GDDR5 memory on the HD 8690M delivers higher bandwidth at 32.00 GB/s versus 28.80 GB/s, even with half the bus width.

Both cards are end-of-life products, released within a month of each other — the HD 8690M on 2013-02-28 and the HD 8730M on 2013-03-31. Both share the same predecessor (London) and successor (Gem System), and both connect via PCIe 3.0 x8.

Head-to-Head Benchmarks

The only available head-to-head benchmark is the Geekbench OpenCL test, and the results are close but decisive. The AMD Radeon HD 8690M scores 6137 against the AMD Radeon HD 8730M's 5955. That is a 3.1% delta in favor of the HD 8690M, giving it one win in the single test category.

A 3.1% margin may seem modest, but the underlying hardware differences make this outcome notable. The HD 8730M has 64 more shading units and 4 more TMUs, yet it loses because its clocks are far lower. The HD 8690M's 975 MHz boost versus the HD 8730M's 700 MHz boost creates a 39% clock advantage that overcomes the shader deficit. The result is a textbook case of clock speed trumping compute unit count in this generation.

The memory subsystems also help explain the score gap. The HD 8690M's GDDR5 memory provides 32.00 GB/s of bandwidth, which is 11% higher than the HD 8730M's DDR3 at 28.80 GB/s. In memory-sensitive OpenCL workloads, this bandwidth advantage can be decisive, even when the competing card has a wider 128-bit bus.

Pixel and texture rates reinforce the HD 8690M's lead. It achieves 7.800 GPixel/s and 19.50 GTexel/s, versus the HD 8730M's 5.600 GPixel/s and 16.80 GTexel/s. The HD 8690M is 39% faster in pixel throughput and 16% faster in texture throughput, despite having fewer texture units.

FP32 compute follows the same pattern. The HD 8690M delivers 624.0 GFLOPS against the HD 8730M's 537.6 GFLOPS, a 16% advantage. This is consistent with the benchmark delta, though the benchmark margin is smaller than the theoretical compute gap suggests.

In percentile terms, the HD 8690M ranks at 36 versus the HD 8730M's 34. Both are close in the overall distribution, but the HD 8690M's position is measurably higher. Its nearest rival, the AMD Radeon HD 6950, sits just 1.2% ahead, while the HD 8730M's closest competitor, the NVIDIA Quadro K620M, is effectively tied at 0% delta.

The Verdict

The data is unambiguous: the AMD Radeon HD 8690M is the faster GPU. It wins the only head-to-head benchmark by 3.1%, holds a higher percentile rank (36 vs 34), and produces higher theoretical throughput in every measurable category — pixel rate, texture rate, and FP32 compute.

The HD 8730M offers more shading units and a wider memory bus, but these advantages do not translate into real-world performance. Its lower clocks and slower DDR3 memory cripple its output. In the Geekbench OpenCL test, the HD 8690M's superior clock speed and memory bandwidth carry the day.

Users should pick the HD 8690M for any compute-oriented task where the benchmark result matters. The 6137 score places it ahead of the Intel Iris Pro Graphics 6200 (6117) and the NVIDIA RTX A400 (6078), making it competitive within its immediate performance neighborhood. The HD 8730M, at 5955, lands just below the HD 8750M (5970) and just above the Intel UHD Graphics 730 (5929).

For users who already own either card, the performance difference is small enough that upgrading from the HD 8730M to the HD 8690M would yield only a marginal gain. The 3.1% delta is within the range of run-to-run variance for many workloads. However, for a fresh choice between the two, the HD 8690M is the clear winner based on every data point in the record.

Neither card is competitive in the broader GPU landscape — both sit in the bottom third of all GPUs. But within this specific comparison, the HD 8690M's combination of higher clocks, faster memory, and better benchmark results makes it the superior option.

Specification Differences

The following table highlights only the fields where the two GPUs differ:

| Specification | AMD Radeon HD 8690M | AMD Radeon HD 8730M |

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

| Chip | Sun | Mars |

| Generation | Solar System (HD 8600M) | Solar System (HD 8700M) |

| Transistors | 690 million | 950 million |

| Die Size | 56 mm² | 77 mm² |

| Base Clock | 925 MHz | 650 MHz |

| Boost Clock | 975 MHz | 700 MHz |

| Memory Clock | 1000 MHz / 4 Gbps effective | 900 MHz / 1800 Mbps effective |

| Memory Type | GDDR5 | DDR3 |

| Bus Width | 64 bit | 128 bit |

| Bandwidth | 32.00 GB/s | 28.80 GB/s |

| Shading Units | 320 | 384 |

| TMUs | 20 | 24 |

| Pixel Rate | 7.800 GPixel/s | 5.600 GPixel/s |

| Texture Rate | 19.50 GTexel/s | 16.80 GTexel/s |

| FP32 | 624.0 GFLOPS | 537.6 GFLOPS |

| Release Date | 2013-02-28 | 2013-03-31 |

| Geekbench OpenCL | 6137 | 5955 |

| Percentile | 36 | 34 |

Both cards share the same ROP count (8), process node (28 nm), foundry (TSMC), transistor density (12.3M / mm²), bus interface (PCIe 3.0 x8), memory size (2 GB), API support (DirectX 12 11_1, OpenGL 4.6, Vulkan 1.2.170), production status (End-of-life), predecessor (London), and successor (Gem System).

Where Each One Wins

The AMD Radeon HD 8690M wins in the single most important category: benchmark performance. Its Geekbench OpenCL score of 6137 beats the HD 8730M by 3.1%. It also wins in clock speed (925 MHz base vs 650 MHz), memory bandwidth (32.00 GB/s vs 28.80 GB/s), pixel rate (7.800 GPixel/s vs 5.600 GPixel/s), texture rate (19.50 GTexel/s vs 16.80 GTexel/s), and FP32 compute (624.0 GFLOPS vs 537.6 GFLOPS).

The HD 8690M is the pick for general compute workloads, OpenCL applications, and any scenario where memory bandwidth matters. Its GDDR5 memory delivers more throughput than the HD 8730M's DDR3, and its higher clocks make it more responsive across the board.

The AMD Radeon HD 8730M wins in only one structural category: raw shader count. It has 384 shading units versus 320, and 24 TMUs versus 20. It also has a wider 128-bit memory bus, though this does not translate into higher bandwidth. These specifications suggest the HD 8730M might theoretically excel in workloads that scale perfectly with shader count and are not memory-bound. However, the benchmark data does not support this hypothesis — the HD 8730M loses the OpenCL test despite its compute advantage.

The HD 8730M also has a larger die (77 mm² vs 56 mm²) and more transistors (950 million vs 690 million), but these are manufacturing characteristics, not performance benefits. In practical terms, the HD 8730M's only real advantage is that it may be found in systems where the HD 8690M is not available, since both are end-of-life products.

For users choosing between these two mobile GPUs, the HD 8690M is the better choice for every measurable performance metric. The HD 8730M's higher shader count is a paper specification that does not overcome its clock and memory penalties. The benchmark record is clear: the HD 8690M wins the head-to-head, and the HD 8730M has no category where it outpaces its competitor in actual performance.