AMD Radeon HD 8690M vs NVIDIA GeForce GTX 680M Comparison

Head-to-Head Benchmarks

The database records a single direct benchmark comparison between these two mobile graphics processors: Geekbench OpenCL. In that test, the NVIDIA GeForce GTX 680M scores 9,230, while the AMD Radeon HD 8690M scores 6,137. The result is a decisive win for the NVIDIA part, with a delta of 50.4 percent. That is not a marginal lead; it is a full one-and-a-half times the AMD score. The GTX 680M does not just win, it dominates the compute workload measured by OpenCL.

Looking at the broader database context, the GTX 680M holds an average benchmark score of 7,023 across all recorded tests, while the HD 8690M averages 6,137. The NVIDIA card sits at the 39th percentile of all GPUs in the database, and the AMD card sits at the 36th percentile. Both are below the median, but the GTX 680M is clearly the stronger performer of the two. The nearest rivals for the GTX 680M include the NVIDIA T600 with an average score of 7,035 (0.2 percent ahead), the AMD Radeon R5 M240 at 6,975 (0.7 percent behind), and the NVIDIA GeForce GTX 675M at 6,946 (1.1 percent behind). The GTX 680M is essentially neck-and-neck with these parts, all within roughly two percent of each other. The only rival that pulls further ahead is the NVIDIA GeForce GTX 970, scoring 7,157, which is 1.9 percent above the GTX 680M. So the GTX 680M sits in a tight cluster of mid-range performers, and its OpenCL score is higher than its average, indicating that specific compute tasks favor it even more.

The HD 8690M, in contrast, is surrounded by a different set of rivals. The Intel Iris Pro Graphics 6200 scores 6,117, which is 0.3 percent behind the AMD part. The NVIDIA RTX A400 scores 6,078 (1.0 percent behind), and the NVIDIA GeForce MX230 scores 6,077 (also 1.0 percent behind). The AMD Radeon HD 6950 scores 6,210, which is 1.2 percent ahead. This places the HD 8690M in a group of entry-level and integrated-class GPUs, all within a narrow band of about two percent. The GTX 680M, by contrast, is grouped with discrete mobile parts that are a full tier higher in performance.

The head-to-head delta of 50.4 percent is the single most important number in this comparison. It is not a case of one card being slightly better in one workload; it is a case of the NVIDIA part being dramatically faster in the only test where both were measured. The GTX 680M's OpenCL score of 9,230 is also higher than its own average, suggesting that the card's compute capabilities are stronger than its overall gaming or graphics average would imply. The HD 8690M's OpenCL score of 6,137 is exactly its average, meaning that test is representative of its overall capability.

Where Each One Wins

The GTX 680M wins the only recorded head-to-head benchmark, and it wins by a wide margin. In Geekbench OpenCL, it is 50.4 percent ahead of the HD 8690M. This is the only test where both cards have data, so the win count is 1 for NVIDIA and 0 for AMD. There is no recorded test where the AMD card comes out ahead. If you are looking for a GPU to handle OpenCL compute tasks, whether that is video encoding, physics simulation, or other general-purpose GPU workloads, the GTX 680M is the clear choice based on the numbers.

The HD 8690M does have a win in the sense of efficiency per transistor or per die area, but that is not a benchmark win. Looking at the specification data, the AMD chip uses a much smaller die: 56 mm² versus 294 mm², and far fewer transistors: 690 million versus 3,540 million. Yet it still delivers 6,137 in OpenCL, which is 66.5 percent of the GTX 680M's score. That means the AMD part is far more efficient in terms of performance per square millimeter of silicon. But in absolute terms, the NVIDIA part wins every workload where both are measured.

For gaming, the database does not include a DirectX or OpenGL gaming benchmark for these two cards. However, the compute scores are often correlated with overall graphics capability, and the GTX 680M has more shading units (1,344 versus 320), more texture mapping units (112 versus 20), and more render output units (32 versus 8). These hardware differences strongly suggest that the GTX 680M would also win in rasterization-heavy tasks, though the database does not record such a test directly. The pixel rate for the GTX 680M is 21.22 GPixel/s versus 7.80 GPixel/s for the AMD card, and the texture rate is 84.90 GTexel/s versus 19.50 GTexel/s. These are not benchmark scores, but they are derived from clock speeds and unit counts, and they show a consistent pattern of NVIDIA superiority.

The HD 8690M does have one advantage: its boost clock is higher (975 MHz versus 758 MHz), and its base clock is higher as well (925 MHz versus 719 MHz). But those raw clock speeds do not translate into performance wins because the NVIDIA card has vastly more execution units. The AMD card's higher clocks are a classic case of a smaller, simpler chip running at higher frequencies to try to close the gap, but it cannot overcome the sheer difference in parallel hardware.

Architecture Differences

The NVIDIA GeForce GTX 680M is built on the Kepler architecture, specifically the GK104 chip, manufactured by TSMC on a 28 nm process. It uses 3,540 million transistors on a die size of 294 mm², giving a transistor density of 12.0 million per mm². The AMD Radeon HD 8690M is built on GCN 1.0 (Graphics Core Next), using the Sun chip, also manufactured by TSMC on the same 28 nm process. It uses 690 million transistors on a die size of 56 mm², giving a transistor density of 12.3 million per mm². The density is nearly identical, which makes sense since both are on the same node from the same foundry. The difference is purely in the scale of the chip: the NVIDIA part is roughly five times larger in die area and has over five times as many transistors.

The memory subsystems are fundamentally different. The GTX 680M has 4 GB of GDDR5 on a 256-bit bus, yielding a bandwidth of 115.2 GB/s. The HD 8690M has 2 GB of GDDR5 on a 64-bit bus, yielding a bandwidth of only 32.00 GB/s. That is a 3.6 times difference in memory bandwidth, which is critical for high-resolution textures and compute workloads that access large datasets. The GTX 680M's memory clock is 900 MHz (3.6 Gbps effective), while the AMD card's memory clock is 1000 MHz (4 Gbps effective). The AMD card has faster memory chips, but the narrow 64-bit bus severely limits the total throughput.

The compute unit counts tell the same story. The GTX 680M has 1,344 shading units, 112 TMUs, and 32 ROPs. The HD 8690M has 320 shading units, 20 TMUs, and 8 ROPs. In terms of raw math, the GTX 680M delivers 2.038 TFLOPS of FP32 compute, while the AMD card delivers 624.0 GFLOPS, which is 0.624 TFLOPS. The NVIDIA part is 3.27 times faster in theoretical FP32 throughput. Neither card has dedicated ray tracing cores or tensor cores, as those features came later. Both support DirectX 12, but the NVIDIA card is at feature level 11_0 while the AMD card is at 11_1. Both support OpenGL 4.6. The NVIDIA driver supports Vulkan 1.2.175, and the AMD driver supports Vulkan 1.2.170, a minor version difference.

The bus interface also differs: the GTX 680M uses MXM-B (3.0), which is a modular mobile graphics standard, while the HD 8690M uses PCIe 3.0 x8. The power connector situation is different as well: the GTX 680M has no power connectors, relying on the MXM slot for power, and its TDP is rated at 100 W. The AMD card has no TDP listed in the database, and no power connector information is available. The GTX 680M is a larger, more power-hungry part, which is typical for a high-end mobile GPU. The AMD card is a smaller, lower-power part designed for thinner and lighter laptops.

The Verdict

Based strictly on the recorded data, the NVIDIA GeForce GTX 680M is the superior GPU in every measurable way. It wins the only head-to-head benchmark by 50.4 percent. It has a higher average benchmark score: 7,023 versus 6,137. It has a higher percentile ranking: 39th versus 36th. It has more memory, more bandwidth, more shading units, more texture units, more ROPs, and more compute throughput. The only numeric advantages the AMD card holds are its higher base and boost clocks, and its smaller die size and transistor count, which are not performance advantages but rather efficiency or cost advantages.

Who should pick the GTX 680M? Anyone who needs maximum compute performance from a mobile GPU. The OpenCL score of 9,230 puts it in a class above its average, suggesting it is particularly strong for general-purpose GPU compute. It also has 4 GB of VRAM, which is useful for larger datasets or higher-resolution textures. The 256-bit memory bus provides 115.2 GB/s of bandwidth, which is essential for memory-heavy workloads. This card is for a gaming laptop or a mobile workstation where performance is the priority.

Who should pick the HD 8690M? The data does not support choosing it for performance reasons. It loses the only benchmark, and it has roughly one-third of the compute throughput. However, if you are looking at a smaller, lighter laptop, the AMD card's much smaller die (56 mm² versus 294 mm²) and lower transistor count (690 million versus 3,540 million) suggest it consumes less power and generates less heat, though the database does not list a TDP for it. For light tasks like office work, media playback, or very light gaming, the HD 8690M may be sufficient, and it will fit in a chassis that cannot accommodate a 100 W MXM module. But if you have the choice and the laptop can physically fit the GTX 680M, the performance gap is overwhelming.

The database shows no scenario where the HD 8690M wins. The verdict is straightforward: the GTX 680M is the faster card, and the HD 8690M is the more efficient card. Choose based on which of those two priorities matters more for your use case.

FAQ

Q: How much faster is the NVIDIA GeForce GTX 680M than the AMD Radeon HD 8690M in the recorded benchmark?

A: In Geekbench OpenCL, the GTX 680M scores 9,230 versus 6,137 for the HD 8690M, a delta of 50.4 percent.

Q: Which card has more memory bandwidth?

A: The GTX 680M has 115.2 GB/s of bandwidth with a 256-bit bus, while the HD 8690M has 32.00 GB/s with a 64-bit bus.

Q: What are the average benchmark scores for each card?

A: The GTX 680M has an average benchmark score of 7,023, and the HD 8690M has an average of 6,137.

Q: Do either of these cards support DirectX 12?

A: Yes, both support DirectX 12. The GTX 680M is at feature level 11_0, and the HD 8690M is at feature level 11_1.

Q: Which card has a higher boost clock?

A: The HD 8690M has a boost clock of 975 MHz, while the GTX 680M has a boost clock of 758 MHz. However, the GTX 680M still performs much better due to its larger number of execution units.

Q: What is the transistor count difference?

A: The GTX 680M has 3,540 million transistors, while the HD 8690M has 690 million transistors.

Specification Differences

| Specification | NVIDIA GeForce GTX 680M | AMD Radeon HD 8690M |

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

| Architecture | Kepler | GCN 1.0 |

| Chip | GK104 | Sun |

| Generation | GeForce 600M | Solar System (HD 8600M) |

| Process Node | 28 nm | 28 nm |

| Transistors | 3,540 million | 690 million |

| Die Size | 294 mm² | 56 mm² |

| Transistor Density | 12.0M / mm² | 12.3M / mm² |

| Base Clock | 719 MHz | 925 MHz |

| Boost Clock | 758 MHz | 975 MHz |

| Memory Clock | 900 MHz (3.6 Gbps effective) | 1000 MHz (4 Gbps effective) |

| Memory Size | 4 GB | 2 GB |

| Memory Type | GDDR5 | GDDR5 |

| Memory Bus Width | 256 bit | 64 bit |

| Memory Bandwidth | 115.2 GB/s | 32.00 GB/s |

| Shading Units | 1344 | 320 |

| TMUs | 112 | 20 |

| ROPs | 32 | 8 |

| Pixel Rate | 21.22 GPixel/s | 7.800 GPixel/s |

| Texture Rate | 84.90 GTexel/s | 19.50 GTexel/s |

| FP32 Compute | 2.038 TFLOPS | 624.0 GFLOPS |

| TDP | 100 W | Not listed |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x8 |

| DirectX Support | 12 (11_0) | 12 (11_1) |

| OpenGL Support | 4.6 | 4.6 |

| Vulkan Support | 1.2.175 | 1.2.170 |

| Release Date | 2012-06-03 | 2013-02-28 |

| Predecessor | GeForce 500M | London |

| Successor | GeForce 700M | Gem System |

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