AMD Radeon HD 8730M vs NVIDIA GeForce GTX 680M Comparison

Head-to-Head Benchmarks

The recorded data contains one direct benchmark comparison between the NVIDIA GeForce GTX 680M and the AMD Radeon HD 8730M: the Geekbench OpenCL test. In this test, the NVIDIA GeForce GTX 680M scores 9230 points, while the AMD Radeon HD 8730M scores 5955 points. This translates to a 55% advantage for the NVIDIA part, a substantial margin that places the two mobile GPUs in clearly different performance tiers.

The GeForce GTX 680M also has a Geekbench Metal score of 4815, a test the AMD Radeon HD 8730M does not appear in. When averaging the available benchmark results, the NVIDIA GPU reaches an average benchmark score of 7023, while the AMD GPU averages 5955. This gap of approximately 17.9% in average score reinforces the OpenCL result, though the average is somewhat tempered by the Metal score being lower than the OpenCL score on the NVIDIA side.

Looking at the broader database context, the GeForce GTX 680M sits at the 39th percentile among all GPUs, while the Radeon HD 8730M sits at the 34th percentile. The percentile difference is modest, but the head-to-head delta is large. This suggests that while both parts occupy the lower half of the overall GPU landscape, the NVIDIA part is meaningfully stronger within that range.

The nearest rivals in the database provide additional context. The GeForce GTX 680M's closest competitors include the NVIDIA T600 with an average score of 7035 (0.2% behind the GTX 680M), the AMD Radeon R5 M240 with 6975 (0.7% behind), the NVIDIA GeForce GTX 675M with 6946 (1.1% behind), and the NVIDIA GeForce GTX 970 with 7157 (1.9% ahead). These tight margins indicate that the GTX 680M performs essentially at parity with a cluster of mid-range GPUs from several generations. The Radeon HD 8730M's nearest rivals are similarly close: the NVIDIA Quadro K620M at 5957 (0% difference), the AMD Radeon HD 8750M at 5970 (0.3% ahead of the HD 8730M), the Intel UHD Graphics 730 at 5929 (0.4% behind), and the NVIDIA Quadro K4000 at 5982 (0.5% ahead). The HD 8730M is thus firmly anchored in a lower performance bracket, with its closest competitors all within a fraction of a percent.

Architecture Differences

The two GPUs come from different architectural lineages. The NVIDIA GeForce GTX 680M is built on the Kepler architecture with the GK104 chip, part of the GeForce 600M generation. The AMD Radeon HD 8730M uses the GCN 1.0 architecture with the Mars chip, belonging to the Solar System (HD 8700M) generation. Both are fabricated by TSMC on a 28 nm process node, so process technology is not a differentiator.

The transistor budgets tell a story of scale. The GK104 packs 3,540 million transistors on a 294 mm² die, yielding a transistor density of 12.0 million transistors per mm². The Mars chip contains 950 million transistors on a 77 mm² die, with a slightly higher density of 12.3 million per mm². The NVIDIA chip is roughly 3.7 times larger in die area and holds about 3.7 times more transistors, yet the density figures are nearly identical, indicating both designs were near the practical limits of the 28 nm node at their respective sizes.

Compute resources diverge sharply. The GeForce GTX 680M features 1344 shading units, 112 texture mapping units, and 32 ROPs. The Radeon HD 8730M has 384 shading units, 24 TMUs, and 8 ROPs. In every category, the NVIDIA part has between 3.5 and 4 times the hardware resources. This explains the large performance gap in the OpenCL test, as the GTX 680M can process far more threads in parallel.

Clock speeds differ as well. The GTX 680M runs at a base clock of 719 MHz with a boost clock of 758 MHz. The HD 8730M operates at 650 MHz base and 700 MHz boost. The NVIDIA GPU thus has both more hardware and higher clocks, a combination that compounds its advantage. The resulting throughput figures reflect this: the GTX 680M delivers 2.038 TFLOPS of FP32 compute, 84.90 GTexel/s of texture fill rate, and 21.22 GPixel/s of pixel rate. The HD 8730M delivers 537.6 GFLOPS, 16.80 GTexel/s, and 5.600 GPixel/s, putting it at roughly one quarter of the NVIDIA part's throughput in each metric.

Memory configuration is another major split. The GTX 680M uses 4 GB of GDDR5 memory on a 256-bit bus, providing 115.2 GB/s of bandwidth. Memory clock is 900 MHz, with an effective data rate of 3.6 Gbps. The HD 8730M uses 2 GB of DDR3 memory on a 128-bit bus, delivering only 28.80 GB/s of bandwidth. Its memory clock is also 900 MHz, but the effective rate is 1800 Mbps. The bandwidth difference is a factor of 4, which matters significantly for texture-heavy workloads and higher-resolution rendering.

API support is broadly similar but not identical. Both support DirectX 12, but the GTX 680M implements version 11_0 while the HD 8730M implements version 11_1. OpenGL 4.6 is available on both. Vulkan support is present on both, with version 1.2.175 on the NVIDIA part and 1.2.170 on the AMD part. These are minor differences in practice for most workloads.

FAQ

Q: Which GPU is faster in the recorded benchmark data?

A: The NVIDIA GeForce GTX 680M wins the only direct head-to-head benchmark, the Geekbench OpenCL test, with a score of 9230 versus 5955 for the AMD Radeon HD 8730M, a 55% advantage.

Q: How do the memory subsystems compare?

A: The GTX 680M has 4 GB of GDDR5 on a 256-bit bus with 115.2 GB/s of bandwidth. The HD 8730M has 2 GB of DDR3 on a 128-bit bus with 28.80 GB/s of bandwidth, exactly one quarter of the NVIDIA part's bandwidth.

Q: Do both GPUs support the same DirectX version?

A: No. The GTX 680M supports DirectX 12 (11_0), while the HD 8730M supports DirectX 12 (11_1), a slightly newer feature level.

Q: What are the closest rivals to each GPU in the database?

A: The GTX 680M's nearest rival is the NVIDIA T600 with an average score of 7035, just 0.2% behind. The HD 8730M's nearest rival is the NVIDIA Quadro K620M with an average score of 5957, a 0% difference.

Q: Which GPU has a higher transistor density?

A: The AMD Radeon HD 8730M has a density of 12.3 million transistors per mm², marginally higher than the GTX 680M's 12.0 million per mm², despite having far fewer total transistors.

Q: Are both GPUs still in production?

A: No. Both the NVIDIA GeForce GTX 680M and the AMD Radeon HD 8730M are marked as end-of-life in the database.

Specification Differences

The following fields differ between the two GPUs:

• Chip: GK104 (NVIDIA) versus Mars (AMD)
• Architecture: Kepler versus GCN 1.0
• Generation: GeForce 600M versus Solar System (HD 8700M)
• Transistors: 3,540 million versus 950 million
• Die Size: 294 mm² versus 77 mm²
• Transistor Density: 12.0M / mm² versus 12.3M / mm²
• Base Clock: 719 MHz versus 650 MHz
• Boost Clock: 758 MHz versus 700 MHz
• Memory Clock: 900 MHz, 3.6 Gbps effective versus 900 MHz, 1800 Mbps effective
• Memory Size: 4 GB versus 2 GB
• Memory Type: GDDR5 versus DDR3
• Memory Bus Width: 256 bit versus 128 bit
• Memory Bandwidth: 115.2 GB/s versus 28.80 GB/s
• Shading Units: 1344 versus 384
• TMUs: 112 versus 24
• ROPs: 32 versus 8
• Pixel Rate: 21.22 GPixel/s versus 5.600 GPixel/s
• Texture Rate: 84.90 GTexel/s versus 16.80 GTexel/s
• FP32: 2.038 TFLOPS versus 537.6 GFLOPS
• TDP: 100 W versus null (not specified)
• Slot Width: MXM Module versus null
• Power Connectors: None versus null
• Bus Interface: MXM-B (3.0) versus PCIe 3.0 x8
• Display Outputs: Portable Device Dependent versus null
• DirectX: 12 (11_0) versus 12 (11_1)
• Vulkan: 1.2.175 versus 1.2.170
• Release Date: 2012-06-03 versus 2013-03-31
• Predecessor: GeForce 500M versus London
• Successor: GeForce 700M versus Gem System
• Benchmarks: Geekbench Metal 4815 and OpenCL 9230 versus OpenCL 5955 only
• Percentile: 39 versus 34
• Average Benchmark Score: 7023 versus 5955

The NVIDIA part is larger, faster, more feature-rich, and was released earlier. The AMD part is smaller, less power-hungry by design philosophy (though its TDP is not recorded), and uses a more conventional PCIe interface.

Where Each One Wins

The NVIDIA GeForce GTX 680M wins in every category where data exists. It wins the sole head-to-head benchmark by 55%. It has higher raw compute throughput, with 2.038 TFLOPS versus 537.6 GFLOPS. It has four times the memory bandwidth, which benefits workloads that stream large textures or geometry. It has more than three times the shading units, TMUs, and ROPs, which translates directly to better performance in pixel-heavy and texture-heavy rendering.

The GTX 680M also holds the only Metal benchmark score in the comparison, 4815, which the HD 8730M cannot contest due to lack of recorded data. Its average benchmark score of 7023 places it in a competitive cluster with the NVIDIA T600, GTX 675M, and GTX 970, all within roughly 2% of each other. This means the GTX 680M is a viable option for workloads that sit in that mid-range performance band.

The AMD Radeon HD 8730M wins no benchmark categories. Its advantage, if any, lies in its smaller physical footprint: the 77 mm² die and 950 million transistors represent a much lighter silicon load, and the absence of a recorded TDP suggests it was positioned for more power-constrained environments. Its PCIe 3.0 x8 interface is also a more standard integration path than the proprietary MXM-B (3.0) connector used by the NVIDIA part. For systems that require a compact, low-complexity GPU with modest performance needs, the HD 8730M is a reasonable fit, but it does not win any recorded performance comparison.

The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 680M is the stronger GPU in every measured metric. Its 55% lead in OpenCL, its 4x memory bandwidth advantage, and its substantially larger compute resource pool make it the clear choice for any workload that stresses the GPU. Users who need 3D rendering, compute acceleration, or higher-resolution gaming should select the GTX 680M.

The AMD Radeon HD 8730M is suited to a narrower set of use cases. Its smaller die and lower transistor count indicate a design aimed at basic display output and light acceleration rather than demanding graphics work. For systems where space, power, and thermal constraints are paramount, and where the workload is limited to everyday tasks, the HD 8730M can suffice. However, the database shows no scenario where it outperforms the GTX 680M.

The competitive context reinforces this conclusion. The GTX 680M sits among rivals like the T600 and GTX 970, all within 2% of its average score, while the HD 8730M sits among much slower parts like the Quadro K620M and Intel UHD Graphics 730. The GTX 680M is effectively two performance brackets above the HD 8730M. Given that both are end-of-life products, the decision hinges on whether the workload justifies the larger, more capable NVIDIA part or whether the lighter AMD part meets the requirement with less silicon. For performance, the choice is the GTX 680M.