AMD Radeon 680M vs NVIDIA GeForce RTX 5090 D V2 Comparison

The NVIDIA GeForce RTX 5090 D V2 and the AMD Radeon 680M occupy opposite ends of the graphics spectrum, and the benchmark data reflects that chasm. In the single shared 3DMark Steel Nomad DX12 test, the RTX 5090 D V2 delivers a score of 16,504 against the Radeon 680M’s 378, a delta of 4,266.1%. However, the Radeon 680M counters with its own benchmark wins in compute and cross-platform APIs, which the RTX 5090 D V2 does not contest. The data shows a discrete flagship designed for maximum rasterization throughput versus an integrated processor aimed at efficiency and versatility.

Where Each One Wins

The RTX 5090 D V2 wins decisively in pure DirectX 12 gaming performance. Its 3DMark Steel Nomad score of 16,504 places it in the 59th percentile of all GPUs, with its nearest rival being the NVIDIA T400 at 16,508 (0% delta) and the AMD Radeon PRO W7500 at 16,415 (0.5% behind). This is a top-tier result for a single benchmark, indicating that the card is built to handle demanding 4K and high-refresh-rate workloads without compromise. The Radeon 680M, by contrast, scores 378 in the same test, placing it in the 57th percentile overall — a figure that reflects its integrated nature rather than a failure of execution.

The Radeon 680M wins in two additional benchmarks that the RTX 5090 D V2 does not appear in: Geekbench OpenCL and Geekbench Vulkan. Its OpenCL score of 23,468 and Vulkan score of 21,965 demonstrate that it is not merely a weak gaming chip but a capable compute and general-purpose processor. These scores are likely aided by its system-shared memory architecture, which allows flexible allocation of resources. In this context, the Radeon 680M wins on software ecosystem coverage, offering measurable performance in APIs where the RTX 5090 D V2 has no listed result.

The use-case split is clear: the RTX 5090 D V2 is for users who need uncompromising DirectX 12 gaming and rendering performance, while the Radeon 680M is for portable or low-power systems where compute versatility and API breadth matter more than raw rasterization. The data shows that each card wins where its architecture is designed to excel.

Architecture Differences

The two GPUs are built on fundamentally different architectural philosophies. The RTX 5090 D V2 uses NVIDIA’s Blackwell 2.0 architecture on a 5 nm TSMC process, with a massive die size of 750 mm² and 92,200 million transistors. This yields a transistor density of 122.9 million per mm². The Radeon 680M uses AMD’s RDNA 2.0 on a 6 nm TSMC process, with a die size of 208 mm² and 13,100 million transistors, resulting in a density of 63.0 million per mm². The RTX 5090 D V2’s die is over 3.6 times larger, and its transistor count is over 7 times higher.

The compute resources differ by an order of magnitude. The RTX 5090 D V2 has 21,760 shading units, 680 texture mapping units, and 176 raster output units, alongside 170 ray tracing cores and 680 tensor cores. The Radeon 680M has 768 shading units, 48 TMUs, and 32 ROPs, with 12 ray tracing cores and no tensor cores. This translates to a pixel rate of 423.6 GPixel/s and a texture rate of 1,636.8 GTexel/s for the RTX 5090 D V2, versus 70.40 GPixel/s and 105.6 GTexel/s for the Radeon 680M.

Memory architecture is another key differentiator. The RTX 5090 D V2 features 24 GB of GDDR7 on a 384-bit bus, delivering 1.34 TB/s of bandwidth. The Radeon 680M uses system-shared memory with system-dependent bandwidth, making its performance highly dependent on the host system's RAM. The RTX 5090 D V2’s FP32 throughput is 104.8 TFLOPS, while the Radeon 680M manages 3.379 TFLOPS — a 31x gap. The Radeon 680M does offer FP16 at 6.758 TFLOPS (2:1 ratio), whereas the RTX 5090 D V2 does FP16 at 1:1 with its FP32 rate, but this advantage is purely theoretical given the raw scale.

Head-to-Head Benchmarks

The single direct comparison is 3DMark Steel Nomad DX12. The RTX 5090 D V2 scores 16,504, while the Radeon 680M scores 378, giving the NVIDIA card a 4,266.1% advantage. This is not a marginal win; it is a categorical obliteration. The RTX 5090 D V2’s score is 43.7 times higher than the Radeon 680M’s, reflecting the difference between a 575 W dual-slot discrete card and a 50 W integrated processor.

To contextualize the RTX 5090 D V2’s result, its nearest rivals in the overall database are the NVIDIA T400 (16,508, 0% delta), AMD Radeon PRO W7500 (16,415, 0.5% behind), and NVIDIA RTX PRO 6000 Blackwell (16,408, 0.6% behind). These are all professional or workstation-class cards, underscoring that the RTX 5090 D V2 is competing at the very top of the performance stack. The Radeon 680M’s nearest rivals in its own average benchmark score include the NVIDIA GeForce GTX 580 (15,283, 0.1% behind) and the NVIDIA GeForce RTX 2060 (15,290, 0.1% behind), which are older discrete cards — yet the Radeon 680M still trails them in the Steel Nomad test specifically.

The Radeon 680M’s wins come in Geekbench OpenCL (23,468) and Vulkan (21,965). These scores are not directly comparable to the RTX 5090 D V2 due to the lack of a head-to-head entry, but they show that the integrated GPU is not a slouch in compute tasks. The Vulkan score of 21,965 is particularly notable, as it suggests the Radeon 680M can handle cross-platform graphics workloads with reasonable efficiency, albeit at a far lower absolute performance level than the RTX 5090 D V2’s DirectX 12 output.

The Verdict

The data is unambiguous. The NVIDIA GeForce RTX 5090 D V2 is the clear choice for any workload that demands maximum DirectX 12 rasterization performance. Its 3DMark Steel Nomad score of 16,504 places it among professional-grade GPUs, and its 4,266.1% lead over the Radeon 680M in that test leaves no room for interpretation. Users who need to drive high-resolution, high-refresh-rate gaming or GPU-accelerated rendering should choose the RTX 5090 D V2 without hesitation.

The AMD Radeon 680M is the right pick for systems where power efficiency and integration are paramount. Its 50 W TDP, IGP form factor, and system-shared memory make it suitable for thin-and-light laptops or compact desktops. Its Geekbench scores (23,468 OpenCL, 21,965 Vulkan) demonstrate that it can handle general compute tasks, and its 57th percentile overall ranking is respectable for an integrated solution. However, the data shows that it cannot compete with the RTX 5090 D V2 in any head-to-head gaming scenario.

The verdict: pick the RTX 5090 D V2 for performance-critical applications, and pick the Radeon 680M for portable, low-power systems where its compute versatility and API coverage are sufficient.

FAQ

Q: How much faster is the RTX 5090 D V2 than the Radeon 680M in 3DMark Steel Nomad?

A: The RTX 5090 D V2 scores 16,504 versus the Radeon 680M’s 378, a 4,266.1% advantage.

Q: Does the Radeon 680M win in any benchmarks?

A: Yes, it wins in Geekbench OpenCL with a score of 23,468 and Geekbench Vulkan with 21,965, though the RTX 5090 D V2 has no listed results in these tests.

Q: What are the RTX 5090 D V2’s nearest rivals in the database?

A: The NVIDIA T400 (16,508, 0% delta), AMD Radeon PRO W7500 (16,415, 0.5% behind), and NVIDIA RTX PRO 6000 Blackwell (16,408, 0.6% behind).

Q: What is the Radeon 680M’s nearest rival by average score?

A: The NVIDIA GeForce GTX 580 (15,283, 0.1% behind) and NVIDIA GeForce RTX 2060 (15,290, 0.1% behind).

Q: What is the transistor count difference between the two GPUs?

A: The RTX 5090 D V2 has 92,200 million transistors, while the Radeon 680M has 13,100 million.

Q: Which GPU has a higher FP32 throughput?

A: The RTX 5090 D V2 at 104.8 TFLOPS, compared to the Radeon 680M’s 3.379 TFLOPS.

Specification Differences

| Specification | NVIDIA GeForce RTX 5090 D V2 | AMD Radeon 680M |

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

| Architecture | Blackwell 2.0 | RDNA 2.0 |

| Process Node | 5 nm | 6 nm |

| Die Size | 750 mm² | 208 mm² |

| Transistors | 92,200 million | 13,100 million |

| Transistor Density | 122.9M / mm² | 63.0M / mm² |

| Base Clock | 2017 MHz | 2000 MHz |

| Boost Clock | 2407 MHz | 2200 MHz |

| Memory Size | 24 GB | System Shared |

| Memory Type | GDDR7 | System Shared |

| Memory Bus Width | 384 bit | System Shared |

| Memory Bandwidth | 1.34 TB/s | System Dependent |

| Shading Units | 21,760 | 768 |

| TMUs | 680 | 48 |

| ROPs | 176 | 32 |

| Ray Tracing Cores | 170 | 12 |

| Tensor Cores | 680 | None |

| Pixel Rate | 423.6 GPixel/s | 70.40 GPixel/s |

| Texture Rate | 1,636.8 GTexel/s | 105.6 GTexel/s |

| FP32 Performance | 104.8 TFLOPS | 3.379 TFLOPS |

| FP16 Performance | 104.8 TFLOPS (1:1) | 6.758 TFLOPS (2:1) |

| TDP | 575 W | 50 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 1x 16-pin | None |

| Bus Interface | PCIe 5.0 x16 | PCIe 4.0 x8 |

| Launch MSRP | 2,299 USD | Not specified |