AMD Radeon 680M vs NVIDIA GeForce RTX 5090 SE Comparison

The AMD Radeon 680M and NVIDIA GeForce RTX 5090 SE sit at opposite ends of the hardware spectrum. The 680M is an integrated graphics processor for mobile devices, while the 5090 SE is a flagship discrete desktop card. The recorded benchmark data is limited, but the available measurements and technical specifications provide a clear picture of their relative positioning.

Head-to-Head Benchmarks

Direct head-to-head benchmark comparisons are absent from the database. However, the AMD Radeon 680M has recorded scores in three separate tests. In the 3DMark Steel Nomad DX12 test, it scored 378 points. Its Geekbench OpenCL score reached 23468, while the Geekbench Vulkan score came in at 21965. These results produce an average benchmark score of 15270 for the 680M.

The NVIDIA GeForce RTX 5090 SE currently has no recorded benchmark scores in the database. Its average benchmark score stands at zero, and there are no head-to-head results available. The percentile data confirms the gap: the 680M sits at the 57th percentile among all GPUs, while the 5090 SE is listed at the 50th percentile, though that figure is based on the absence of measured data rather than actual performance.

The 680M's nearest rivals in the database help contextualize its scores. The NVIDIA GeForce GTX 580 has an average score of 15283, a delta of -0.1% compared to the 680M. The NVIDIA GeForce RTX 2060 scores 15290, also -0.1% relative to the 680M. The RTX 3050 OEM trails by 0.5% with a score of 15199, and the AMD Radeon RX 7600 is 0.7% behind at 15171. These results indicate the integrated 680M performs roughly on par with older mid-range discrete graphics cards from several generations ago, a notable achievement for an IGP.

Architecture Differences

The two processors share a common TSMC foundry but use different process nodes. The AMD Radeon 680M uses a 6 nm node, while the NVIDIA GeForce RTX 5090 SE uses a 5 nm node. The transistor counts differ dramatically: the 680M contains 13,100 million transistors on a 208 mm² die, yielding a transistor density of 63.0M per mm². The 5090 SE packs 92,200 million transistors onto a 750 mm² die, achieving 122.9M per mm², nearly double the density.

Architecturally, the 680M is built on RDNA 2.0, part of the Navi II IGP generation for Rembrandt Mobile. The 5090 SE uses Blackwell 2.0, belonging to the GeForce 50 series. The 680M's chip is codenamed Rembrandt+, while the 5090 SE uses the GB202 chip. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API compatibility is equivalent.

The memory subsystems are fundamentally different. The 680M relies on System Shared memory, with a System Shared type, bus width, and System Dependent bandwidth. The 5090 SE has 24 GB of GDDR7 memory on a 384 bit bus, delivering 1.34 TB/s of bandwidth. The 680M's memory clock is listed as System Shared, while the 5090 SE runs at 1750 MHz with 28 Gbps effective speed.

Clock speeds show a closer relationship than the rest of the specifications might suggest. The 680M has a base clock of 2000 MHz and a boost clock of 2200 MHz. The 5090 SE starts lower at 1740 MHz base but boosts higher to 2377 MHz. The integrated part actually has a higher base clock, though the discrete card's boost ceiling is higher.

Where Each One Wins

The AMD Radeon 680M wins in power efficiency and physical footprint. Its TDP is 50 W, and it is an IGP with no slot width, no power connectors, and no suggested PSU requirement. It uses a PCIe 4.0 x8 bus interface. Display outputs are listed as Portable Device Dependent, meaning the 680M is designed for laptops and compact mobile systems where space and power are at a premium. Its 6 nm process and smaller die also contribute to its suitability for integrated applications.

The NVIDIA GeForce RTX 5090 SE wins decisively in raw computational resources. It has 14080 shading units compared to 768 on the 680M, 440 TMUs versus 48, and 160 ROPs versus 32. The 5090 SE includes 110 RT cores and 440 tensor cores, while the 680M has 12 RT cores and no tensor cores. The pixel rate for the 5090 SE is 380.3 GPixel/s against 70.40 GPixel/s for the 680M. Texture rate is 1,045.9 GTexel/s versus 105.6 GTexel/s. FP32 performance reaches 66.94 TFLOPS on the 5090 SE, compared to 3.379 TFLOPS on the 680M. FP16 output is 66.94 TFLOPS (1:1) on the NVIDIA card, while the AMD part delivers 6.758 TFLOPS (2:1).

The 5090 SE also wins on memory capacity and bandwidth. The 24 GB GDDR7 configuration dwarfs the System Shared memory of the 680M. The 1.34 TB/s bandwidth is a fundamental advantage for high-resolution textures and large datasets. The NVIDIA card uses a PCIe 5.0 x16 interface, double the lanes and a newer standard than the 680M's PCIe 4.0 x8.

The power envelope tells the rest of the story. The 5090 SE has a 500 W TDP, requires a dual-slot cooler, uses a 1x 16-pin power connector, and suggests a 900 W PSU. It measures 267 mm in length, 111 mm in height, and 40 mm in width. The 680M requires none of that infrastructure.

FAQ

Q: How does the AMD Radeon 680M compare to its nearest rivals in the database?

A: The 680M scores 15270 on average. The GTX 580 is 0.1% higher at 15283, the RTX 2060 is 0.1% higher at 15290, the RTX 3050 OEM is 0.5% lower at 15199, and the RX 7600 is 0.7% lower at 15171.

Q: What is the transistor density difference between the two cards?

A: The 680M has a density of 63.0M transistors per mm², while the 5090 SE achieves 122.9M per mm². The 5090 SE also holds more total transistors: 92,200 million versus 13,100 million.

Q: Do both cards support the same DirectX version?

A: Yes, both support DirectX 12 Ultimate (12_2). They also both support OpenGL 4.6 and Vulkan 1.4.

Q: What memory configuration does each card use?

A: The 680M uses System Shared memory with System Dependent bandwidth. The 5090 SE has 24 GB of GDDR7 on a 384 bit bus with 1.34 TB/s bandwidth.

Q: How do the boost clocks compare?

A: The 680M boosts to 2200 MHz, while the 5090 SE boosts to 2377 MHz. The 680M has a higher base clock at 2000 MHz versus 1740 MHz.

Q: What is the power requirement difference?

A: The 680M has a 50 W TDP with no power connectors. The 5090 SE has a 500 W TDP, requires a 1x 16-pin power connector, and suggests a 900 W PSU.

The Verdict

The data shows two products designed for entirely different use cases. The AMD Radeon 680M delivers integrated graphics performance that lands within 1% of older discrete cards like the GTX 580 and RTX 2060. Its 57th percentile ranking among all GPUs, combined with a 50 W TDP and no additional power hardware, makes it a sensible choice for portable systems where space and battery life take priority.

The NVIDIA GeForce RTX 5090 SE is a different class of hardware altogether. Its 66.94 TFLOPS FP32 output is roughly 19.8 times the 680M's 3.379 TFLOPS. The 5090 SE carries 14080 shading units, 110 RT cores, and 440 tensor cores, none of which the 680M can approach. Its 1.34 TB/s memory bandwidth and 24 GB GDDR7 capacity serve workloads that the 680M's System Shared memory cannot handle. The 5090 SE's 380.3 GPixel/s pixel rate and 1,045.9 GTexel/s texture rate indicate it is built for high-resolution rendering and compute-heavy tasks.

The 5090 SE occupies the 50th percentile in the database, but that figure reflects the absence of benchmark results rather than actual performance. A card with 14080 shading units and 92,200 million transistors is positioned for desktop workstations and high-end gaming rigs. The 680M's 768 shading units and 13,100 million transistors are appropriate for integrated mobile graphics.

The choice depends on the platform. The 680M fits systems with no expansion slots, no discrete power delivery, and tight thermal budgets. The 5090 SE requires a dual-slot chassis, a 16-pin power connector, and a 900 W PSU suggestion. The 680M's PCIe 4.0 x8 interface suits bandwidth-constrained mobile designs, while the 5090 SE's PCIe 5.0 x16 interface takes full advantage of modern desktop motherboards. The 680M is the only option for a portable device, while the 5090 SE is the only option for maximum throughput.

Specification Differences

The table below lists only the fields where the two differ, based on the recorded data.

| Specification | AMD Radeon 680M | NVIDIA GeForce RTX 5090 SE |

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

| Series | None | GeForce 50-series |

| Architecture | RDNA 2.0 | Blackwell 2.0 |

| Generation | Navi II IGP (Rembrandt Mobile) | GeForce 50 |

| Process Node | 6 nm | 5 nm |

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

| Die Size | 208 mm² | 750 mm² |

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

| Base Clock | 2000 MHz | 1740 MHz |

| Boost Clock | 2200 MHz | 2377 MHz |

| Memory Clock | System Shared | 1750 MHz 28 Gbps effective |

| Memory Size | System Shared | 24 GB |

| Memory Type | System Shared | GDDR7 |

| Memory Bus Width | System Shared | 384 bit |

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

| Shading Units | 768 | 14080 |

| TMUs | 48 | 440 |

| ROPs | 32 | 160 |

| RT Cores | 12 | 110 |

| Tensor Cores | None | 440 |

| Pixel Rate | 70.40 GPixel/s | 380.3 GPixel/s |

| Texture Rate | 105.6 GTexel/s | 1,045.9 GTexel/s |

| FP32 | 3.379 TFLOPS | 66.94 TFLOPS |

| FP16 | 6.758 TFLOPS (2:1) | 66.94 TFLOPS (1:1) |

| TDP | 50 W | 500 W |

| Slot Width | IGP | Dual-slot |

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

| Suggested PSU | None | 900 W |

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

| Display Outputs | Portable Device Dependent | 1x HDMI 2.1b3x DisplayPort 2.1b |

| Dimensions | None | 267 mm 10.5 inches, 111 mm 4.4 inches, 40 mm 1.6 inches |

| Release Date | 2023-01-02 | 2025-12-31 |

| Predecessor | Vega II IGP | GeForce 40 |

| Successor | Navi III IGP | GeForce 60 |

| Launch MSRP | None | 1,499 USD |

| Benchmarks | 378, 23468, 21965 | None |

| Average Benchmark Score | 15270 | 0 |

| Percentile | 57 | 50 |