AMD Radeon 680M vs NVIDIA GeForce GTX 580 Comparison

The Verdict

The data is clear: the AMD Radeon 680M is the stronger GPU in raw compute, winning the only direct head-to-head OpenCL benchmark by a decisive 34.9% margin. The NVIDIA GeForce GTX 580, despite being a 2010-era discrete card, remains competitive in aggregate scores but is outclassed in modern workloads. For users needing legacy PCIe 2.0 compatibility, dual-slot discrete mounting, or a fixed 1536 MB GDDR5 frame buffer, the GTX 580 is the only choice. For anyone else — especially those in portable systems or constrained power envelopes — the Radeon 680M is the data-backed pick, delivering 3.379 TFLOPS FP32 versus the GTX 580’s 1.581 TFLOPS while consuming only 50 W against 244 W.

The average benchmark scores are nearly identical — 15283 for the GTX 580 and 15270 for the Radeon 680M — placing both within 0.1% of each other and within 0.7% of the NVIDIA GeForce RTX 3050 OEM. This near-tie in aggregate hides a stark split: the Radeon 680M wins the OpenCL test outright with 23468 points, while the GTX 580’s entire benchmark profile rests on a single 15283 OpenCL result. The Radeon 680M also posts a Vulkan score of 21965 and a 3DMark Steel Nomad DX12 score of 378, neither of which the GTX 580 can attempt due to its lack of Vulkan support and older DirectX 12 (11_0) feature level.

Architecture Differences

The two GPUs represent opposite ends of the hardware spectrum. The NVIDIA GeForce GTX 580 uses the GF110 chip on Fermi 2.0 architecture, built on TSMC’s 40 nm process with 3,000 million transistors on a 520 mm² die. The AMD Radeon 680M uses Rembrandt+ silicon on RDNA 2.0 architecture, fabricated on TSMC’s 6 nm node with 13,100 million transistors packed into a 208 mm² die. This yields a dramatic transistor density gap: 5.8M transistors per mm² for Fermi versus 63.0M per mm² for RDNA 2.0 — an 11x density advantage for the AMD part.

The Radeon 680M is an integrated graphics processor (IGP) with no standalone dimensions, no power connectors, and a 50 W TDP. It draws memory from the system pool, making bandwidth "System Dependent" with a 2000 MHz base and 2200 MHz boost clock. The GTX 580 is a dual-slot discrete card measuring 267 mm in length and 111 mm in height, requiring a 550 W suggested PSU, one 6-pin and one 8-pin power connector, and featuring 1536 MB of GDDR5 on a 384-bit bus with 192.4 GB/s bandwidth.

Core configuration favors the Radeon 680M in shading units (768 versus 512) but the GTX 580 leads in texture mapping units (64 versus 48) and render output units (48 versus 32). The Radeon 680M adds 12 ray tracing cores, a feature entirely absent from the Fermi-based GTX 580. The AMD part also supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 580 is limited to DirectX 12 (11_0) and OpenGL 4.6 with no Vulkan support.

Head-to-Head Benchmarks

The sole direct comparison available is the Geekbench OpenCL test, and the result is lopsided. The AMD Radeon 680M scores 23468, while the NVIDIA GeForce GTX 580 scores 15283 — a 34.9% advantage for the Radeon. This is not a marginal win; it is a substantial margin that reflects the architectural gulf between a 2010 Fermi part and a 2023 RDNA 2.0 IGP.

The GTX 580’s OpenCL score of 15283 places it within 0.1% of the Radeon 680M’s average score of 15270, yet that average is pulled down by the 680M’s other benchmarks. The Radeon 680M’s Vulkan score of 21965 and 3DMark Steel Nomad DX12 score of 378 show it handles modern APIs far better than the GTX 580 could, given the latter’s lack of Vulkan and older DirectX feature level.

The aggregate averages — 15283 for GTX 580, 15270 for Radeon 680M — are statistically indistinguishable, with a deltaPct of 0.1% in favor of the GTX 580 when comparing their average scores directly. But this parity is misleading: the GTX 580 has exactly one benchmark result, while the Radeon 680M has three, including two tests the GTX 580 cannot run. The data indicates the Radeon 680M is the more versatile and future-proof part.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 580 averages 15283, while the AMD Radeon 680M averages 15270. The difference is 0.1%, with the GTX 580 technically ahead by 13 points.

Q: How much faster is the AMD Radeon 680M in OpenCL?

A: The Radeon 680M scores 23468 in Geekbench OpenCL versus the GTX 580’s 15283, making it 34.9% faster in that specific test.

Q: Does the GTX 580 support Vulkan?

A: No. The GTX 580 lists no Vulkan API support in its specifications, while the Radeon 680M explicitly supports Vulkan 1.4.

Q: What is the power consumption difference?

A: The GTX 580 has a 244 W TDP and requires a 550 W suggested PSU with dual power connectors. The Radeon 680M has a 50 W TDP, no power connectors, and is an IGP.

Q: Can the Radeon 680M handle ray tracing?

A: Yes, it includes 12 ray tracing cores. The GTX 580 has no ray tracing cores at all.

Q: Which GPU has more shading units?

A: The Radeon 680M has 768 shading units, compared to 512 on the GTX 580. However, the GTX 580 has more TMUs (64 versus 48) and more ROPs (48 versus 32).

Where Each One Wins

The AMD Radeon 680M wins in raw compute throughput, delivering 3.379 TFLOPS FP32 and 6.758 TFLOPS FP16 (2:1) against the GTX 580’s 1.581 TFLOPS FP32 and no listed FP16 capability. Its pixel rate of 70.40 GPixel/s and texture rate of 105.6 GTexel/s dwarf the GTX 580’s 24.70 GPixel/s and 49.41 GTexel/s. The Radeon 680M also wins on process technology, transistor density, API support, ray tracing, and power efficiency — all while being an active production part released in 2023 versus the end-of-life GTX 580 from 2010.

The NVIDIA GeForce GTX 580 wins in memory configuration specifics: 1536 MB of dedicated GDDR5 on a 384-bit bus with 192.4 GB/s fixed bandwidth, versus the Radeon 680M’s system-shared memory with dependent bandwidth. The GTX 580 also offers a discrete form factor with dual-slot cooling and a 267 mm length, which some desktop builders may prefer for airflow or aesthetic reasons. Its PCIe 2.0 x16 interface, while older, is more universally compatible with legacy motherboards than the Radeon 680M’s PCIe 4.0 x8.

In benchmark breadth, the GTX 580 has no wins in the head-to-head comparison — the Radeon 680M takes the sole victory. But the aggregate score parity means the GTX 580 is not obsolete; it simply peaks in a narrower range of workloads, primarily those that favor its fixed memory bandwidth and higher ROP count.

Specification Differences

| Specification | NVIDIA GeForce GTX 580 | AMD Radeon 680M |

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

| Chip | GF110 | Rembrandt+ |

| Architecture | Fermi 2.0 | RDNA 2.0 |

| Process Node | 40 nm | 6 nm |

| Transistors | 3,000 million | 13,100 million |

| Die Size | 520 mm² | 208 mm² |

| Transistor Density | 5.8M / mm² | 63.0M / mm² |

| Base Clock | None listed | 2000 MHz |

| Boost Clock | None listed | 2200 MHz |

| Memory Size | 1536 MB | System Shared |

| Memory Type | GDDR5 | System Shared |

| Memory Bus | 384 bit | System Shared |

| Memory Bandwidth | 192.4 GB/s | System Dependent |

| Shading Units | 512 | 768 |

| TMUs | 64 | 48 |

| ROPs | 48 | 32 |

| RT Cores | None | 12 |

| Pixel Rate | 24.70 GPixel/s | 70.40 GPixel/s |

| Texture Rate | 49.41 GTexel/s | 105.6 GTexel/s |

| FP32 | 1.581 TFLOPS | 3.379 TFLOPS |

| FP16 | None listed | 6.758 TFLOPS (2:1) |

| TDP | 244 W | 50 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 1x 6-pin + 1x 8-pin | None |

| Suggested PSU | 550 W | None listed |

| Bus Interface | PCIe 2.0 x16 | PCIe 4.0 x8 |

| DirectX | 12 (11_0) | 12 Ultimate (12_2) |

| Vulkan | None listed | 1.4 |

| Dimensions | 267 mm x 111 mm | None listed |

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

| Release Date | 2010-11-08 | 2023-01-02 |

| Launch MSRP | 499 USD | None listed |