AMD Radeon Pro 560X vs NVIDIA GeForce GTX 680 Comparison

# AMD Radeon Pro 560X vs NVIDIA GeForce GTX 680

The AMD Radeon Pro 560X and NVIDIA GeForce GTX 680 represent two very different eras of GPU design, and their benchmark results reflect that divergence clearly. The Radeon Pro 560X, a 14 nm Polaris 21 part from 2018, and the GTX 680, a 28 nm GK104 Kepler chip from 2012, each claim victories in different API tests. The data shows a split decision: the AMD card dominates in Metal, while the NVIDIA card counters in OpenCL and Vulkan. This is not a simple "newer is better" story, and the underlying numbers reveal why.

Where Each One Wins

The Radeon Pro 560X claims its single decisive victory in the geekbench_metal test, posting a score of 18,763 against the GTX 680's 7,897. That is a 137.6% advantage, a margin so large it suggests the GTX 680 is nearly outclassed in Apple's graphics API. The GTX 680, in contrast, wins the geekbench_opencl test with 16,906 versus 9,663, a 42.8% lead for NVIDIA. The Vulkan test is far closer: GTX 680 scores 17,646 against 16,819 for the Radeon, a 4.7% edge. These results indicate the Radeon Pro 560X is the stronger choice for Metal-centric workloads, which dominate macOS environments, while the GTX 680 remains competitive in OpenCL and edges ahead in Vulkan. The overall win count favors the GTX 680 at 2 wins to 1, but the magnitude of the Metal loss is striking.

Architecture Differences

The two GPUs are built on fundamentally different foundations. The Radeon Pro 560X uses the Polaris 21 chip with GCN 4.0 architecture, fabricated on a 14 nm process by GlobalFoundries. It packs 3,000 million transistors onto a 123 mm² die, yielding a transistor density of 24.4M per mm². The GTX 680 uses the GK104 chip with Kepler architecture, built on TSMC's 28 nm process, with 3,540 million transistors on a much larger 294 mm² die, giving a density of just 12.0M per mm². The smaller, denser Polaris 21 die is a direct consequence of the newer process node. The Radeon supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3, while the GTX 680 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Radeon also features FP16 at 2.056 TFLOPS (1:1 ratio), while the GTX 680 lists no FP16 capability. The GTX 680 has a 256-bit memory bus and 192.3 GB/s bandwidth, compared to the Radeon's 128-bit bus and 94.08 GB/s. Both are end-of-life products, but the GTX 680's predecessor and successor are listed as GeForce 500 and GeForce 700, respectively, while the Radeon Pro 560X has no listed predecessor or successor.

FAQ

Q: Which card wins in Metal performance?

A: The AMD Radeon Pro 560X wins decisively, scoring 18,763 in geekbench_metal versus the GTX 680's 7,897. That is a 137.6% advantage for the Radeon.

Q: Is the GTX 680 better in OpenCL?

A: Yes. The GTX 680 scores 16,906 in geekbench_opencl, while the Radeon Pro 560X scores 9,663. NVIDIA leads by 42.8% in this test.

Q: How close is Vulkan performance between these two?

A: Very close. The GTX 680 scores 17,646 and the Radeon Pro 560X scores 16,819, a difference of just 4.7% in NVIDIA's favor.

Q: What is the memory size and type for each card?

A: The Radeon Pro 560X has 4 GB of GDDR5 memory on a 128-bit bus, while the GTX 680 has 2 GB of GDDR5 memory on a 256-bit bus.

Q: Which card has a higher transistor count?

A: The GTX 680 has 3,540 million transistors, compared to the Radeon Pro 560X's 3,000 million. However, the Radeon's die is far smaller at 123 mm² versus 294 mm².

Q: What are the power requirements?

A: The Radeon Pro 560X has a 75 W TDP and requires no power connectors, while the GTX 680 has a 195 W TDP and needs two 6-pin connectors.

Specification Differences

| Specification | AMD Radeon Pro 560X | NVIDIA GeForce GTX 680 |

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

| Process Node | 14 nm | 28 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 3,000 million | 3,540 million |

| Die Size | 123 mm² | 294 mm² |

| Transistor Density | 24.4M / mm² | 12.0M / mm² |

| Base Clock | Not listed | 1006 MHz |

| Boost Clock | Not listed | 1058 MHz |

| Memory Clock | 1470 MHz (5.9 Gbps effective) | 1502 MHz (6 Gbps effective) |

| Memory Size | 4 GB | 2 GB |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 94.08 GB/s | 192.3 GB/s |

| Shading Units | 1024 | 1536 |

| TMUs | 64 | 128 |

| ROPs | 16 | 32 |

| Pixel Rate | 16.06 GPixel/s | 33.86 GPixel/s |

| Texture Rate | 64.26 GTexel/s | 135.4 GTexel/s |

| FP32 | 2.056 TFLOPS | 3.250 TFLOPS |

| FP16 | 2.056 TFLOPS (1:1) | Not listed |

| TDP | 75 W | 195 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 2x 6-pin |

| Suggested PSU | Not listed | 450 W |

| Bus Interface | PCIe 3.0 x8 | PCIe 3.0 x16 |

| Display Outputs | Portable Device Dependent | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| DirectX | 12 (12_0) | 12 (11_0) |

| Vulkan | 1.3 | 1.2.175 |

| Dimensions | Not listed | 256 mm (10.1 inches) long, 111 mm (4.4 inches) high, 38 mm (1.5 inches) wide |

| Release Date | 2018-07-15 | 2012-03-21 |

| Launch MSRP | Not listed | 499 USD |

Head-to-Head Benchmarks

The geekbench_metal test delivers the largest swing in this comparison. The Radeon Pro 560X posts 18,763 against the GTX 680's 7,897, a 137.6% delta. This is not a marginal improvement; it is a near-doubling of the NVIDIA card's score. The result strongly suggests that the Radeon's GCN 4.0 architecture is far better optimized for Metal than Kepler, which launched nearly six years earlier. For any Metal-based application, the Radeon Pro 560X is clearly the superior option.

The geekbench_opencl test flips the script. The GTX 680 achieves 16,906, while the Radeon Pro 560X manages only 9,663. The 42.8% deficit for AMD is substantial and somewhat surprising given the Radeon's newer architecture. The GTX 680's higher shading unit count (1536 versus 1024), TMU count (128 versus 64), and ROP count (32 versus 16) likely contribute to this result, as does its much wider 256-bit memory bus. In raw compute throughput, the GTX 680's FP32 of 3.250 TFLOPS also outpaces the Radeon's 2.056 TFLOPS.

The geekbench_vulkan test is the closest contest. The GTX 680 wins with 17,646 against 16,819, a 4.7% margin. This small delta indicates that both cards are reasonably competitive in Vulkan, despite the architectural differences. The Radeon's newer Vulkan 1.3 support versus the GTX 680's 1.2.175 does not translate into a win here. The GTX 680's higher memory bandwidth (192.3 GB/s vs 94.08 GB/s) may be the deciding factor in this test.

The Verdict

The data points to a clear split based on workload. The AMD Radeon Pro 560X is the card to choose for Metal-centric applications, where its 137.6% lead over the GTX 680 is overwhelming. It also offers double the memory (4 GB vs 2 GB), a far lower TDP (75 W vs 195 W), and a much smaller physical footprint as an IGP with no power connectors. The GTX 680, however, takes the crown in OpenCL with a 42.8% advantage and narrowly wins in Vulkan with a 4.7% edge. It also has higher raw specs in almost every category: more shading units, TMUs, ROPs, higher clock speeds, and more than double the memory bandwidth. The GTX 680's 3.250 TFLOPS FP32 performance is also 58% higher than the Radeon's 2.056 TFLOPS. The GTX 680's launch MSRP was 499 USD. Users on modern Mac platforms will almost certainly prefer the Radeon Pro 560X, while those prioritizing OpenCL compute or Vulkan gaming performance on older systems might find the GTX 680 more capable. The average benchmark scores reflect this tension: the Radeon Pro 560X averages 15,082 across all tests, while the GTX 680 averages 14,150. The Radeon's percentile ranking is 57 versus the GTX 680's 55, a small but consistent edge. Ultimately, the choice hinges on which API matters most, and the data is unambiguous about that trade-off.