NVIDIA GeForce GTX 1660 SUPER vs NVIDIA GeForce GTX 580 Comparison

Head-to-Head Benchmarks

The only direct benchmark comparison recorded in the database is the Geekbench OpenCL test. The GTX 1660 SUPER scores 52,490, while the GTX 580 scores 15,283. This translates to a 70.9% deficit for the older card, meaning the GTX 1660 SUPER delivers over three times the raw compute performance in this specific workload. The margin is decisive and reflects the generational gap between the two architectures.

The GTX 580's single benchmark result places it at the 58th percentile of all GPUs in the database. Its nearest rivals include the RTX 2060 at 15,290 (0% delta), the AMD Radeon 680M at 15,270 (0.1% delta), the RTX 3050 OEM at 15,199 (0.6% delta), and the RX 7600 at 15,171 (0.7% delta). These are remarkably tight margins, with the GTX 580 effectively matching modern entry-level and midrange parts in OpenCL. The data suggests that despite its age, the GTX 580 retains a surprising level of compute capability in this legacy test.

The GTX 1660 SUPER, by contrast, sits at the 53rd percentile overall, with an average benchmark score of 12,986 across all its recorded tests. Its nearest rivals in the aggregate include the Tesla M2090 at 13,075 (-0.7% delta), the RTX 3050 Ti Mobile at 12,940 (0.4% delta), the RX 580 at 12,928 (0.4% delta), and the Radeon 740M at 12,870 (0.9% delta). The GTX 1660 SUPER's average score is lower than its OpenCL result because it also runs DirectX 9, 10, 11, and 12 tests, Vulkan, and compute workloads, many of which score lower. For example, its Passmark DirectX 9 score is 189, DirectX 11 is 104, and DirectX 12 is 50. The OpenCL score, however, is the headline number for this head-to-head.

The GTX 580 has only the OpenCL score recorded, so its average benchmark score equals that single result at 15,283. The GTX 1660 SUPER wins the head-to-head by the sole available metric, with a 1-0 win count. No other shared test exists in the database, so this comparison rests entirely on the OpenCL result. The magnitude of the win, though, is substantial enough to define the overall performance relationship.

Architecture Differences

The GTX 580 uses the GF110 chip on a 40 nm process from TSMC, with 3,000 million transistors packed into a 520 mm² die. Its transistor density is 5.8 million per mm². The GTX 1660 SUPER uses the TU116 chip on a 12 nm process, also from TSMC, with 6,600 million transistors on a 284 mm² die. That yields a density of 23.2 million per mm², which is four times higher. The newer card fits more than twice the transistors into roughly half the silicon area.

The GTX 580 belongs to the Fermi 2.0 architecture and the GeForce 500 generation. The GTX 1660 SUPER belongs to the Turing architecture and the GeForce 16 generation. DirectX support differs: the GTX 580 lists DirectX 12 (11_0), while the GTX 1660 SUPER lists DirectX 12 (12_1). Both support OpenGL 4.6. The GTX 1660 SUPER adds Vulkan 1.4 support, while the GTX 580 has no Vulkan entry in the database.

Shader and texture resources differ sharply. The GTX 580 has 512 shading units, 64 texture mapping units, and 48 ROPs. The GTX 1660 SUPER has 1,408 shading units, 88 texture mapping units, and 48 ROPs. The ROP count is identical, but the newer card has 2.75 times the shaders and 1.375 times the TMUs. Pixel rate on the GTX 580 is 24.70 GPixel/s, versus 85.68 GPixel/s on the GTX 1660 SUPER. Texture rate is 49.41 GTexel/s versus 157.1 GTexel/s. FP32 throughput is 1.581 TFLOPS versus 5.027 TFLOPS. The GTX 1660 SUPER also records FP16 performance at 10.05 TFLOPS (2:1 ratio), while the GTX 580 has no FP16 entry.

Memory configuration differs as well. The GTX 580 uses 1,536 MB of GDDR5 on a 384-bit bus, yielding 192.4 GB/s bandwidth. The GTX 1660 SUPER uses 6 GB of GDDR6 on a 192-bit bus, yielding 336.0 GB/s bandwidth. The newer card has four times the capacity and 74.6% more bandwidth despite a narrower bus. Memory clocks are listed as 4 Gbps effective for the GTX 580 and 14 Gbps effective for the GTX 1660 SUPER.

Where Each One Wins

The GTX 580 wins no recorded benchmark in this comparison. Its single OpenCL score is far below the GTX 1660 SUPER, and no other test exists where it takes the lead. The data indicates that the older card cannot claim a single performance victory in the available measurements.

The GTX 1660 SUPER wins the only head-to-head test. Beyond that, its broader benchmark suite shows strengths in specific areas. Its Passmark G3D score is 12,699, which is high relative to its other Passmark results. Its Geekbench Vulkan score is 57,102, slightly higher than its OpenCL score of 52,490. Its Passmark GPU compute score is 5,076. The DirectX 9 score of 189 is the highest among its legacy DirectX tests, while DirectX 12 scores only 50 and DirectX 10 scores 65. This pattern suggests the card is strongest in modern compute and Vulkan workloads, weaker in legacy DirectX rasterization tests.

The GTX 580's nearest rivals in OpenCL are all within 0.7% of its score, meaning it performs on par with modern integrated graphics like the Radeon 680M and entry-level discrete cards like the RTX 3050 OEM. That places its compute capability in a specific niche: acceptable for basic OpenCL tasks but far behind any modern midrange card. The GTX 1660 SUPER, by contrast, sits near the RTX 3050 Ti Mobile and RX 580 in average score, indicating it competes with established midrange parts from the same era.

The use-case split is clear from the data. The GTX 580 is competitive only in the narrow OpenCL test it shares with the newer card, and even there it loses by a wide margin. The GTX 1660 SUPER has no such limitation; it records competitive scores across DirectX 9 through 12, Vulkan, OpenCL, and compute tests. The only domain where the GTX 580 might claim relevance is legacy compatibility, given its Fermi architecture and 2010-era feature set, but no benchmark result supports a performance advantage.

FAQ

Q: Which card has the higher FP32 compute throughput?

A: The GTX 1660 SUPER delivers 5.027 TFLOPS, while the GTX 580 delivers 1.581 TFLOPS. The newer card is roughly 3.2 times faster in FP32.

Q: How do the memory capacities compare?

A: The GTX 580 has 1,536 MB of GDDR5, while the GTX 1660 SUPER has 6 GB of GDDR6. The newer card offers four times the capacity and higher bandwidth at 336.0 GB/s versus 192.4 GB/s.

Q: What is the transistor count difference?

A: The GTX 580 has 3,000 million transistors on a 520 mm² die, while the GTX 1660 SUPER has 6,600 million transistors on a 284 mm² die. The newer card uses a denser 12 nm process versus 40 nm.

Q: Does the GTX 580 support Vulkan?

A: No Vulkan support is listed for the GTX 580. The GTX 1660 SUPER lists Vulkan 1.4 support.

Q: Which card wins the recorded head-to-head benchmark?

A: The GTX 1660 SUPER wins the Geekbench OpenCL test with a score of 52,490 versus 15,283, a 70.9% margin.

Q: How does the GTX 580 compare to its nearest rivals in OpenCL?

A: The GTX 580 scores 15,283, within 0.7% of the RTX 2060 (15,290), Radeon 680M (15,270), RTX 3050 OEM (15,199), and RX 7600 (15,171). It effectively matches those cards in that specific test.

The Verdict

The data points to a lopsided comparison. The GTX 1660 SUPER wins the only shared benchmark by 70.9%, and its broader suite shows competitive performance across modern APIs. The GTX 580, despite being a flagship from its generation, scores no wins and sits at the 58th percentile with a single OpenCL result that places it near integrated graphics. The GTX 580's 1.581 TFLOPS FP32, 192.4 GB/s bandwidth, and 1,536 MB memory are all far below the GTX 1660 SUPER's 5.027 TFLOPS, 336.0 GB/s, and 6 GB.

The GTX 1660 SUPER also offers a much more efficient package. Its TDP is 125 W with a suggested PSU of 300 W, versus 244 W and 550 W for the GTX 580. Both are dual-slot cards, but the newer card is shorter at 229 mm versus 267 mm. The GTX 580 uses 1x 6-pin plus 1x 8-pin power connectors, while the GTX 1660 SUPER uses a single 8-pin. The newer card also supports PCIe 3.0 x16 versus PCIe 2.0 x16, and its display outputs include HDMI 2.0 and DisplayPort 1.4a, while the GTX 580 offers DVI and mini-HDMI 1.3a.

For a user choosing between these two, the GTX 1660 SUPER is the clear performer. It wins every measurable category except the ROP count, which ties at 48. The GTX 580's only advantage is its 384-bit memory bus, which does not compensate for the lower bandwidth. The verdict from the recorded data is unambiguous: the GTX 1660 SUPER is the superior card in compute, memory, efficiency, and API support.

Specification Differences

| Specification | NVIDIA GeForce GTX 580 | NVIDIA GeForce GTX 1660 SUPER |

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

| Chip | GF110 | TU116 |

| Architecture | Fermi 2.0 | Turing |

| Generation | GeForce 500 | GeForce 16 |

| Process Node | 40 nm | 12 nm |

| Transistors | 3,000 million | 6,600 million |

| Die Size | 520 mm² | 284 mm² |

| Transistor Density | 5.8M / mm² | 23.2M / mm² |

| Base Clock | Not listed | 1530 MHz |

| Boost Clock | Not listed | 1785 MHz |

| Memory Clock | 4 Gbps effective | 14 Gbps effective |

| Memory Size | 1536 MB | 6 GB |

| Memory Type | GDDR5 | GDDR6 |

| Memory Bus Width | 384 bit | 192 bit |

| Memory Bandwidth | 192.4 GB/s | 336.0 GB/s |

| Shading Units | 512 | 1408 |

| TMUs | 64 | 88 |

| ROPs | 48 | 48 |

| Pixel Rate | 24.70 GPixel/s | 85.68 GPixel/s |

| Texture Rate | 49.41 GTexel/s | 157.1 GTexel/s |

| FP32 | 1.581 TFLOPS | 5.027 TFLOPS |

| FP16 | Not listed | 10.05 TFLOPS (2:1) |

| TDP | 244 W | 125 W |

| Power Connectors | 1x 6-pin + 1x 8-pin | 1x 8-pin |

| Suggested PSU | 550 W | 300 W |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Display Outputs | 2x DVI, 1x mini-HDMI 1.3a | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a |

| DirectX | 12 (11_0) | 12 (12_1) |

| Vulkan | Not listed | 1.4 |

| Length | 267 mm | 229 mm |

| Height | 111 mm | 111 mm |

| Width | Not listed | 35 mm |

| Release Date | 2010-11-08 | 2019-10-28 |

| Launch MSRP | 499 USD | 229 USD |