NVIDIA GeForce GTX 580 vs NVIDIA GeForce RTX 2060 SUPER Comparison

Head-to-Head Benchmarks

The recorded data contains exactly one direct head-to-head comparison between the NVIDIA GeForce RTX 2060 SUPER and the NVIDIA GeForce GTX 580: the OpenCL workload within Geekbench. In that test, the RTX 2060 SUPER scored 76,957 points against the GTX 580's 15,283 points. That is a delta of 403.5%, meaning the newer card delivered roughly five times the raw compute throughput in this particular API workload. The win is decisive, but it is also the only shared benchmark in the database, so conclusions must be drawn carefully from a single data point.

The RTX 2060 SUPER also holds a substantial edge in the broader benchmark averages recorded for each card. Its average benchmark score sits at 18,093, while the GTX 580 averages 15,283. The difference is about 18.4%, a notable gap but far smaller than the OpenCL disparity suggests. This discrepancy indicates that the GTX 580 remains competitive in certain legacy or lower-level workloads, even as it falls far behind in compute-heavy OpenCL scenarios. The GTX 580's average score is almost exactly matched by the baseline NVIDIA GeForce RTX 2060, which posts an average of 15,290, a delta of 0%. In other words, the GTX 580 performs on par with a non-SUPER RTX 2060 in the database's aggregate measurements, despite being several generations older.

The RTX 2060 SUPER's percentile standing among all GPUs is 62, while the GTX 580 sits at 58. Both cards are mid-pack in the overall distribution, but the SUPER card is positioned above a larger share of the field. The RTX 2060 SUPER's nearest rivals in the database are the NVIDIA GeForce RTX 3060 Mobile (average 18,159, delta -0.4%), the AMD Radeon Pro 5700 (average 18,189, delta -0.5%), the AMD Radeon RX 460 (average 18,373, delta -1.5%), and the Intel Arc A770M (average 18,383, delta -1.6%). These deltas are all negative, indicating the RTX 2060 SUPER trails each of these cards by less than 2%, a very tight cluster. Meanwhile, the GTX 580's nearest rivals include the RTX 2060 (delta 0%), the AMD Radeon 680M (delta 0.1%), the NVIDIA GeForce RTX 3050 OEM (delta 0.6%), and the AMD Radeon RX 7600 (delta 0.7%). The GTX 580 is effectively tied with the RTX 2060 and slightly ahead of the others, but all of these deltas are within 1%, showing that the old Fermi card still trades blows with much newer hardware in the database's aggregate scoring.

The OpenCL result is the only test where both cards appear in the same workload, and it is overwhelmingly one-sided. Still, the aggregate averages tell a more nuanced story: the RTX 2060 SUPER is clearly faster overall, but the GTX 580 is not an outlier when viewed against its nearest rivals. It sits comfortably in a cluster of modern and mid-range parts, which is remarkable for a card from the GeForce 500 generation.

Where Each One Wins

The RTX 2060 SUPER wins the only direct head-to-head benchmark, and it wins by a massive margin in OpenCL compute. If the workload is compute-heavy, particularly in an OpenCL context, the data strongly favors the newer card. The 403.5% delta is not a marginal improvement; it is a generational leap in raw throughput. The RTX 2060 SUPER also has a higher average benchmark score (18,093 versus 15,283) and a higher percentile rank (62 versus 58), so in aggregate performance across the database's full test suite, it is the stronger card.

The GTX 580 does not win any recorded head-to-head test, but its aggregate score places it in a competitive position relative to modern mid-range parts. Its average of 15,283 is essentially identical to the RTX 2060 (15,290), and it edges out the AMD Radeon 680M (15,270), the RTX 3050 OEM (15,199), and the AMD Radeon RX 7600 (15,171) by fractions of a percent. For workloads that resemble the aggregate mix of the database's benchmark suite, the GTX 580 can still hold its own against a wide range of newer GPUs. It is not a card to dismiss outright, even though it lacks the compute muscle of the RTX 2060 SUPER.

Use-case splitting is straightforward from the data. The RTX 2060 SUPER is the choice for OpenCL compute tasks and for any scenario where the highest possible average benchmark score matters. The GTX 580 is the choice for legacy compatibility or for situations where a card from the GeForce 500 generation is required, since its aggregate performance is within 1% of several modern cards. Neither card has a clear advantage in the other's territory, because the only shared test is OpenCL.

Architecture Differences

The two cards come from completely different architectural eras. The RTX 2060 SUPER uses the TU106 chip, built on Turing architecture, fabricated by TSMC on a 12 nm process. The GTX 580 uses the GF110 chip, built on Fermi 2.0 architecture, also fabricated by TSMC but on a 40 nm process. The transistor counts reflect the generational gap: the RTX 2060 SUPER packs 10,800 million transistors on a 445 mm² die, giving a transistor density of 24.3 million per square millimeter. The GTX 580 has 3,000 million transistors on a 520 mm² die, a density of 5.8 million per square millimeter. The newer card is smaller in physical area but contains more than three times as many transistors.

The memory subsystems differ substantially. The RTX 2060 SUPER has 8 GB of GDDR6 memory on a 256-bit bus, delivering 448.0 GB/s of bandwidth. The GTX 580 has 1536 MB of GDDR5 memory on a 384-bit bus, delivering 192.4 GB/s. The newer card has more than twice the bandwidth despite a narrower bus, thanks to the faster GDDR6 memory running at 1750 MHz (14 Gbps effective) versus the GTX 580's 1002 MHz (4 Gbps effective). The GTX 580's wider bus is not enough to overcome the slower memory technology.

Compute resources also diverge sharply. The RTX 2060 SUPER has 2,176 shading units, 136 texture mapping units, and 64 raster output units. It also includes 34 ray tracing cores and 272 tensor cores, features that simply do not exist on the GTX 580. The GTX 580 has 512 shading units, 64 TMUs, and 48 ROPs. Pixel and texture rates tell the story: the RTX 2060 SUPER reaches 105.6 GPixel/s and 224.4 GTexel/s, while the GTX 580 manages 24.70 GPixel/s and 49.41 GTexel/s. FP32 throughput is 7.181 TFLOPS for the RTX 2060 SUPER versus 1.581 TFLOPS for the GTX 580. The RTX 2060 SUPER also supports FP16 at 14.36 TFLOPS (2:1 ratio), while the GTX 580 has no recorded FP16 capability.

API support differs as well. The RTX 2060 SUPER supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GTX 580 supports DirectX 12 (11_0) and OpenGL 4.6, but has no recorded Vulkan support. The RTX 2060 SUPER also features a PCIe 3.0 x16 bus interface and a USB Type-C display output, neither of which appear on the GTX 580, which uses PCIe 2.0 x16 and outputs only 2x DVI and 1x mini-HDMI 1.3a.

Power requirements and physical dimensions also differ. The RTX 2060 SUPER has a TDP of 175 W with a single 8-pin power connector and a suggested PSU of 450 W. The GTX 580 has a TDP of 244 W, requiring a 6-pin plus 8-pin configuration and a suggested PSU of 550 W. The RTX 2060 SUPER is shorter at 229 mm (9 inches) and taller at 113 mm (4.4 inches) with a width of 35 mm (1.4 inches). The GTX 580 is longer at 267 mm (10.5 inches) and has the same 111 mm (4.4 inches) height, with no recorded width. Both are dual-slot cards.

FAQ

Q: How much faster is the RTX 2060 SUPER than the GTX 580 in the OpenCL benchmark?

A: The RTX 2060 SUPER scored 76,957 in Geekbench OpenCL, while the GTX 580 scored 15,283, a delta of 403.5% in favor of the RTX 2060 SUPER.

Q: Does the GTX 580 beat any modern cards in the database?

A: In aggregate average score, the GTX 580 (15,283) is slightly ahead of the AMD Radeon 680M (15,270, delta 0.1%), the NVIDIA GeForce RTX 3050 OEM (15,199, delta 0.6%), and the AMD Radeon RX 7600 (15,171, delta 0.7%). It is tied with the NVIDIA GeForce RTX 2060 (15,290, delta 0%).

Q: What is the average benchmark score difference between the two cards?

A: The RTX 2060 SUPER has an average benchmark score of 18,093, while the GTX 580 has an average of 15,283, a difference of about 18.4% in favor of the RTX 2060 SUPER.

Q: Does the GTX 580 support ray tracing or tensor cores?

A: No. The GTX 580 has no ray tracing cores and no tensor cores. The RTX 2060 SUPER has 34 ray tracing cores and 272 tensor cores.

Q: Which card has higher memory bandwidth?

A: The RTX 2060 SUPER has 448.0 GB/s of bandwidth from 8 GB of GDDR6 on a 256-bit bus. The GTX 580 has 192.4 GB/s from 1536 MB of GDDR5 on a 384-bit bus.

Q: How do the percentile ranks compare?

A: The RTX 2060 SUPER ranks at the 62nd percentile among all GPUs, while the GTX 580 ranks at the 58th percentile.

Specification Differences

| Specification | RTX 2060 SUPER | GTX 580 |

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

| Architecture | Turing | Fermi 2.0 |

| Chip | TU106 | GF110 |

| Process node | 12 nm | 40 nm |

| Transistors | 10,800 million | 3,000 million |

| Die size | 445 mm² | 520 mm² |

| Transistor density | 24.3M / mm² | 5.8M / mm² |

| Memory size | 8 GB | 1536 MB |

| Memory type | GDDR6 | GDDR5 |

| Memory bus width | 256 bit | 384 bit |

| Memory bandwidth | 448.0 GB/s | 192.4 GB/s |

| Memory clock | 1750 MHz, 14 Gbps effective | 1002 MHz, 4 Gbps effective |

| Shading units | 2176 | 512 |

| TMUs | 136 | 64 |

| ROPs | 64 | 48 |

| RT cores | 34 | null |

| Tensor cores | 272 | null |

| Pixel rate | 105.6 GPixel/s | 24.70 GPixel/s |

| Texture rate | 224.4 GTexel/s | 49.41 GTexel/s |

| FP32 | 7.181 TFLOPS | 1.581 TFLOPS |

| FP16 | 14.36 TFLOPS (2:1) | null |

| TDP | 175 W | 244 W |

| Power connectors |