NVIDIA GeForce GTX 570 vs NVIDIA GeForce GTX 660 Ti Comparison

Head-to-Head Benchmarks

The database contains a single direct comparison between these two cards, the Geekbench OpenCL test. In that test, the NVIDIA GeForce GTX 660 Ti scored 15,113 points, while the NVIDIA GeForce GTX 570 scored 13,515 points. This gives the GTX 660 Ti a clear victory with an 11.8% performance advantage. The data shows a decisive win for the newer Kepler-based card in this compute-oriented workload.

The GTX 660 Ti’s average benchmark score across all recorded tests is 15,063, which places it in the 57th percentile of all GPUs. The GTX 570, with its single recorded score of 13,515, sits in the 54th percentile. While both cards are in the lower half of the performance distribution, the GTX 660 Ti holds a modest but consistent position above its predecessor-generation rival.

Looking at the nearest rivals for the GTX 660 Ti, the data shows it is very closely matched with the AMD Radeon Pro 560X, which scores 15,082, a negligible 0.1% difference. The AMD Radeon RX 7600 scores 15,171, which is 0.7% higher, and the NVIDIA GeForce RTX 3050 OEM scores 15,199, 0.9% higher. The AMD Radeon 680M integrated graphics scores 15,270, putting it 1.4% ahead. This indicates that the GTX 660 Ti sits in a tight cluster of performance, where the 11.8% lead over the GTX 570 is far more significant than the differences between the GTX 660 Ti and its own nearest rivals.

For the GTX 570, its nearest rivals paint a similarly tight picture. The AMD Radeon HD 7770M scores 13,536, a 0.2% difference, and the AMD Radeon RX 9070 XT scores 13,543, also a 0.2% difference. The NVIDIA P106-090 scores 13,470, which is 0.3% lower, and the AMD Radeon Pro 555 scores 13,407, which is 0.8% lower. The GTX 570 is therefore surrounded by a mix of mobile and desktop parts, but the GTX 660 Ti stands out as a full 11.8% ahead of it in the head-to-head comparison.

Architecture Differences

The two cards come from different architectural generations. The GTX 660 Ti uses the GK104 chip, built on the Kepler architecture, which is part of the GeForce 600 generation. The GTX 570 uses the GF110 chip, built on the Fermi 2.0 architecture, part of the GeForce 500 generation. This generational gap explains most of the performance difference.

The manufacturing process differs substantially. The GTX 660 Ti is fabricated on a 28 nm process at TSMC, while the GTX 570 uses a 40 nm process, also at TSMC. The newer process allows the GTX 660 Ti to pack 3,540 million transistors into a die size of 294 mm². This yields a transistor density of 12.0 million transistors per mm². The GTX 570, in contrast, has 3,000 million transistors on a much larger die of 520 mm², giving it a density of only 5.8 million transistors per mm². The data shows the Kepler chip is significantly more efficient in terms of transistor packing.

The compute resources also differ drastically. The GTX 660 Ti has 1,344 shading units, 112 texture mapping units, and 24 render output units. The GTX 570 has 480 shading units, 60 texture mapping units, and 40 render output units. The GTX 660 Ti has nearly three times the shading units and almost double the texture units, but fewer ROPs. These numbers explain the fill rate figures: the GTX 660 Ti achieves a pixel rate of 27.44 GPixel/s and a texture rate of 109.8 GTexel/s, while the GTX 570 achieves 21.96 GPixel/s and 43.92 GTexel/s respectively.

The floating-point performance follows the shading unit count. The GTX 660 Ti delivers 2.634 TFLOPS of FP32 compute, while the GTX 570 delivers 1,405.4 GFLOPS. That is a significant gap in raw compute throughput, consistent with the OpenCL benchmark results.

Clock speeds also differ. The GTX 660 Ti has a base clock of 915 MHz and a boost clock of 980 MHz. The GTX 570 has no recorded base or boost clocks in the database, but its memory runs at 950 MHz, translating to 3.8 Gbps effective. The GTX 660 Ti’s memory runs at 1,502 MHz, translating to 6 Gbps effective, which is considerably faster.

Memory configuration is another key architectural difference. The GTX 660 Ti has 2 GB of GDDR5 memory on a 192-bit bus, yielding a bandwidth of 144.2 GB/s. The GTX 570 has 1,280 MB of GDDR5 memory on a wider 320-bit bus, yielding a slightly higher bandwidth of 152.0 GB/s. Despite the narrower bus, the higher clock speed on the GTX 660 Ti’s memory nearly closes the gap, and the larger capacity gives it an advantage in memory-heavy workloads.

The bus interface also differs: the GTX 660 Ti uses PCIe 3.0 x16, while the GTX 570 uses PCIe 2.0 x16. The newer interface offers higher bandwidth for data transfer between the card and the system.

API support is similar for DirectX and OpenGL, with both cards supporting DirectX 12 (11_0) and OpenGL 4.6. However, the GTX 660 Ti supports Vulkan 1.2.175, while the GTX 570 has no recorded Vulkan support in the database. This is a notable feature difference for modern applications.

FAQ

Q: Which card is faster in the recorded benchmarks?

A: The NVIDIA GeForce GTX 660 Ti is faster, scoring 15,113 in Geekbench OpenCL compared to 13,515 for the GTX 570, an 11.8% lead.

Q: How much memory does each card have?

A: The GTX 660 Ti has 2 GB of GDDR5 memory, while the GTX 570 has 1,280 MB of GDDR5 memory.

Q: What are the power requirements for each card?

A: The GTX 660 Ti has a TDP of 150 W and a suggested PSU of 450 W, while the GTX 570 has a TDP of 219 W and a suggested PSU of 550 W.

Q: Do both cards support Vulkan?

A: No. The GTX 660 Ti supports Vulkan 1.2.175, while the GTX 570 has no recorded Vulkan support in the database.

Q: What is the transistor density of each chip?

A: The GTX 660 Ti’s GK104 chip has a density of 12.0 million transistors per mm², while the GTX 570’s GF110 chip has a density of 5.8 million transistors per mm².

Q: Which card has a higher pixel fill rate?

A: The GTX 660 Ti has a pixel rate of 27.44 GPixel/s, which is higher than the GTX 570’s 21.96 GPixel/s.

Specification Differences

| Specification | NVIDIA GeForce GTX 660 Ti | NVIDIA GeForce GTX 570 |

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

| Architecture | Kepler | Fermi 2.0 |

| Generation | GeForce 600 | GeForce 500 |

| Process Node | 28 nm | 40 nm |

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

| Die Size | 294 mm² | 520 mm² |

| Transistor Density | 12.0M / mm² | 5.8M / mm² |

| Base Clock | 915 MHz | Not recorded |

| Boost Clock | 980 MHz | Not recorded |

| Memory Clock | 1502 MHz, 6 Gbps effective | 950 MHz, 3.8 Gbps effective |

| Memory Size | 2 GB | 1280 MB |

| Memory Bus Width | 192 bit | 320 bit |

| Memory Bandwidth | 144.2 GB/s | 152.0 GB/s |

| Shading Units | 1344 | 480 |

| TMUs | 112 | 60 |

| ROPs | 24 | 40 |

| Pixel Rate | 27.44 GPixel/s | 21.96 GPixel/s |

| Texture Rate | 109.8 GTexel/s | 43.92 GTexel/s |

| FP32 | 2.634 TFLOPS | 1,405.4 GFLOPS |

| TDP | 150 W | 219 W |

| Suggested PSU | 450 W | 550 W |

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

| Display Outputs | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 | 2x DVI, 1x mini-HDMI 1.3a |

| Vulkan Support | 1.2.175 | Not recorded |

| Length | 241 mm (9.5 inches) | 267 mm (10.5 inches) |

| Release Date | 2012-08-15 | 2010-12-06 |

Where Each One Wins

The GTX 660 Ti wins in the only recorded head-to-head benchmark, the Geekbench OpenCL test, with an 11.8% advantage. It also has a higher average benchmark score (15,063 vs 13,515) and a higher percentile ranking (57th vs 54th). Its strengths lie in raw compute throughput: it has more shading units, higher clock speeds, and a more efficient manufacturing process. The larger 2 GB memory capacity is also an advantage for modern workloads that require more video memory. The GTX 660 Ti supports Vulkan, which the GTX 570 does not, making it more compatible with newer applications and games that rely on this API.

The GTX 570 does hold a few advantages in specific specifications. It has a wider memory bus (320 bit vs 192 bit), which gives it a higher memory bandwidth (152.0 GB/s vs 144.2 GB/s) despite slower memory clocks. It also has more ROPs (40 vs 24), which can help with certain rasterization tasks, though the GTX 660 Ti still achieves a higher pixel rate due to its higher clocks. The GTX 570 has a lower power draw per ROP, but its overall TDP is higher at 219 W versus 150 W.

In terms of physical footprint, the GTX 660 Ti is shorter at 241 mm compared to the GTX 570’s 267 mm, which may matter for smaller cases. However, both are dual-slot cards and use the same 2x 6-pin power connectors.

The Verdict

The data points to a clear winner for most use cases: the NVIDIA GeForce GTX 660 Ti. It outperforms the GTX 570 in the only direct benchmark comparison, with an 11.8% higher OpenCL score. It also has a higher average score and a higher percentile ranking. For compute-heavy tasks, the GTX 660 Ti’s 2.634 TFLOPS of FP32 performance versus 1,405.4 GFLOPS is a substantial advantage. Its 2 GB of memory is more future-proof than the GTX 570’s 1,280 MB, and its support for Vulkan makes it the more versatile option for modern software.

The GTX 570, however, is not without merit. Its wider memory bus and higher bandwidth could be beneficial in specific bandwidth-bound scenarios, and its higher ROP count may help in certain rendering workloads. But the recorded data shows that these theoretical advantages do not translate into a benchmark win.

For a user choosing between these two end-of-life cards, the GTX 660 Ti is the stronger choice for general performance, compute tasks, and modern API support. The GTX 570 remains viable only in niche cases where its specific memory characteristics are more important than the overall compute advantage of the GTX 660 Ti. The GTX 660 Ti also has a lower TDP of 150 W versus 219 W, meaning it achieves its higher performance while drawing less power. The verdict from the database is straightforward: the GTX 660 Ti is the superior card.