NVIDIA GeForce GTX 465 vs NVIDIA GeForce GTX 660 Comparison

The NVIDIA GeForce GTX 465 and the NVIDIA GeForce GTX 660 represent two distinct eras of GPU design, with the former being a first-generation Fermi part and the latter a mature Kepler design. While both cards are now end-of-life products, the benchmark data indicates a clear generational leap in performance, with the GTX 660 holding a significant advantage in the available OpenCL test. This analysis will break down the architectural shifts, benchmark results, and practical use cases for each card based strictly on the provided data.

FAQ

Q: Which card is faster in the available benchmark data?

A: The NVIDIA GeForce GTX 660 is faster. In the Geekbench OpenCL test, the GTX 660 scored 11,347 points, while the GTX 465 scored 9,294 points. This represents an 18.1% lead for the GTX 660, making it the clear winner in compute performance.

Q: How does the GTX 465 compare to its nearest rivals?

A: The GTX 465's average benchmark score of 9,294 places it in the 46th percentile of all GPUs. Its closest rival is the NVIDIA GeForce GTX 850M, which scores 9,302, a negligible 0.1% difference. It also trails the AMD Radeon R7 M380 by 0.2% but leads the NVIDIA GeForce GTX 960 by 0.2% and the AMD Radeon Vega 8 by 0.8%.

Q: What is the performance context for the GTX 660?

A: The GTX 660's average benchmark score is 9,022, placing it in the 45th percentile of all GPUs. Its nearest rival is the NVIDIA TITAN V CEO Edition, which scores 9,037, a 0.2% difference. It also trails the NVIDIA GeForce GTX 560 by 0.4% but leads the AMD Radeon 550X by 1.2% and the AMD Radeon Pro WX 5100 by 1.8%.

Q: What are the key memory differences between the two cards?

A: The GTX 660 has a clear advantage in memory capacity and bandwidth. It features 2 GB of GDDR5 memory on a 192-bit bus, providing 144.2 GB/s of bandwidth. The GTX 465 has 1 GB of GDDR5 memory on a 256-bit bus, yielding 102.7 GB/s of bandwidth. The GTX 660 also runs its memory at a higher effective speed of 6 Gbps compared to the GTX 465's 3.2 Gbps.

Q: Do both cards support the same modern APIs?

A: Both cards support DirectX 12 (11_0) and OpenGL 4.6. However, the GTX 660 is the only one of the two with Vulkan support, listed at version 1.2.175, while the GTX 465 has no Vulkan support listed.

Q: What are the power requirements for each card?

A: The GTX 465 has a significantly higher thermal design power (TDP) of 200 W and requires a 550 W power supply with two 6-pin connectors. The GTX 660 is much more efficient, with a TDP of 140 W, a suggested 300 W power supply, and only a single 6-pin connector.

Architecture Differences

The two cards are built on fundamentally different architectures and manufacturing processes. The GTX 465 uses the GF100 chip based on the Fermi architecture, fabricated on a 40 nm process at TSMC. This results in a large die size of 529 mm² and a transistor count of 3,100 million, yielding a transistor density of 5.9 million transistors per square millimeter. In contrast, the GTX 660 uses the GK106 chip based on the Kepler architecture, also from TSMC but on a more advanced 28 nm process. This allows for a much smaller die size of 221 mm² with 2,540 million transistors, achieving a higher density of 11.5 million transistors per square millimeter.

The compute core configurations differ dramatically. The GTX 465 features 352 shading units, 44 texture mapping units (TMUs), and 32 render output units (ROPs). The GTX 660, despite its smaller die, packs 960 shading units, 80 TMUs, and 24 ROPs. This means the GTX 660 has nearly three times the shading units and almost double the TMUs, while having fewer ROPs. These changes reflect a shift in design philosophy from Fermi, which favored more complex, flexible cores, to Kepler, which prioritized many simpler, more efficient cores.

The memory subsystems also differ. The GTX 465 uses a 256-bit memory bus with 1 GB of GDDR5, while the GTX 660 uses a narrower 192-bit bus but compensates with a larger 2 GB capacity and faster memory. The GTX 660's effective memory speed is 6 Gbps, double that of the GTX 465's 3.2 Gbps, resulting in higher bandwidth of 144.2 GB/s versus 102.7 GB/s. The GTX 660 also includes a base clock of 980 MHz and a boost clock of 1032 MHz, whereas the GTX 465 has no listed base or boost clocks.

Head-to-Head Benchmarks

The only direct benchmark comparison available is the Geekbench OpenCL test, which measures raw compute performance. In this test, the NVIDIA GeForce GTX 660 scored 11,347 points, while the NVIDIA GeForce GTX 465 scored 9,294 points. The delta percentage of -18.1% indicates that the GTX 465's score is 18.1% lower than the GTX 660's score. This is a substantial margin and confirms the GTX 660's superior compute throughput.

This result is consistent with the architectural differences. The GTX 660's 960 shading units and higher clock speeds allow it to perform significantly more floating-point operations. Its FP32 performance is rated at 1.981 TFLOPS, compared to the GTX 465's 855.4 GFLOPS. This more than doubling of compute capability translates directly into the OpenCL score. The GTX 660 also achieves a higher texture rate of 82.56 GTexel/s versus the GTX 465's 26.75 GTexel/s, and a higher pixel rate of 20.64 GPixel/s versus 13.38 GPixel/s, indicating advantages in both texturing and pixel processing.

While the GTX 660 wins the only head-to-head test, the data shows 0 wins for the GTX 465 and 1 win for the GTX 660. It is importantly the GTX 660 also has additional benchmark scores for Metal (4,305) and Vulkan (11,415) APIs, which are not available for the GTX 465, suggesting the newer card has broader API support for different workloads.

The Verdict

Based strictly on the data, the NVIDIA GeForce GTX 660 is the superior performer. Its 18.1% lead in the OpenCL benchmark is decisive, and its architectural advantages in shading units, texture units, memory capacity, and bandwidth make it the more capable card for modern workloads. The GTX 465, while a powerful card in its time, is limited by its older Fermi architecture and smaller memory pool.

The GTX 660 is the clear choice for users who prioritize compute performance and want a card that can handle more demanding tasks. Its higher average benchmark score of 9,022 versus 9,294 for the GTX 465 is a nuanced point; while the GTX 465 has a higher average score across all its benchmarks, the GTX 660's single OpenCL score is much higher. The GTX 660 also offers Vulkan support, which the GTX 465 lacks, making it more future-proof for certain applications.

The GTX 465 should only be considered if the specific lack of Vulkan support in the GTX 660 is not a concern, or if the user has a specific need for the GTX 465's unique configuration. However, the data overwhelmingly favors the GTX 660 as the better performing and more efficient option.

Specification Differences

The table below highlights the key specifications where the two cards differ, based solely on the provided data.

| Specification | NVIDIA GeForce GTX 465 | NVIDIA GeForce GTX 660 |

|:--- |:--- |:--- |

| Architecture | Fermi | Kepler |

| Chip | GF100 | GK106 |

| Generation | GeForce 400 | GeForce 600 |

| Process Node | 40 nm | 28 nm |

| Transistors | 3,100 million | 2,540 million |

| Die Size | 529 mm² | 221 mm² |

| Transistor Density | 5.9M / mm² | 11.5M / mm² |

| Base Clock | N/A | 980 MHz |

| Boost Clock | N/A | 1032 MHz |

| Memory Clock | 802 MHz (3.2 Gbps) | 1502 MHz (6 Gbps) |

| Memory Size | 1024 MB | 2 GB |

| Memory Bus Width | 256 bit | 192 bit |

| Memory Bandwidth | 102.7 GB/s | 144.2 GB/s |

| Shading Units | 352 | 960 |

| TMUs | 44 | 80 |

| ROPs | 32 | 24 |

| Pixel Rate | 13.38 GPixel/s | 20.64 GPixel/s |

| Texture Rate | 26.75 GTexel/s | 82.56 GTexel/s |

| FP32 Performance | 855.4 GFLOPS | 1.981 TFLOPS |

| TDP | 200 W | 140 W |

| Power Connectors | 2x 6-pin | 1x 6-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 | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| Vulkan Support | N/A | 1.2.175 |

| Launch MSRP | 279 USD | 229 USD |

Where Each One Wins

The benchmark and specification data point to distinct areas where each card has an advantage.

The NVIDIA GeForce GTX 660 wins in raw compute performance, as shown by its 18.1% lead in the OpenCL benchmark. Its higher FP32 throughput (1.981 TFLOPS vs 855.4 GFLOPS) and superior texture rate (82.56 GTexel/s vs 26.75 GTexel/s) indicate it is better suited for tasks that are heavily parallel, such as video encoding, physics simulations, and modern game rendering that leverages many shader cores. The GTX 660 also wins on memory capacity (2 GB vs 1 GB) and bandwidth (144.2 GB/s vs 102.7 GB/s), which is crucial for higher-resolution textures and data-intensive workloads. Finally, the GTX 660 is the only card with Vulkan support, making it the better choice for applications that use this modern API.

The NVIDIA GeForce GTX 465 has only one notable advantage: its higher number of ROPs (32 vs 24). This could imply a theoretical advantage in raw pixel fill rate at lower resolutions, though its overall pixel rate (13.38 GPixel/s) is actually lower than the GTX 660's (20.64 GPixel/s). The GTX 465 also has a wider 256-bit memory bus, which could be beneficial in specific, older workloads that are sensitive to bus width rather than total bandwidth. However, given its lower bandwidth, this is not a practical win. Ultimately, the data suggests the GTX 465's strengths are mostly historical, while the GTX 660 provides a more compelling and modern feature set.