NVIDIA GeForce GTX 560 SE vs NVIDIA GeForce GTX 750 Comparison

The NVIDIA GeForce GTX 750 and NVIDIA GeForce GTX 560 SE are two end-of-life graphics cards from different generations, and the benchmark data reveals a clear, if narrow, hierarchy between them. While both cards land in the same performance tier according to their average benchmark scores, the GeForce GTX 750 demonstrates a decisive advantage in the one test where they go head-to-head, suggesting that architectural efficiency can outweigh raw specifications like memory bandwidth and texture units.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, and the results are unambiguous. The GeForce GTX 750 scores 9315, while the GeForce GTX 560 SE scores 7171. This represents a 29.9% advantage for the GTX 750, a substantial margin that firmly establishes it as the faster card in compute workloads. This is not a marginal victory; the GTX 750 is nearly a third faster, which is a significant generational leap.

Looking at the broader average benchmark scores, the GTX 750's average is 7222, which is only 0.7% higher than the GTX 560 SE's average of 7171. This seems contradictory at first glance. The head-to-head shows a 29.9% lead, but the averages are nearly identical. The explanation lies in the number of benchmarks each card was tested on. The GTX 750 has three benchmark scores (Geekbench Metal, OpenCL, and Vulkan), while the GTX 560 SE only has one (Geekbench OpenCL). The GTX 750's average is pulled down by its lower scores in the Metal (4274) and Vulkan (8078) tests, even though its OpenCL score is far superior. The data suggests that the GTX 560 SE's single data point is not representative of its overall capability, or perhaps the OpenCL test is particularly well-optimized for the Maxwell architecture.

The deltaPct values in the nearestRivals lists confirm this ranking. For the GTX 750, the GTX 560 SE is listed as a rival with a deltaPct of 0.7, meaning the GTX 750 is 0.7% faster on average. Conversely, for the GTX 560 SE, the GTX 750 is listed with a deltaPct of -0.7, meaning the GTX 560 SE is 0.7% slower on average. While the overall averages are close, the single head-to-head result is the more reliable indicator of direct performance, and it heavily favors the GTX 750.

Where Each One Wins

Based strictly on the data, the GeForce GTX 750 wins the only direct benchmark comparison. It is the clear victor in the Geekbench OpenCL test, which is a general-purpose compute benchmark. This suggests that for tasks leveraging OpenCL, such as certain video encoding, physics simulations, or other GPU-accelerated applications, the GTX 750 will deliver noticeably better performance. The 29.9% lead is not trivial; it could translate into meaningfully shorter render times or smoother real-time effects.

The GeForce GTX 560 SE does not win any of the recorded benchmarks. Its single score of 7171 in Geekbench OpenCL is its only data point, and it is significantly lower than the GTX 750's score in the same test. However, its average benchmark score of 7171 places it in the same percentile tier (39th) as the GTX 750 (40th percentile). This indicates that in the broader ecosystem of all GPUs, the GTX 560 SE is not an outlier. It is a mid-to-low-range card from its era, and its performance is consistent with that positioning. The GTX 750, while also in the 40th percentile, clearly has a higher ceiling in compute performance.

Therefore, the GTX 750 is the winner for any workload that is compute-bound. The GTX 560 SE's potential strengths, such as its higher texture fill rate (35.33 GTexel/s vs 34.72 GTexel/s), do not manifest in any recorded benchmark victory. The data simply does not support a scenario where the GTX 560 SE outperforms the GTX 750.

Architecture Differences

The two cards are built on fundamentally different architectures, which explains the performance disparity. The GTX 750 uses the GM107 chip, based on the Maxwell architecture, manufactured on a 28 nm process at TSMC. In contrast, the GTX 560 SE uses the GF114 chip, based on the older Fermi 2.0 architecture, manufactured on a larger 40 nm process, also at TSMC.

The transistor counts are similar—1,870 million for the GTX 750 and 1,950 million for the GTX 560 SE—but the die size tells a different story. The GTX 750's die is a compact 148 mm², while the GTX 560 SE's die is a sprawling 332 mm². This results in a transistor density of 12.6M / mm² for the GTX 750 versus just 5.9M / mm² for the GTX 560 SE. The Maxwell architecture is far more efficient, packing nearly twice the transistors into the same area. This efficiency is a major factor in the GTX 750's superior performance-per-watt, though the data only shows the TDP figures.

The memory subsystems also differ significantly. Both cards have 1024 MB of GDDR5 memory, but the GTX 750 uses a 128-bit bus, while the GTX 560 SE uses a wider 192-bit bus. Despite the narrower bus, the GTX 750's memory runs at 1253 MHz (5 Gbps effective), yielding a bandwidth of 80.19 GB/s. The GTX 560 SE's memory runs at 957 MHz (3.8 Gbps effective), but the wider bus gives it a higher bandwidth of 91.87 GB/s. So the GTX 560 SE has a memory bandwidth advantage, but it does not translate into a performance win.

The compute resources are arranged differently as well. The GTX 750 has 512 shading units, 32 TMUs, and 16 ROPs. The GTX 560 SE has fewer shading units (288) but more TMUs (48) and ROPs (24). This configuration gives the GTX 560 SE a slightly higher texture fill rate (35.33 GTexel/s vs 34.72 GTexel/s) but a much lower pixel rate (8.832 GPixel/s vs 17.36 GPixel/s). The GTX 750's FP32 performance is 1,111.0 GFLOPS, significantly higher than the GTX 560 SE's 847.9 GFLOPS. The massive difference in pixel rate (nearly double) is a strong indicator of the GTX 750's architectural efficiency.

The Verdict

The data points to a single, clear conclusion: the NVIDIA GeForce GTX 750 is the superior card. It wins the only direct benchmark comparison by a 29.9% margin in OpenCL, and it achieves this with a lower TDP (55 W vs 150 W), a smaller die, and no power connectors required. The GTX 750 is also a single-slot card, whereas the GTX 560 SE is a dual-slot card requiring two 6-pin power connectors.

For a user looking at these two historical cards, the GTX 750 is the obvious choice. It offers significantly better compute performance, consumes far less power, and is physically easier to install. The GTX 560 SE's only advantages are a higher memory bandwidth (91.87 GB/s vs 80.19 GB/s) and a slightly higher texture fill rate, but these do not appear in any benchmark win. The GTX 750's architecture is simply more advanced and efficient. The data suggests that anyone choosing the GTX 560 SE would be sacrificing a large amount of performance for no measurable benefit. The GTX 750 is the definitive winner in this comparison.

FAQ

Q: Which GPU is faster in the Geekbench OpenCL test?

A: The NVIDIA GeForce GTX 750 is faster, scoring 9315 compared to the GTX 560 SE's 7171, a 29.9% difference.

Q: How do their average benchmark scores compare?

A: The GTX 750 has an average benchmark score of 7222, which is 0.7% higher than the GTX 560 SE's average of 7171.

Q: What are the key architectural differences?

A: The GTX 750 uses the Maxwell architecture (GM107 chip) on a 28 nm process, while the GTX 560 SE uses the Fermi 2.0 architecture (GF114 chip) on a 40 nm process. The GTX 750 has a die size of 148 mm², while the GTX 560 SE has a die size of 332 mm².

Q: Which card has more shading units?

A: The GTX 750 has 512 shading units, while the GTX 560 SE has 288 shading units.

Q: Does the GTX 560 SE have any performance advantage?

A: The data shows no benchmark wins for the GTX 560 SE. It does have higher memory bandwidth (91.87 GB/s vs 80.19 GB/s) and a slightly higher texture fill rate (35.33 GTexel/s vs 34.72 GTexel/s), but these do not translate into a higher benchmark score.

Q: What is the TDP of each card?

A: The GTX 750 has a TDP of 55 W, while the GTX 560 SE has a TDP of 150 W.

Specification Differences

| Specification | NVIDIA GeForce GTX 750 | NVIDIA GeForce GTX 560 SE |

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

| Chip | GM107 | GF114 |

| Architecture | Maxwell | Fermi 2.0 |

| Process Node | 28 nm | 40 nm |

| Transistors | 1,870 million | 1,950 million |

| Die Size | 148 mm² | 332 mm² |

| Transistor Density | 12.6M / mm² | 5.9M / mm² |

| Base Clock | 1020 MHz | null |

| Boost Clock | 1085 MHz | null |

| Memory Clock | 1253 MHz (5 Gbps effective) | 957 MHz (3.8 Gbps effective) |

| Memory Bus Width | 128 bit | 192 bit |

| Memory Bandwidth | 80.19 GB/s | 91.87 GB/s |

| Shading Units | 512 | 288 |

| TMUs | 32 | 48 |

| ROPs | 16 | 24 |

| Pixel Rate | 17.36 GPixel/s | 8.832 GPixel/s |

| Texture Rate | 34.72 GTexel/s | 35.33 GTexel/s |

| FP32 | 1,111.0 GFLOPS | 847.9 GFLOPS |

| TDP | 55 W | 150 W |

| Slot Width | Single-slot | Dual-slot |

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

| Suggested PSU | 250 W | 450 W |

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

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

| Vulkan Support | 1.4 | null |

| Release Date | 2014-02-17 | 2012-02-19 |