NVIDIA GeForce GT 645M vs NVIDIA GeForce GTX 560M Comparison

Head-to-Head Benchmarks

The recorded benchmark data contains a single direct comparison between these two mobile graphics processors, and the result is decisively lopsided. In the Geekbench OpenCL test, the NVIDIA GeForce GTX 560M scores 4,855 points, while the NVIDIA GeForce GT 645M manages only 2,680 points. That works out to an 81.2% advantage for the older Fermi-based part, a gap so wide that it essentially settles the compute-performance question before any other architectural factors are considered.

This result is worth pausing over. The GTX 560M is the earlier product, released in May 2011, while the GT 645M arrived a full year and a half later, in September 2012. Yet the older chip wins the only head-to-head test in the database by a margin that rivals see as a generational leap, not a small iterative gain. The delta of 81.2% means the GTX 560M delivers nearly double the OpenCL score of the GT 645M. For users running OpenCL workloads, the choice is not close.

However, the GT 645M does not appear in the head-to-head comparison with a total loss across all tests, because the database records only one shared benchmark. The GTX 560M takes the single win, giving it a 1-0 record in direct matchups. That does not mean the GT 645M is without merit elsewhere; it simply means the shared test pool is thin. The GT 645M does have its own benchmark results in Geekbench Metal and Geekbench Vulkan, but no corresponding scores exist for the GTX 560M in those APIs, so a direct comparison there is impossible from the recorded data.

Looking at the broader context, the GTX 560M's average benchmark score of 4,855 places it in the 28th percentile of all GPUs in the database. Its nearest rivals include the NVIDIA GeForce 940MX at 4,844 points, a 0.2% gap in favor of the GTX 560M, and the AMD Radeon R6 M255DX at 4,867 points, which edges it out by 0.2%. The GT 645M, meanwhile, sits at a lower average of 4,411 points, in the 26th percentile. Its closest competitor is the NVIDIA GeForce 930M at 4,388 points, which trails by 0.5%, and the Intel Iris Pro Graphics 5200 at 4,360 points, which sits 1.2% behind. Interestingly, the GT 645M also lands within 1.8% of the NVIDIA GeForce RTX 4070 GDDR6 in this particular scoring metric, though that result likely reflects the narrow scoring range of the benchmark rather than any practical equivalence.

The takeaway from the head-to-head data is straightforward: in the one test where both GPUs appear, the GTX 560M dominates the GT 645M by a massive margin. The GT 645M does not win a single recorded comparison, and its average score across all its benchmarks (4,411) is still lower than the GTX 560M's single OpenCL result (4,855). That is a notable consistency: even when averaging in its Metal and Vulkan scores, the GT 645M cannot surpass the GTX 560M's lone OpenCL number.

FAQ

Q: Which GPU wins the only direct benchmark comparison in the database?

A: The NVIDIA GeForce GTX 560M wins the Geekbench OpenCL test with a score of 4,855 against the NVIDIA GeForce GT 645M's 2,680, a delta of 81.2% in favor of the GTX 560M.

Q: How does the GTX 560M compare to its nearest rivals in average benchmark score?

A: The GTX 560M's average score is 4,855. It is 0.2% ahead of the NVIDIA GeForce 940MX (4,844), 0.2% behind the AMD Radeon R6 M255DX (4,867), 0.8% ahead of the NVIDIA GeForce GTS 450 (4,893, which actually scores higher), and 0.9% ahead of the NVIDIA GeForce RTX 5060 Ti 8 GB (4,901, which also scores higher).

Q: Does the GT 645M have any benchmark wins over the GTX 560M?

A: No. The only shared benchmark, Geekbench OpenCL, is won by the GTX 560M. The GT 645M's wins tally is 0 in direct head-to-head matchups.

Q: What is the GT 645M's average benchmark score and percentile?

A: The GT 645M has an average benchmark score of 4,411 and sits in the 26th percentile of all GPUs. Its closest rival, the NVIDIA GeForce 930M, scores 4,388, which is 0.5% lower.

Q: How do the two GPUs compare in memory bandwidth?

A: The GTX 560M uses GDDR5 memory with a 192-bit bus, delivering 60.00 GB/s of bandwidth. The GT 645M uses DDR3 memory with a 128-bit bus, delivering 28.80 GB/s, less than half the bandwidth.

Q: Are there any benchmark results for the GTX 560M in Metal or Vulkan?

A: The database lists no Metal or Vulkan scores for the GTX 560M. The GT 645M does have scores in those APIs (5,679 in Metal and 4,875 in Vulkan), but without a GTX 560M result, no comparison is possible.

Architecture Differences

The two GPUs represent distinct architectural generations from NVIDIA, and the recorded specifications highlight how much changed between them. The GTX 560M is built on the Fermi 2.0 architecture, using the GF116 chip fabricated on a 40 nm process at TSMC. The GT 645M, by contrast, uses the Kepler architecture with the GK107 chip on a 28 nm process, also from TSMC. This process shrink is significant: the GTX 560M packs 1,170 million transistors into a 238 mm² die, yielding a transistor density of 4.9 million per square millimeter. The GT 645M fits 1,270 million transistors into just 118 mm², for a density of 10.8 million per square millimeter. The Kepler part crams more transistors into roughly half the silicon area, a direct reflection of the newer manufacturing node.

The shading unit counts differ dramatically. The GTX 560M has 192 shading units, while the GT 645M has 384, exactly double. However, this doubling does not translate into a compute advantage in the recorded OpenCL test, where the GTX 560M wins by 81.2%. This suggests that Fermi's per-shader efficiency, or its memory subsystem, more than compensates for the raw unit count. Both GPUs have 32 texture mapping units, so texture throughput is nearly identical: 24.80 GTexel/s for the GTX 560M versus 24.96 GTexel/s for the GT 645M. The pixel rate is also close, with the GTX 560M at 6.200 GPixel/s and the GT 645M at 6.240 GPixel/s. The GT 645M has fewer ROPs (16 versus 24), yet still achieves a marginally higher pixel rate, likely due to its higher clocks.

Clock speeds tell a similar story of architectural divergence. The GTX 560M's memory runs at 625 MHz, with an effective data rate of 2.5 Gbps. The GT 645M's memory runs at 900 MHz, but with an effective rate of only 1800 Mbps, because it uses DDR3 rather than GDDR5. The GT 645M does have explicit base and boost clocks of 709 MHz and 780 MHz, respectively, while the GTX 560M's base and boost clocks are not recorded in the database. Despite the GT 645M's higher memory clock, its bandwidth is far lower: 28.80 GB/s versus 60.00 GB/s for the GTX 560M. The GTX 560M's 192-bit bus and GDDR5 memory provide a substantial bandwidth advantage that likely explains its OpenCL dominance.

The FP32 compute figures are surprisingly close. The GTX 560M delivers 595.2 GFLOPS, while the GT 645M delivers 599.0 GFLOPS. That is a difference of less than 1%, yet the benchmark results show an 81.2% gap. This points to memory bandwidth, not raw compute, as the limiting factor in the OpenCL workload. The GTX 560M's 60.00 GB/s bandwidth versus 28.80 GB/s for the GT 645M is the most plausible explanation for the lopsided result.

Power consumption is another major differentiator. The GTX 560M has a TDP of 75 W and uses an MXM Module slot, with no power connectors. The GT 645M has a TDP of just 32 W and is an IGP (integrated graphics processor) with a PCIe 3.0 x16 bus interface. The GT 645M is clearly designed for lower-power laptops, while the GTX 560M targets more substantial mobile systems. The GTX 560M's bus interface is MXM-B (3.0), which is a modular mobile form factor, while the GT 645M uses standard PCIe.

API support also differs slightly. Both support DirectX 12 (11_0) and OpenGL 4.6. However, the GT 645M explicitly supports Vulkan 1.2.175, while the GTX 560M's Vulkan support is not recorded in the database. This means the newer Kepler part has a software feature that the older Fermi part may lack, though the absence of a recorded Vulkan score for the GTX 560M makes it impossible to verify its Vulkan performance.

The Verdict

The data points to a clear winner for raw compute performance: the NVIDIA GeForce GTX 560M. It wins the only head-to-head benchmark by 81.2%, and its average score of 4,855 exceeds the GT 645M's average of 4,411 by roughly 10%. For any workload that relies on OpenCL, the GTX 560M is the superior choice, despite being the older product.

However, the GT 645M has its own advantages that the benchmark data alone does not capture. Its TDP of 32 W is less than half the GTX 560M's 75 W, making it far more suitable for thin-and-light laptops where power efficiency is critical. Its Kepler architecture is newer, supports Vulkan 1.2.175, and its 2 GB of DDR3 memory offers more capacity than the GTX 560M's 1536 MB, even if the bandwidth is far lower. The GT 645M also has explicit base and boost clocks, which suggests more predictable performance scaling in a thermally constrained chassis.

The GTX 560M is the pick for users who prioritize compute performance and are willing to accept higher power draw. The GT 645M is the pick for users who need a low-power solution with modern API support and more memory capacity. The benchmark data does not support the GT 645M as a compute performer, but its architectural efficiency and feature set make it a reasonable choice for specific use cases.

Specification Differences

| Specification | NVIDIA GeForce GTX 560M | NVIDIA GeForce GT 645M |

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

| Architecture | Fermi 2.0 | Kepler |

| Chip | GF116 | GK107 |

| Process Node | 40 nm | 28 nm |

| Die Size | 238 mm² | 118 mm² |

| Transistor Density | 4.9M / mm² | 10.8M / mm² |

| Memory Size | 1536 MB | 2 GB |

| Memory Type | GDDR5 | DDR3 |

| Memory Bus Width | 192 bit | 128 bit |

| Memory Bandwidth | 60.00 GB/s | 28.80 GB/s |

| Memory Clock | 625 MHz (2.5 Gbps effective) | 900 MHz (1800 Mbps effective) |

| Shading Units | 192 | 384 |

| ROPs | 24 | 16 |

| Pixel Rate | 6.200 GPixel/s | 6.240 GPixel/s |

| Texture Rate | 24.80 GTexel/s | 24.96 GTexel/s |

| FP32 | 595.2 GFLOPS | 599.0 GFLOPS |

| TDP | 75 W | 32 W |

| Slot Width | MXM Module | IGP |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x16 |

| Vulkan Support | Not recorded | 1.2.175 |

| Release Date | 2011-05-29 | 2012-09-30 |

| Predecessor | GeForce 400M | GeForce 500M |

| Successor | GeForce 600M | GeForce 700M |

Where Each One Wins

The GTX 560M wins decisively in compute-bound scenarios. Its OpenCL score of 4,855 versus 2,680 for the GT 645M represents an 81.2% advantage, and its memory bandwidth of 60.00 GB/s is more than double the GT 645M's 28.80 GB/s. For tasks like GPU-accelerated rendering, physics simulation, or any OpenCL-based workload, the GTX 560M is the clear choice. Its 192-bit GDDR5 memory interface is a major asset, and its higher TDP of 75 W suggests it can sustain performance under load without the thermal throttling that might affect a 32 W part.

The GT 645M wins in efficiency and modern feature support. Its 32 W TDP makes it suitable for systems where battery life and cooling are paramount. Its Kepler architecture is a generation newer, and it explicitly supports Vulkan 1.2.175, which the GTX 560M does not record. The GT 645M also offers 2 GB of memory versus 1536 MB, which can be helpful for larger texture sets or higher-resolution framebuffers, even if the DDR3 bandwidth is a bottleneck. Its 384 shading units, double the GTX 560M's count, may also benefit workloads that are shading-unit-bound rather than bandwidth-bound, though the recorded OpenCL test does not reflect this.

In terms of raw throughput, the two GPUs are nearly identical. The GTX 560M's pixel rate of 6.200 GPixel/s is only 0.6% lower than the GT 645M's 6.240 GPixel/s, and the texture rates differ by just 0.6% as well. FP32 compute is also a tie, with 595.2 GFLOPS versus 599.0 GFLOPS. The GTX 560M does have more ROPs (24 versus 16), but the GT 645M's higher clocks compensate in the pixel-rate calculation.

The GTX 560M is the winner for anyone who needs maximum compute performance in a mobile form factor and can tolerate higher power consumption. The GT 645M is the winner for anyone who prioritizes low power, modern API support, and memory capacity over raw OpenCL speed. The benchmark data favors the GTX 560M in the only direct test, but the GT 645M's architectural advantages suggest it is not without its own niche.