NVIDIA GeForce GTX 675MX vs NVIDIA GeForce GTX 960 Comparison

Head-to-Head Benchmarks

The recorded data shows a clear, though not uniform, performance gap between the NVIDIA GeForce GTX 960 and the NVIDIA GeForce GTX 675MX. Across the two shared benchmark tests, the GTX 960 wins decisively in both, with no test favoring the older mobile part. The most dramatic divergence appears in the Geekbench OpenCL test, where the GTX 960 scores 18,925 against the GTX 675MX's 10,723. That is a 76.5% advantage for the GTX 960, a massive margin that speaks to far superior compute throughput under this particular workload.

The Geekbench Metal test narrows the gap somewhat but still shows a substantial lead for the GTX 960. Here, the GTX 960 posts 8,773 points while the GTX 675MX manages 6,131, a 43.1% difference. While the OpenCL result might exaggerate the differences in real-world gaming scenarios, the Metal result reinforces that the GTX 960 is fundamentally faster in general-purpose GPU compute. The GTX 960 also holds an additional benchmark result, a 3DMark Steel Nomad DX12 score of 162, which the GTX 675MX does not have a recorded score for in the database. Similarly, the GTX 960 has a Geekbench Vulkan score of 9,231, further demonstrating its broader API support and modern feature set.

When looking at average benchmark scores across all recorded tests, the GTX 960 averages 9,273 points, placing it in the 45th percentile of all GPUs. The GTX 675MX, by contrast, averages only 8,427 points, which lands it in the 43rd percentile. The delta between these averages is about 10%, which is substantial but less dramatic than the peak OpenCL gap. The nearest rivals for the GTX 960 include the NVIDIA GeForce GTX 465, which scores 9,294 (a 0.2% difference), and the AMD Radeon R7 M380 at 9,313 (0.4% difference). For the GTX 675MX, the closest competitor is the AMD Radeon 880M, which scores 8,436, a 0.1% difference. These rivalries show that the GTX 960 is competing in a higher performance tier than the GTX 675MX, which sits closer to modern integrated graphics or low-end mobile parts.

Architecture Differences

The underlying silicon tells a story of two very different design philosophies and eras. The GTX 960 is built on the GM206 chip, using NVIDIA's Maxwell 2.0 architecture, which was introduced with the GeForce 900 generation. In contrast, the GTX 675MX uses the GK104 chip, based on the older Kepler architecture, from the GeForce 600M generation. Both are manufactured on the same 28 nm process node at TSMC, but they diverge significantly in transistor counts and die size. The GTX 960 contains 2,940 million transistors on a 228 mm² die, resulting in a transistor density of 12.9 million per square millimeter. The GTX 675MX packs more transistors overall at 3,540 million, but on a much larger 294 mm² die, yielding a lower density of 12.0 million per square millimeter.

These architectural changes directly impact execution resources. The GTX 960 has 1,024 shading units, 64 texture mapping units, and 32 ROPs. The GTX 675MX has fewer shading units at 960, but more TMUs at 80, and the same 32 ROPs. Despite having more TMUs, the GTX 675MX's lower clock speeds (the data does not list base or boost clocks, only memory clocks) result in a texture rate of 52.32 GTexel/s, which is significantly lower than the GTX 960's 75.39 GTexel/s. The pixel rate shows a similar story: the GTX 960 achieves 37.70 GPixel/s, while the GTX 675MX only manages 13.08 GPixel/s. These numbers indicate that the Maxwell architecture in the GTX 960 is far more efficient at issuing and completing work per clock cycle.

Memory subsystems differ as well. Both cards have 2 GB of GDDR5 memory, but the GTX 960 uses a 128-bit memory bus, while the GTX 675MX uses a wider 256-bit bus. The GTX 960 compensates for its narrower bus with a much higher memory clock: 1,753 MHz (7 Gbps effective) versus the GTX 675MX's 900 MHz (3.6 Gbps effective). The result is that the GTX 960 delivers 112.2 GB/s of bandwidth, while the GTX 675MX delivers 115.2 GB/s. The wider bus on the older card nearly offsets the slower memory speed, but the GTX 960 still wins on raw compute performance.

API support also reflects the architectural leap. The GTX 960 supports DirectX 12 (12_1), while the GTX 675MX only supports DirectX 12 (11_0). Both support OpenGL 4.6, but the GTX 960 supports Vulkan 1.4, whereas the GTX 675MX is limited to Vulkan 1.2.175. The GTX 960 also has a higher FP32 (single-precision) compute rate at 2.413 TFLOPS, compared to the GTX 675MX's 1,255.7 GFLOPS. This nearly 2x difference in raw compute power is the root cause of the massive OpenCL benchmark gap.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 960 has an average benchmark score of 9,273, placing it in the 45th percentile of all GPUs. The NVIDIA GeForce GTX 675MX averages 8,427, which lands in the 43rd percentile.

Q: How does the GTX 960 compare to the GTX 675MX in the Geekbench Metal test?

A: The GTX 960 scores 8,773, which is 43.1% higher than the GTX 675MX's score of 6,131.

Q: What is the memory configuration difference between the two cards?

A: Both have 2 GB of GDDR5 memory. The GTX 960 uses a 128-bit bus with a 1,753 MHz memory clock (7 Gbps effective), yielding a bandwidth of 112.2 GB/s. The GTX 675MX uses a 256-bit bus with a 900 MHz memory clock (3.6 Gbps effective), yielding a bandwidth of 115.2 GB/s.

Q: Which GPU supports more modern DirectX versions?

A: The GTX 960 supports DirectX 12 (12_1), while the GTX 675MX only supports DirectX 12 (11_0).

Q: What is the difference in shading units?

A: The GTX 960 has 1,024 shading units, while the GTX 675MX has 960 shading units. The GTX 960 also has a higher texture rate (75.39 GTexel/s vs 52.32 GTexel/s) and pixel rate (37.70 GPixel/s vs 13.08 GPixel/s).

Q: Is the GTX 675MX a mobile or desktop GPU?

A: The GTX 675MX is a mobile GPU, as indicated by its MXM Module slot width, its MXM-B (3.0) bus interface, and its "Portable Device Dependent" display outputs. The GTX 960 is a desktop card with a dual-slot design and a PCIe 3.0 x16 interface.

The Verdict

The data is unambiguous. The NVIDIA GeForce GTX 960 is the stronger performer across every recorded benchmark. Its average score is 8,273, which is roughly 5.5% higher than the GTX 675MX's 8,427. The biggest win for the GTX 960 is in OpenCL, where it leads by 76.5%, followed by a 43.1% lead in Metal. The architecture is also more modern, with higher shading unit counts, higher pixel and texture rates, and better API support (DirectX 12_1 vs 11_0, and Vulkan 1.4 vs 1.2.175). The GTX 960 also has a higher FP32 compute throughput (2.413 TFLOPS vs 1,255.7 GFLOPS). The only metric where the GTX 675MX shows a theoretical advantage is memory bandwidth, but it comes at the cost of wider bus, and the data shows that this does not translate into real-world performance wins.

For a user strictly looking for the highest frame rates in modern titles, the GTX 960 is the obvious choice. The GTX 675MX, while being a capable mobile part in its day, falls behind. The GTX 960's modern API support and superior compute performance make it better suited for both current games and compute-intensive tasks. The GTX 675MX is an older part, and its performance profile aligns more with its position in the 43rd percentile, while the GTX 960 sits in the 45th percentile. The verdict is a clear win for the GTX 960.

Specification Differences

| Specification | NVIDIA GeForce GTX 960 | NVIDIA GeForce GTX 675MX | Difference |

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

| Architecture | Maxwell 2.0 | Kepler | Eras apart |

| Generation | GeForce 900 | GeForce 600M | 2 generations newer |

| Chip | GM206 | GK104 | Different dies |

| Transistors | 2,940 million | 3,540 million | GTX 675MX has more |

| Die Size | 228 mm² | 294 mm² | GTX 675MX is larger |

| Transistor Density | 12.9M/mm² | 12.0M/mm² | GTX 960 is denser |

| Base Clock | 1,127 MHz | N/A | GTX 960 has a base clock |

| Memory Clock | 1,753 MHz (7 Gbps) | 900 MHz (3.6 Gbps) | GTX 960 is faster |

| Memory Bus | 128 bit | 256 bit | GTX 675MX is wider |

| Memory Bandwidth | 112.2 GB/s | 115.2 GB/s | GTX 675MX has slight edge |

| Shading Units | 1,024 | 960 | GTX 960 has more |

| TMUs | 64 | 80 | GTX 675MX has more |

| Pixel Rate | 37.70 GPixel/s | 13.08 GPixel/s | GTX 960 is 3x faster |

| Texture Rate | 75.39 GTexel/s | 52.32 GTexel/s | GTX 960 is faster |

| FP32 Compute | 2.413 TFLOPS | 1,255.7 GFLOPS | GTX 960 is ~2x faster |

| TDP | 120 W | 100 W | GTX 675MX is lower power |

| Slot Width | Dual-slot | MXM Module | Desktop vs Mobile |

| Power Connectors | 1x 6-pin | None | GTX 960 requires external power |

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

| DirectX | 12 (12_1) | 12 (11_0) | GTX 960 is more modern |

| Vulkan | 1.4 | 1.2.175 | GTX 960 is newer |

| Length | 241 mm (9.5 inches) | N/A | GTX 960 is defined |

Where Each One Wins

The GTX 960 wins in pure compute performance. Its 2.413 TFLOPS of FP32 throughput, compared to the 1,255.7 GFLOPS of the GTX 675MX, results in a massive 76.5% lead in the OpenCL benchmark. For users running general-purpose compute workloads, such as video encoding, physics simulations, or OpenCL-accelerated applications, the GTX 960 is the clear winner. Its higher pixel rate (37.70 GPixel/s vs 13.08 GPixel/s) also suggests it can handle more demanding resolution and geometry tasks, though the benchmark data does not directly measure gaming framerates.

The GTX 675MX, despite its age, holds a small advantage in memory bandwidth (115.2 GB/s vs 112.2 GB/s) due to its wider 256-bit bus. This could theoretically help in bandwidth-bound scenarios, but the recorded benchmark data shows no test where this translates into a win. Its lower TDP of 100 W compared to 120 W makes it more power-efficient on paper, and it has no power connector requirement, which is an advantage for thin mobile designs. However, the actual benchmark results show that the GTX 675MX loses in every shared test.

In terms of API support, the GTX 960 is the only one that supports DirectX 12_1 and Vulkan 1.4, which means it is more future-proof for modern titles that use these features. The GTX 675MX is limited to DirectX 12 (11_0) and Vulkan 1.2.175, which may not expose all the rendering features of newer games. The recorded data suggests that for any user prioritizing performance, modern API support, or raw compute power, the GTX 960 is the only logical choice. The GTX 675MX only "wins" in the sense of directly lower power consumption and a physically smaller footprint, but no benchmark data supports a performance win for it.