NVIDIA GeForce GTX 650 Ti vs NVIDIA GeForce GTX 675MX Comparison

The NVIDIA GeForce GTX 675MX and the NVIDIA GeForce GTX 650 Ti are both Kepler-based graphics cards from the same generation, but they target different segments of the market. The GTX 675MX is a mobile-class processor for laptops, while the GTX 650 Ti is a desktop card. Benchmark data shows the GTX 675MX holds a clear performance edge in the one shared test, but the GTX 650 Ti counters with a more robust feature set for desktop use. The GTX 675MX is the pick for raw compute performance, while the GTX 650 Ti is the pick for desktop connectivity and a lower power draw profile relative to its compute output.

The Verdict

The data points to a straightforward choice based on the single available head-to-head benchmark. The GTX 675MX wins the only direct comparison, the Geekbench OpenCL test, with a score of 10,723 against the GTX 650 Ti’s 7,877. That is a 36.1% lead for the mobile card. This is a decisive margin, suggesting that the GTX 675MX is the stronger compute performer. Its average benchmark score across all tests is 8,427, which places it in the 43rd percentile of all GPUs. The GTX 650 Ti averages 8,053, sitting in the 42nd percentile. While the percentile difference is small, the raw average score difference is 374 points in favor of the GTX 675MX.

For users prioritizing raw compute throughput, the GTX 675MX is the superior choice. Its closest rivals include the AMD Radeon 880M, which scores 8,436 (a 0.1% difference), and the NVIDIA GeForce MX330, at 8,458 (a 0.4% difference). The GTX 650 Ti, meanwhile, sits near the NVIDIA GeForce GTX 880M (8,040, a 0.2% delta) and the NVIDIA GeForce GTX 650 Ti Boost (8,067, a 0.2% delta). This places the two cards in adjacent performance tiers, but the GTX 675MX is consistently ahead. The GTX 650 Ti’s only advantage in the data is its desktop form factor, which includes a standard PCIe 3.0 x16 interface and fixed display outputs, making it the logical pick for a stationary desktop build. The GTX 675MX, with its MXM-B (3.0) interface, is locked to portable devices.

Architecture Differences

Both cards use the Kepler architecture and are built on the 28 nm process node at TSMC, but they are not identical silicon. The GTX 675MX is based on the GK104 chip, a large die measuring 294 mm² and containing 3,540 million transistors. The GTX 650 Ti uses the GK106S chip, which is smaller at 221 mm² and has 2,540 million transistors. This size difference is significant: the GK104 die is 73 mm² larger, and the transistor count is 1,000 million higher on the GTX 675MX. Transistor density is close, with the GTX 675MX at 12.0 million transistors per mm² and the GTX 650 Ti at 11.5 million per mm², but the larger die gives the GTX 675MX a structural advantage.

Memory configurations also diverge. The GTX 675MX has 2 GB of GDDR5 memory on a 256-bit bus, delivering 115.2 GB/s of bandwidth. The GTX 650 Ti has 1,024 MB of GDDR5 on a 128-bit bus, providing 86.40 GB/s. The GTX 675MX thus offers double the memory capacity and 28.8 GB/s more bandwidth. The memory clock differs as well: the GTX 675MX runs at 900 MHz (3.6 Gbps effective), while the GTX 650 Ti runs at 1350 MHz (5.4 Gbps effective). The higher clock on the GTX 650 Ti partially compensates for its narrower bus, but the GTX 675MX still wins on total bandwidth.

Feature-wise, both cards support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. They both lack ray tracing and tensor cores. The physical formats are the major differentiator: the GTX 675MX is an MXM module with no power connectors and a 100 W TDP, while the GTX 650 Ti is a single-slot card with a 1x 6-pin power connector and a 110 W TDP. The GTX 650 Ti also lists a suggested PSU of 300 W, a detail absent for the GTX 675MX.

Where Each One Wins

The GTX 675MX wins decisively in compute benchmarks. Its OpenCL score of 10,723 is the only head-to-head result, and it wins that outright. Its average benchmark score of 8,427 is also higher than the GTX 650 Ti’s 8,053. The GTX 675MX also holds advantages in memory capacity (2 GB vs 1,024 MB) and memory bandwidth (115.2 GB/s vs 86.40 GB/s). These specs suggest it will handle larger textures and higher-resolution workloads more comfortably. Its pixel rate is 13.08 GPixel/s, and its texture rate is 52.32 GTexel/s.

The GTX 650 Ti wins in raw clock speeds and certain throughput metrics. Its memory clock is 1350 MHz versus 900 MHz, and its pixel rate is 14.85 GPixel/s, which is 1.77 GPixel/s higher than the GTX 675MX. Its texture rate is 59.39 GTexel/s, beating the GTX 675MX’s 52.32 GTexel/s by 7.07 GTexel/s. Its FP32 compute is 1,425.4 GFLOPS, which is 169.7 GFLOPS higher than the GTX 675MX’s 1,255.7 GFLOPS. This is a curious split: the GTX 650 Ti has higher compute and fill rates on paper, yet loses the OpenCL test. The GTX 650 Ti also has a lower TDP (110 W vs 100 W), but that is a minor difference.

The GTX 650 Ti wins on connectivity. It offers 2x DVI and 1x mini-HDMI 1.4a outputs, while the GTX 675MX is listed as "portable device dependent," meaning its outputs vary by laptop design. The GTX 650 Ti uses a standard PCIe 3.0 x16 interface, which is universal in desktop motherboards, whereas the GTX 675MX uses MXM-B (3.0), a proprietary mobile standard.

FAQ

Q: Which card has the higher average benchmark score?

A: The GTX 675MX has an average benchmark score of 8,427, while the GTX 650 Ti scores 8,053, giving the GTX 675MX a 374-point lead.

Q: What is the margin of victory in the only direct head-to-head test?

A: The GTX 675MX wins the Geekbench OpenCL test with a score of 10,723 versus 7,877 for the GTX 650 Ti, a delta of 36.1%.

Q: Do both cards support the same APIs?

A: Yes, both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, and neither has ray tracing or tensor cores.

Q: Which card has more memory bandwidth?

A: The GTX 675MX has 115.2 GB/s of bandwidth, compared to 86.40 GB/s for the GTX 650 Ti, due to its wider 256-bit bus.

Q: What is the form factor difference?

A: The GTX 675MX is an MXM module with no power connectors, while the GTX 650 Ti is a single-slot desktop card with a 1x 6-pin power connector and a 300 W suggested PSU.

Q: Which card has a higher pixel rate?

A: The GTX 650 Ti has a pixel rate of 14.85 GPixel/s, which is higher than the GTX 675MX’s 13.08 GPixel/s.

Head-to-Head Benchmarks

The only direct benchmark between these two cards is Geekbench OpenCL, and it is a decisive win for the GTX 675MX. The GTX 675MX scores 10,723, while the GTX 650 Ti scores 7,877. That is a 36.1% difference, which is a substantial margin in any compute workload. This result is notable because the GTX 650 Ti has a higher FP32 rating (1,425.4 GFLOPS vs 1,255.7 GFLOPS) and a higher texture rate (59.39 GTexel/s vs 52.32 GTexel/s). The GTX 675MX wins despite lower raw clock-derived metrics, suggesting that its larger memory bus (256-bit vs 128-bit) and double the memory capacity (2 GB vs 1,024 MB) provide a real-world advantage in OpenCL tasks that are sensitive to memory bandwidth and capacity.

Looking at the broader benchmark context, the GTX 675MX also has a Geekbench Metal score of 6,131, a test the GTX 650 Ti does not have a result for. The GTX 650 Ti counters with a Geekbench Vulkan score of 8,229, a test absent for the GTX 675MX. This means each card has one exclusive benchmark, but the shared OpenCL test is the only apples-to-apples comparison. The GTX 675MX wins that test by a large margin, and its average score across all benchmarks (8,427) is higher than the GTX 650 Ti’s (8,053). The GTX 675MX also outperforms its nearest rivals in the data: it is 1.2% ahead of the AMD Radeon R9 M375X (8,325) and 0.1% behind the AMD Radeon 880M (8,436). The GTX 650 Ti, by contrast, is 0.2% ahead of the NVIDIA GeForce GTX 880M (8,040) but 0.2% behind the NVIDIA GeForce GTX 650 Ti Boost (8,067). This places the GTX 675MX in a slightly higher competitive tier.

Specification Differences

The two cards differ on nearly every hardware specification except for architecture, manufacturer, process node, foundry, and API support. The GTX 675MX uses the GK104 chip, while the GTX 650 Ti uses the GK106S. The GTX 675MX has 960 shading units, 80 TMUs, and 32 ROPs; the GTX 650 Ti has 768 shading units, 64 TMUs, and 16 ROPs. The GTX 675MX thus has 192 more shading units, 16 more TMUs, and double the ROPs. Transistor count is 3,540 million versus 2,540 million, and die size is 294 mm² versus 221 mm². The GTX 675MX’s memory is 2 GB on a 256-bit bus with 115.2 GB/s bandwidth; the GTX 650 Ti’s memory is 1,024 MB on a 128-bit bus with 86.40 GB/s. The memory clock is 900 MHz for the GTX 675MX and 1350 MHz for the GTX 650 Ti.

Compute and fill rates go the other way. The GTX 650 Ti has a higher pixel rate (14.85 GPixel/s vs 13.08 GPixel/s), a higher texture rate (59.39 GTexel/s vs 52.32 GTexel/s), and higher FP32 (1,425.4 GFLOPS vs 1,255.7 GFLOPS). The GTX 675MX has a lower TDP of 100 W versus 110 W. The GTX 650 Ti is a single-slot card with a 1x 6-pin power connector, a 300 W suggested PSU, and a length of 145 mm (5.7 inches); the GTX 675MX is an MXM module with no power connectors. The bus interface is MXM-B (3.0) for the GTX 675MX and PCIe 3.0 x16 for the GTX 650 Ti. Display outputs are portable-device-dependent for the GTX 675MX and 2x DVI plus 1x mini-HDMI 1.4a for the GTX 650 Ti. The GTX 650 Ti has a launch MSRP of 149 USD, while the GTX 675MX has none listed. Both are end-of-life products, released within days of each other in 2012.