NVIDIA GeForce GTX 675MX vs NVIDIA GRID K2 Comparison

The GeForce GTX 675MX and GRID K2 share a surprising amount of DNA, both being built on NVIDIA’s Kepler architecture with the GK104 chip. However, they are tuned for entirely different purposes. The GTX 675MX is a mobile gaming part, while the GRID K2 is a datacenter virtualization card. Based strictly on benchmark data, the GTX 675MX holds a narrow but consistent lead in raw compute tests, while the GRID K2 offers double the memory and significantly higher bandwidth. The choice between them hinges on whether you need a portable, self-contained graphics solution or a server-oriented card with a large frame buffer and no display outputs.

The Verdict

Pick the NVIDIA GeForce GTX 675MX if you need a functional mobile GPU with display outputs and the best raw compute scores in this comparison. The data shows it wins both head-to-head benchmarks, with an average benchmark score of 8427 against the GRID K2’s 8080. Its Geekbench Metal score of 6131 is 10.3% higher than the GRID K2’s 5557, and its OpenCL score of 10723 edges out the GRID K2’s 10602 by 1.1%. It also occupies a slightly higher percentile slot at 43, versus 42 for the GRID K2.

Pick the NVIDIA GRID K2 if your priority is memory capacity and bandwidth rather than peak compute. The GRID K2 offers 4 GB of GDDR5 across a 256-bit bus, delivering 160.0 GB/s of bandwidth, compared to the GTX 675MX’s 2 GB and 115.2 GB/s. It also has a much higher shader count, with 1536 shading units versus 960, and a higher texture rate of 95.36 GTexel/s. However, it has no display outputs, so it is only suitable for headless server workloads. The data shows it trails in every benchmark test, but its memory advantage could be decisive for workloads that exceed 2 GB.

Architecture Differences

Both cards are built on the same 28 nm TSMC process with identical GK104 chips, 3,540 million transistors, and a 294 mm² die size. The transistor density is identical at 12.0M / mm². They share the same Kepler architecture, DirectX 12 (11_0) API support, OpenGL 4.6, and Vulkan 1.2.175.

The core configuration differs significantly. The GRID K2 has 1536 shading units and 128 TMUs, while the GTX 675MX has 960 shading units and 80 TMUs. Both have 32 ROPs. This gives the GRID K2 a theoretical pixel rate of 23.84 GPixel/s and texture rate of 95.36 GTexel/s, versus 13.08 GPixel/s and 52.32 GTexel/s for the GTX 675MX. The FP32 compute is also higher on the GRID K2 at 2.289 TFLOPS, compared to 1,255.7 GFLOPS for the GTX 675MX.

Memory is another key divider. The GRID K2 runs its GDDR5 at 1250 MHz (5 Gbps effective) for 160.0 GB/s bandwidth across 4 GB. The GTX 675MX runs its GDDR5 at 900 MHz (3.6 Gbps effective) for 115.2 GB/s across 2 GB. Both use a 256-bit bus. The GRID K2 is a dual-slot PCIe 3.0 x16 card with a 267 mm length and requires a 550 W suggested PSU, plus 1x 6-pin and 1x 8-pin power connectors. The GTX 675MX is an MXM module with no power connectors and is portable-device dependent for display output. The GRID K2 has no outputs at all.

FAQ

Q: Which card has better raw compute performance?

A: The GTX 675MX wins both head-to-head benchmarks. It scores 6131 in Geekbench Metal versus 5557 for the GRID K2, and 10723 in Geekbench OpenCL versus 10602. Despite the GRID K2 having higher theoretical FP32, texture, and pixel rates, the GTX 675MX’s actual benchmark scores are higher.

Q: Can I use the GRID K2 in a desktop PC for gaming?

A: No. The GRID K2 has no display outputs, so it cannot drive a monitor directly. It is designed for headless server environments. The GTX 675MX, by contrast, has portable-device dependent outputs, meaning it relies on the laptop or mobile device to provide display connectivity.

Q: How much memory do they have, and does it matter?

A: The GRID K2 has 4 GB of GDDR5, while the GTX 675MX has 2 GB. The GRID K2 also has higher bandwidth at 160.0 GB/s versus 115.2 GB/s. For workloads that require large frame buffers, such as virtualized desktops, the GRID K2’s memory capacity is a clear advantage.

Q: Which card has a better average benchmark score?

A: The GTX 675MX has an average benchmark score of 8427, placing it in the 43rd percentile of all GPUs. The GRID K2 has an average of 8080, placing it in the 42nd percentile. The GTX 675MX is roughly 4.3% faster on average.

Q: Are they the same chip?

A: Yes, both use the GK104 chip on TSMC’s 28 nm process, with identical transistor counts (3,540 million) and die size (294 mm²). The differences come from how the chip is configured: the GRID K2 has more shaders and TMUs enabled, but runs at lower benchmark efficiency in these specific tests.

Q: What are the power requirements?

A: The GRID K2 has a TDP of 225 W and requires a 550 W suggested PSU with 1x 6-pin and 1x 8-pin power connectors. The GTX 675MX has a TDP of 100 W and uses no power connectors, as it is an MXM module. The GRID K2 is a dual-slot card, 267 mm long.

Specification Differences

| Specification | NVIDIA GeForce GTX 675MX | NVIDIA GRID K2 |

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

| Memory Size | 2 GB | 4 GB |

| Memory Clock | 900 MHz (3.6 Gbps effective) | 1250 MHz (5 Gbps effective) |

| Memory Bandwidth | 115.2 GB/s | 160.0 GB/s |

| Shading Units | 960 | 1536 |

| TMUs | 80 | 128 |

| Pixel Rate | 13.08 GPixel/s | 23.84 GPixel/s |

| Texture Rate | 52.32 GTexel/s | 95.36 GTexel/s |

| FP32 Performance | 1,255.7 GFLOPS | 2.289 TFLOPS |

| TDP | 100 W | 225 W |

| Slot Width | MXM Module | Dual-slot |

| Power Connectors | None | 1x 6-pin + 1x 8-pin |

| Suggested PSU | None | 550 W |

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

| Display Outputs | Portable Device Dependent | No outputs |

| Dimensions | Not specified | 267 mm (10.5 inches) |

| Release Date | 2012-09-30 | 2013-05-10 |

| Launch MSRP | Not available | 5,199 USD |

Head-to-Head Benchmarks

The head-to-head data shows a clean sweep for the GTX 675MX, but the margins tell different stories. In Geekbench Metal, the GTX 675MX scores 6131 against the GRID K2’s 5557, a decisive 10.3% lead. This is the largest performance gap in the comparison. The OpenCL test is much closer: 10723 versus 10602, a mere 1.1% difference. The GTX 675MX wins both, giving it a 2-0 record in head-to-head matches.

These results are notable because the GRID K2 has significantly higher theoretical specs. It has 60% more shading units (1536 vs 960), 60% more TMUs (128 vs 80), and nearly double the FP32 throughput (2.289 TFLOPS vs 1,255.7 GFLOPS). Yet in practice, the GTX 675MX outperforms it. The data suggests that the GRID K2’s higher clocks and core counts do not translate into better real-world compute in these specific Geekbench workloads. The GTX 675MX’s average benchmark score of 8427 also beats the GRID K2’s 8080, a 4.3% difference.

Looking at nearest rivals reinforces the positioning. The GTX 675MX sits within 0.4% of the AMD Radeon 880M and NVIDIA GeForce MX330, and 1.2% ahead of the AMD Radeon R9 M375X. The GRID K2, meanwhile, is within 0.5% of the NVIDIA GeForce GTX 880M and 0.3% of the GTX 650 Ti, and 0.2% faster than the GTX 650 Ti Boost. This places the GTX 675MX in a slightly faster peer group, despite the GRID K2 having more hardware resources.

Where Each One Wins

The GTX 675MX wins every benchmark in this comparison, making it the clear choice for compute-bound tasks. Its 10.3% lead in Geekbench Metal and 1.1% lead in OpenCL indicate that it handles general-purpose GPU workloads more efficiently. This makes it the sensible pick for any application where the card’s score on Geekbench is a proxy for performance, such as mobile workstations or gaming laptops. It also has the advantage of display outputs, meaning it can drive a screen directly, and it consumes far less power at 100 W versus 225 W. Its MXM form factor means it fits in portable devices, though it lacks the GRID K2’s memory capacity.

The GRID K2 wins on memory capacity and bandwidth, doubling the GTX 675MX’s 2 GB to 4 GB and increasing bandwidth from 115.2 GB/s to 160.0 GB/s. This makes it more suitable for workloads that need to hold large datasets on the GPU, such as virtualized graphics or server-side rendering. It also has a higher pixel rate (23.84 vs 13.08 GPixel/s) and texture rate (95.36 vs 52.32 GTexel/s), which the benchmark data does not fully reflect. However, the absence of display outputs and the higher power draw (225 W, 550 W PSU requirement) limit its use to headless server environments. The GRID K2’s launch MSRP is 5,199 USD, a figure that reflects its enterprise positioning. For raw compute per benchmark score, the GTX 675MX is the winner; for memory-heavy server tasks, the GRID K2 offers a distinct hardware advantage.