NVIDIA GeForce GTX 675MX vs NVIDIA GeForce GTX 950M Comparison

# NVIDIA GeForce GTX 675MX vs NVIDIA GeForce GTX 950M

The GTX 675MX and GTX 950M represent two distinct approaches to mobile graphics from the same vendor, separated by roughly two and a half years of development. The older Kepler-based GTX 675MX edges out the newer Maxwell-based GTX 950M in the only shared benchmark, scoring 10% higher in Geekbench OpenCL, yet the GTX 950M counters with a lower power draw, more memory, and newer API support. The data paints a picture of two GPUs optimized for different priorities: raw compute throughput versus efficiency and modern feature support.

Where Each One Wins

The GTX 675MX wins the only direct head-to-head comparison available. In Geekbench OpenCL, it scores 10,723 against the GTX 950M's 9,745, a 10% advantage that reflects its larger memory bus and higher shading-unit count. This GPU is built for compute-heavy workloads that can utilize its wider 256-bit memory interface and 960 shading units. Its average benchmark score of 8,427 across all tests places it in the 43rd percentile of all GPUs, slightly ahead of the GTX 950M's 8,135 average and 42nd percentile ranking.

The GTX 950M, despite losing the OpenCL test, wins on architectural efficiency and feature completeness. It draws 75 W versus the GTX 675MX's 100 W, a 25% reduction in power consumption that makes it better suited for thinner laptops with smaller cooling solutions. It also offers double the memory capacity at 4 GB versus 2 GB, which benefits texture-heavy games and applications that exceed the 2 GB frame buffer limit. The GTX 950M supports Vulkan 1.4, while the GTX 675MX is limited to Vulkan 1.2.175, giving the newer card access to newer graphics API features.

In the broader benchmark landscape, the GTX 675MX's nearest rivals include the AMD Radeon 880M (average score 8,436, only 0.1% higher) and the NVIDIA GeForce MX330 (8,458, 0.4% higher). The GTX 950M's nearest competitors are the AMD Radeon R9 M360 (8,129, 0.1% lower) and the NVIDIA GeForce 945M (8,099, 0.4% lower). These close margins indicate both cards sit in a tightly contested mid-range segment where small architectural differences translate into minor performance deltas.

Architecture Differences

The GTX 675MX uses the GK104 chip built on TSMC's 28 nm process, the same silicon that powered desktop Kepler parts. It packs 3,540 million transistors into a 294 mm² die, yielding a transistor density of 12.0 million per square millimeter. The GTX 950M, by contrast, uses the GM107 chip from the Maxwell family, also on 28 nm but with only 1,870 million transistors on a 148 mm² die, achieving a slightly higher density of 12.6 million per square millimeter. The Maxwell design achieves better performance per watt through architectural improvements rather than raw transistor count.

The compute resources differ substantially. The GTX 675MX fields 960 shading units, 80 texture mapping units, and 32 ROPs, while the GTX 950M has 640 shading units, 40 TMUs, and 16 ROPs. Despite having 33% fewer shading units, the GTX 950M achieves higher peak FP32 throughput at 1,438.7 GFLOPS versus 1,255.7 GFLOPS for the GTX 675MX, thanks to its higher clock speeds (993 MHz base, 1,124 MHz boost versus no listed base/boost for the 675MX). The Maxwell architecture's improved instruction efficiency allows it to extract more work per shader per clock.

Memory configurations diverge sharply. The GTX 675MX uses 2 GB of GDDR5 on a 256-bit bus, delivering 115.2 GB/s of bandwidth. The GTX 950M opts for 4 GB of DDR3 on a 128-bit bus, yielding only 28.80 GB/s. This is a 75% reduction in memory bandwidth, a severe bottleneck for the GTX 950M in bandwidth-sensitive scenarios. The GTX 675MX's GDDR5 memory runs at 900 MHz with 3.6 Gbps effective data rate, while the GTX 950M's DDR3 runs at 900 MHz with 1,800 Mbps effective. The older card's wider bus more than compensates for the same base memory clock.

Other differences include the bus interface: the GTX 675MX uses MXM-B (3.0) as a modular MXM board, while the GTX 950M is an integrated IGP with a PCIe 3.0 x8 interface. The GTX 675MX belongs to the GeForce 600M generation with a release date of September 2012, while the GTX 950M is from the GeForce 900M generation launched in March 2015. Both support DirectX 12 (11_0) and OpenGL 4.6, but the GTX 950M's Vulkan 1.4 support exceeds the GTX 675MX's Vulkan 1.2.175.

Head-to-Head Benchmarks

The only direct benchmark comparison in the data is Geekbench OpenCL, where the GTX 675MX scores 10,723 against the GTX 950M's 9,745. This 10% lead is notable given the GTX 950M's higher peak FP32 rating. The explanation lies in memory bandwidth: the GTX 675MX's 115.2 GB/s versus the GTX 950M's 28.80 GB/s means the older card can feed its compute units far more effectively in OpenCL workloads that stress memory access patterns. The 4x bandwidth advantage more than offsets the GTX 950M's superior theoretical compute throughput.

Both cards also have additional benchmark results that cannot be directly compared. The GTX 675MX scores 6,131 in Geekbench Metal, a test the GTX 950M does not have a result for. The GTX 950M scores 6,525 in Geekbench Vulkan, a test absent from the GTX 675MX's results. These are useful for context: the GTX 675MX's Metal score suggests its compute capabilities translate well to Apple's graphics API, while the GTX 950M's Vulkan score demonstrates its newer API support in action.

The aggregate benchmark picture shows the GTX 675MX with an average score of 8,427 across two tests, while the GTX 950M averages 8,135 across two different tests. The GTX 675MX's average is 3.6% higher, but the differing test suites make this a rough comparison. The percentile rankings tell a similar story: 43rd percentile for the GTX 675MX versus 42nd for the GTX 950M, placing both in the lower-middle tier of all GPUs.

FAQ

Q: Which GPU is faster in OpenCL compute?

A: The GTX 675MX scores 10,723 in Geekbench OpenCL, 10% higher than the GTX 950M's 9,745. The older card's 256-bit GDDR5 memory bus provides 115.2 GB/s bandwidth versus 28.80 GB/s on the GTX 950M.

Q: Does the GTX 950M have better API support?

A: Yes. The GTX 950M supports Vulkan 1.4, while the GTX 675MX is limited to Vulkan 1.2.175. Both support DirectX 12 (11_0) and OpenGL 4.6.

Q: How do their memory configurations differ?

A: The GTX 675MX has 2 GB of GDDR5 on a 256-bit bus, while the GTX 950M has 4 GB of DDR3 on a 128-bit bus. The GTX 675MX offers 115.2 GB/s bandwidth; the GTX 950M offers 28.80 GB/s.

Q: Which card consumes less power?

A: The GTX 950M has a 75 W TDP, 25% lower than the GTX 675MX's 100 W TDP. This makes the GTX 950M more suitable for compact laptop designs.

Q: What are their average benchmark scores?

A: The GTX 675MX averages 8,427 across its two benchmark results, placing it in the 43rd percentile. The GTX 950M averages 8,135 across its two results, placing it in the 42nd percentile.

Q: Which card has more shading units?

A: The GTX 675MX has 960 shading units, 80 TMUs, and 32 ROPs. The GTX 950M has 640 shading units, 40 TMUs, and 16 ROPs.

The Verdict

The data supports a clear split: choose the GTX 675MX for compute-heavy applications that benefit from high memory bandwidth. Its 10% OpenCL advantage, 115.2 GB/s bandwidth, and larger 256-bit bus make it the stronger performer in raw throughput scenarios. The 960 shading units and 80 TMUs provide ample parallel processing resources, and its 2 GB GDDR5 frame buffer is adequate for most workloads at its performance level.

Choose the GTX 950M for modern gaming and efficiency-conscious laptop designs. Its 75 W TDP allows for thinner, lighter machines with smaller batteries, and its 4 GB memory capacity handles higher-resolution textures better than the 2 GB limit of the GTX 675MX. Vulkan 1.4 support ensures compatibility with newer graphics APIs, and its higher peak FP32 of 1,438.7 GFLOPS suggests better raw compute efficiency despite the narrower memory bus.

The GTX 675MX's average benchmark score of 8,427 and 43rd percentile ranking slightly edge out the GTX 950M's 8,135 and 42nd percentile, but the GTX 950M's nearest rivals include the NVIDIA GeForce GTX 980 (0.4% higher average), indicating it competes in a tougher performance tier. Both cards are end-of-life products, but the GTX 950M's Maxwell architecture is the more refined design, delivering competitive compute with significantly lower power draw and double the memory capacity.

Specification Differences

| Specification | GTX 675MX | GTX 950M |

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

| Chip | GK104 | GM107 |

| Architecture | Kepler | Maxwell |

| Generation | GeForce 600M | GeForce 900M |

| Process Node | 28 nm | 28 nm |

| Transistors | 3,540 million | 1,870 million |

| Die Size | 294 mm² | 148 mm² |

| Transistor Density | 12.0M / mm² | 12.6M / mm² |

| Base Clock | — | 993 MHz |

| Boost Clock | — | 1,124 MHz |

| Memory Clock | 900 MHz / 3.6 Gbps effective | 900 MHz / 1,800 Mbps effective |

| Memory Size | 2 GB | 4 GB |

| Memory Type | GDDR5 | DDR3 |

| Memory Bus Width | 256 bit | 128 bit |

| Memory Bandwidth | 115.2 GB/s | 28.80 GB/s |

| Shading Units | 960 | 640 |

| TMUs | 80 | 40 |

| ROPs | 32 | 16 |

| Pixel Rate | 13.08 GPixel/s | 17.98 GPixel/s |

| Texture Rate | 52.32 GTexel/s | 44.96 GTexel/s |

| FP32 Performance | 1,255.7 GFLOPS | 1,438.7 GFLOPS |

| TDP | 100 W | 75 W |

| Slot Width | MXM Module | IGP |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x8 |

| Vulkan Version | 1.2.175 | 1.4 |

| Release Date | 2012-09-30 | 2015-03-12 |

| Predecessor | GeForce 500M | GeForce 800M |

| Successor | GeForce 700M | GeForce 10 Mobile |