GPU Comparison

# FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 670MX leads with an average benchmark score of 5721, compared to 5317 for the GeForce 930A. This places the GTX 670MX in the 33rd percentile of all GPUs, while the 930A sits in the 31st percentile.

Q: How do the two compare in OpenCL performance?

A: The GTX 670MX scores 6125 in Geekbench OpenCL, which is 13.2% higher than the 930A’s 5317. This is the only head-to-head benchmark available, and the GTX 670MX wins it decisively.

Q: What are the nearest rivals for each card, and how do they stack up?

A: The 930A’s closest rival is the GeForce 840M, which scores 5322 (0.1% higher), followed by the GTX 980M at 5308 (0.2% lower) and the 940M at 5284 (0.6% lower). For the GTX 670MX, the Intel Iris Pro P6300 scores 5712 (0.2% lower), the GTX 550 Ti scores 5731 (0.2% higher), and the Quadro M500M scores 5604 (2.1% lower).

Q: Which card has more shading units and texture mapping units?

A: The GTX 670MX has 960 shading units and 80 TMUs, while the 930A has only 384 shading units and 24 TMUs. The GTX 670MX also has 24 ROPs versus 8 on the 930A.

Q: How do memory specifications differ between the two?

A: The GTX 670MX uses 3 GB of GDDR5 memory on a 192-bit bus, delivering 67.20 GB/s bandwidth. The 930A has 2 GB of DDR3 memory on a 64-bit bus, providing only 16.02 GB/s bandwidth.

Q: What is the power consumption difference?

A: The GTX 670MX has a TDP of 75 W, while the 930A consumes just 33 W. Both use no external power connectors, and both are end-of-life products.

# Where Each One Wins

The data clearly favors the GTX 670MX in raw compute and memory throughput. It wins the only head-to-head benchmark (Geekbench OpenCL) with a 13.2% margin, and its average benchmark score of 5721 is 7.6% higher than the 930A’s 5317. The GTX 670MX also excels in memory-intensive workloads due to its 67.20 GB/s bandwidth, more than four times the 930A’s 16.02 GB/s, and its larger 3 GB frame buffer.

However, the 930A is not without its strengths. Its 33 W TDP makes it far more power-efficient, and its Maxwell architecture supports Vulkan 1.4, whereas the GTX 670MX’s Kepler architecture is limited to Vulkan 1.2.175. For portable devices where battery life and thermal headroom are critical, the 930A’s lower power draw and IGP slot width (integrated graphics package) give it an edge in thin-and-light laptops.

In terms of raw performance per watt, the 930A delivers 161.1 GFLOPS per watt (722.7 GFLOPS / 33 W), while the GTX 670MX manages 15.4 GFLOPS per watt (1,153.9 GFLOPS / 75 W). The 930A is over 10 times more efficient in this metric, making it the clear choice for power-constrained environments.

# Architecture Differences

The two GPUs come from different NVIDIA architectures. The 930A is built on Maxwell (chip GM108), while the GTX 670MX uses Kepler (chip GK104). Both are manufactured on a 28 nm process at TSMC, but the transistor counts differ dramatically: the GTX 670MX packs 3,540 million transistors on a 294 mm² die, while the 930A has just 1,020 million transistors on a 77 mm² die. This results in similar transistor densities (13.2M/mm² for the 930A, 12.0M/mm² for the GTX 670MX), but the GTX 670MX’s much larger die enables significantly more compute resources.

The GTX 670MX has 960 shading units, 80 TMUs, and 24 ROPs, versus 384 shading units, 24 TMUs, and 8 ROPs on the 930A. This 2.5x difference in shading units and 3x difference in ROPs explains why the GTX 670MX achieves a pixel rate of 12.02 GPixel/s and a texture rate of 48.08 GTexel/s, compared to 7.528 GPixel/s and 22.58 GTexel/s on the 930A.

In terms of API support, both cards support DirectX 12 (11_0) and OpenGL 4.6. However, the 930A’s Maxwell architecture supports Vulkan 1.4, while the GTX 670MX’s Kepler architecture is capped at Vulkan 1.2.175. This gives the 930A a modern-API advantage despite its older release date (March 2015 vs. September 2012 for the GTX 670MX).

Memory architecture also differs fundamentally: the 930A uses 2 GB of DDR3 with a 64-bit bus, while the GTX 670MX uses 3 GB of GDDR5 with a 192-bit bus. The GDDR5 memory runs at 700 MHz (2.8 Gbps effective) compared to 1001 MHz (2 Gbps effective) for DDR3, but the wider bus is the dominant factor in the GTX 670MX’s 4.2x higher bandwidth.

# Specification Differences

The table below highlights only the fields where the two GPUs differ:

| Specification | GeForce 930A | GeForce GTX 670MX |

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

| Chip | GM108 | GK104 |

| Architecture | Maxwell | Kepler |

| Generation | GeForce 900A | GeForce 600M |

| Transistors | 1,020 million | 3,540 million |

| Die Size | 77 mm² | 294 mm² |

| Transistor Density | 13.2M / mm² | 12.0M / mm² |

| Base Clock | 928 MHz | 601 MHz |

| Boost Clock | 941 MHz | 601 MHz |

| Memory Clock | 1001 MHz (2 Gbps) | 700 MHz (2.8 Gbps) |

| Memory Size | 2 GB | 3 GB |

| Memory Type | DDR3 | GDDR5 |

| Memory Bus | 64 bit | 192 bit |

| Memory Bandwidth | 16.02 GB/s | 67.20 GB/s |

| Shading Units | 384 | 960 |

| TMUs | 24 | 80 |

| ROPs | 8 | 24 |

| Pixel Rate | 7.528 GPixel/s | 12.02 GPixel/s |

| Texture Rate | 22.58 GTexel/s | 48.08 GTexel/s |

| FP32 | 722.7 GFLOPS | 1,153.9 GFLOPS |

| TDP | 33 W | 75 W |

| Slot Width | IGP | null |

| Bus Interface | PCIe 3.0 x8 | PCIe 3.0 x16 |

| Vulkan | 1.4 | 1.2.175 |

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

| Predecessor | GeForce 800A | GeForce 500M |

| Successor | null | GeForce 700M |

| Avg Benchmark Score | 5317 | 5721 |

| Percentile | 31 | 33 |

# Head-to-Head Benchmarks

The only direct benchmark comparison available is Geekbench OpenCL, where the GTX 670MX scores 6125 against the 930A’s 5317. This represents a 13.2% advantage for the GTX 670MX, which is substantial given the 930A’s higher clock speeds (928 MHz base vs. 601 MHz base). The GTX 670MX compensates for its lower clocks with 2.5x more shading units and 3x more ROPs, resulting in 59.7% higher FP32 throughput (1,153.9 GFLOPS vs. 722.7 GFLOPS).

In synthetic compute terms, the GTX 670MX’s pixel rate advantage is 59.6% (12.02 GPixel/s vs. 7.528 GPixel/s), and its texture rate advantage is 112.9% (48.08 GTexel/s vs. 22.58 GTexel/s). Memory bandwidth is where the gap is most pronounced: the GTX 670MX’s 67.20 GB/s is 319.5% higher than the 930A’s 16.02 GB/s. This makes the GTX 670MX vastly superior for large textures, high-resolution frame buffers, and compute workloads that require frequent memory access.

Looking at nearest rivals for context, the 930A scores nearly identically to the GeForce 840M (5322, only 0.1% higher) and the GTX 980M (5308, 0.2% lower), indicating that its performance is tightly clustered around that performance tier. The GTX 670MX, meanwhile, sits between the Iris Pro P6300 (5712, 0.2% lower) and the GTX 550 Ti (5731, 0.2% higher), showing it is competitive with desktop-class GPUs from the same era.

# The Verdict

The data points to a clear winner for raw performance: the NVIDIA GeForce GTX 670MX. It wins the only head-to-head benchmark by 13.2%, has a 7.6% higher average benchmark score, and dominates in every compute metric, FP32 (1,153.9 GFLOPS vs. 722.7 GFLOPS), pixel rate (12.02 vs. 7.528 GPixel/s), texture rate (48.08 vs. 22.58 GTexel/s), and memory bandwidth (67.20 vs. 16.02 GB/s). Its 3 GB GDDR5 frame buffer is also more future-proof for larger textures and higher resolutions.

However, the 930A is the better choice for specific use cases. Its 33 W TDP is less than half the GTX 670MX’s 75 W, making it ideal for ultra-portable devices where battery life and thermals are primary concerns. The 930A also supports Vulkan 1.4, which is a newer API version than the GTX 670MX’s 1.2.175, potentially offering better compatibility with modern game engines that leverage newer Vulkan features.

For users who prioritize maximum frame rates, higher resolution textures, or GPU compute performance, the GTX 670MX is the only rational choice. Its 4.2x memory bandwidth advantage alone will prevent stuttering and texture pop-in in memory-heavy scenarios. For users who need a low-power GPU for basic gaming, media playback, or light productivity, the 930A’s efficiency and modern API support make it a sensible pick.

The GTX 670MX’s release date (September 2012) predates the 930A by over two years, yet it still outperforms it significantly. This reflects the GTX 670MX’s positioning as a higher-tier mobile GPU with a much larger die and more compute resources. The 930A, by contrast, is a budget-oriented chip that maximizes efficiency over performance. Ultimately, the benchmark data shows no scenario where the 930A beats the GTX 670MX in raw speed, the only advantages are power consumption and Vulkan version support.