NVIDIA GeForce 940M vs NVIDIA GeForce GTX 670MX Comparison

Head-to-Head Benchmarks

The recorded data places the NVIDIA GeForce GTX 670MX ahead of the NVIDIA GeForce 940M in both available benchmark tests. The most significant gap appears in the Geekbench Vulkan test, where the GTX 670MX scores 5316 against the 940M's 4549. That is a 16.9% advantage, a substantial margin that suggests the older Kepler-based part handles Vulkan workloads noticeably better, likely due to its wider memory bus and higher raw shading throughput.

In the Geekbench OpenCL test, the difference narrows considerably. The GTX 670MX scores 6125, while the 940M posts 6018, a delta of only 1.8%. This near-tie indicates that in compute tasks that scale well with clock speed and modern architecture, the Maxwell-based 940M can almost keep pace with its older sibling. The 940M's higher base clock of 1020 MHz versus 601 MHz helps close the gap despite having fewer shading units.

The overall average benchmark score reinforces this split. The GTX 670MX averages 5721 across all recorded tests, placing it in the 33rd percentile of all GPUs. The 940M averages 5284, sitting in the 31st percentile. While the percentile difference is small, the raw score gap of 437 points (roughly 8.3%) is meaningful. Interestingly, the GTX 670MX's closest rival, the Intel Iris Pro Graphics P6300, scores 5712, a mere 0.2% behind. The 940M's nearest rival, the NVIDIA GeForce GTX 980M, scores 5308, which is 0.4% behind the 940M. This context shows that both cards sit in a crowded mid-range field, but the GTX 670MX edges out its competition more consistently.

What stands out is the Vulkan result. A 16.9% lead in Vulkan is not trivial. It suggests that the GTX 670MX's 192-bit memory interface and 67.20 GB/s of bandwidth provide a structural advantage for graphics-heavy workloads, whereas the 940M's 64-bit bus and 14.40 GB/s bandwidth become a bottleneck despite its faster clocks. The OpenCL result, however, shows that pure compute performance is more balanced, with the 940M's Maxwell architecture squeezing out efficiency from its 512 shading units.

FAQ

Q: Which GPU wins in OpenCL performance?

A: The NVIDIA GeForce GTX 670MX wins the Geekbench OpenCL test with a score of 6125, compared to the 940M's 6018. The margin is 1.8%, making this a close contest where architecture differences almost cancel out.

Q: How large is the Vulkan performance gap?

A: The GTX 670MX leads by 16.9% in Geekbench Vulkan, scoring 5316 versus 4549 for the 940M. This is the largest single-test advantage in the head-to-head data.

Q: Which GPU has a higher average benchmark score?

A: The GTX 670MX averages 5721 across all recorded benchmarks, while the 940M averages 5284. The GTX 670MX also holds a better percentile ranking at 33rd versus 31st for the 940M.

Q: How does the GTX 670MX compare to its nearest rival?

A: The GTX 670MX's closest competitor is the Intel Iris Pro Graphics P6300, which scores 5712, a 0.2% difference. It also sits near the NVIDIA GeForce GTX 550 Ti (5731, 0.2% behind) and the AMD Radeon HD 8790M (5691, 0.5% ahead).

Q: What is the 940M's nearest rival and how close is it?

A: The 940M's nearest rival is the NVIDIA GeForce GTX 980M, which averages 5308, placing it 0.4% behind the 940M. Other close competitors include the NVIDIA GeForce 930A (5317, 0.6% behind) and the NVIDIA GeForce 840M (5322, 0.7% behind).

Q: Does the 940M win any benchmark categories?

A: No. The recorded data shows the GTX 670MX winning both the OpenCL and Vulkan tests. The 940M has zero wins in the head-to-head comparison.

Architecture Differences

The two GPUs come from different architectural generations, and the data reflects a clear design philosophy shift. The GTX 670MX is built on Kepler (chip GK104), fabricated on a 28 nm process at TSMC. It packs 3,540 million transistors onto a die size of 294 mm², yielding a transistor density of 12.0M per mm². This is a large, power-hungry design by modern standards, but it offers substantial raw resources: 960 shading units, 80 texture mapping units, and 24 ROPs.

The 940M, in contrast, uses Maxwell (chip GM107), also on a 28 nm process at TSMC, but with only 1,870 million transistors on a 148 mm² die. That gives it a slightly higher transistor density of 12.6M per mm², indicating a more efficient layout. Its execution resources are halved: 512 shading units, 32 TMUs, and 16 ROPs. Despite fewer cores, the 940M compensates with much higher clocks: 1020 MHz base and 1098 MHz boost, versus the GTX 670MX's flat 601 MHz (no boost clock separation).

The memory subsystems diverge sharply. The GTX 670MX uses 3 GB of GDDR5 on a 192-bit bus, delivering 67.20 GB/s of bandwidth. The 940M uses 2 GB of DDR3 on a 64-bit bus, yielding just 14.40 GB/s. This 4.7x bandwidth difference is the single largest architectural gap and explains why the GTX 670MX dominates in Vulkan. The 940M's memory clock of 900 MHz (1800 Mbps effective) cannot compensate for the narrow bus.

Both GPUs support DirectX 12 (11_0) and OpenGL 4.6, but their Vulkan support differs: the GTX 670MX reports Vulkan 1.2.175, while the 940M reports Vulkan 1.4. Interestingly, the newer Vulkan version on the 940M does not translate into better Vulkan performance in the recorded tests, suggesting that hardware resources matter more than API version support.

Specification Differences

The following fields differ between the two GPUs, based solely on the recorded data:

• Chip: GK104 (GTX 670MX) versus GM107 (940M)
• Architecture: Kepler versus Maxwell
• Generation: GeForce 600M versus GeForce 900M
• Transistors: 3,540 million versus 1,870 million
• Die Size: 294 mm² versus 148 mm²
• Transistor Density: 12.0M / mm² versus 12.6M / mm²
• Base Clock: 601 MHz versus 1020 MHz
• Boost Clock: 601 MHz versus 1098 MHz
• Memory Clock: 700 MHz (2.8 Gbps effective) versus 900 MHz (1800 Mbps effective)
• Memory Size: 3 GB versus 2 GB
• Memory Type: GDDR5 versus DDR3
• Memory Bus Width: 192 bit versus 64 bit
• Memory Bandwidth: 67.20 GB/s versus 14.40 GB/s
• Shading Units: 960 versus 512
• TMUs: 80 versus 32
• ROPs: 24 versus 16
• Pixel Rate: 12.02 GPixel/s versus 17.57 GPixel/s
• Texture Rate: 48.08 GTexel/s versus 35.14 GTexel/s
• FP32: 1,153.9 GFLOPS versus 1,124.4 GFLOPS
• Slot Width: null versus MXM Module
• Bus Interface: PCIe 3.0 x16 versus MXM-B (3.0)
• Vulkan Version: 1.2.175 versus 1.4
• Release Date: 2012-09-30 versus 2015-03-12
• Predecessor: GeForce 500M versus GeForce 800M
• Successor: GeForce 700M versus GeForce 10 Mobile

Both GPUs share a 28 nm process, TSMC foundry, 75 W TDP, no power connectors, and portable-device-dependent display outputs.

The Verdict

The data points to a clear winner for graphics-heavy tasks: the NVIDIA GeForce GTX 670MX. It wins both head-to-head benchmarks, with a decisive 16.9% margin in Vulkan. Its 192-bit GDDR5 memory configuration provides 67.20 GB/s of bandwidth, which is essential for high-resolution textures and complex scenes. The 940M's 14.40 GB/s DDR3 bandwidth is a severe limitation that no amount of clock speed can overcome in memory-intensive workloads.

However, the verdict is not one-sided. The 940M posts a nearly identical OpenCL score (6018 versus 6125, just 1.8% behind), which indicates that its Maxwell architecture is more efficient per transistor. Its higher pixel rate (17.57 GPixel/s versus 12.02 GPixel/s) also suggests better fill-rate performance in certain rendering scenarios. For users prioritizing compute tasks that are not bandwidth-bound, the 940M is a respectable alternative.

The GTX 670MX also holds a better position in the broader GPU landscape. Its 33rd percentile and 5721 average score place it above the 940M's 31st percentile and 5284 average. The GTX 670MX's nearest rival, the Intel Iris Pro Graphics P6300, is nearly identical in score (5712), while the 940M's nearest rival, the GTX 980M, trails by 0.4%. This suggests the GTX 670MX is the stronger all-rounder in the direct comparison.

Where Each One Wins

NVIDIA GeForce GTX 670MX wins in:

• Vulkan graphics workloads, with a 16.9% lead (5316 versus 4549), likely due to its superior memory bandwidth.
• OpenCL compute, though narrowly, with a 1.8% advantage (6125 versus 6018).
• Overall average benchmark score, 5721 versus 5284.
• Memory bandwidth, texture fill rate (48.08 GTexel/s versus 35.14 GTexel/s), and FP32 throughput (1,153.9 GFLOPS versus 1,124.4 GFLOPS).
• Having more shading units (960 versus 512), TMUs (80 versus 32), and ROPs (24 versus 16).

NVIDIA GeForce 940M wins in:

• Pixel fill rate, posting 17.57 GPixel/s versus the GTX 670MX's 12.02 GPixel/s, indicating faster rasterization per clock.
• Clock speeds, with a base of 1020 MHz and boost of 1098 MHz, versus the GTX 670MX's fixed 601 MHz.
• Transistor density, at 12.6M / mm² versus 12.0M / mm², showing better use of silicon area.
• Newer Vulkan API support (1.4 versus 1.2.175), though this does not translate into a benchmark win.
• A later release date (2015 versus 2012), meaning longer software optimization potential.

For gaming and 3D rendering, the GTX 670MX is the safer choice based on its Vulkan dominance. For lightweight workloads or systems where pixel throughput matters more than memory bandwidth, the 940M's higher clocks and pixel rate offer an edge. The data suggests that the GTX 670MX is the more capable GPU overall, but the 940M is not without its strengths in specific metrics.