NVIDIA GeForce 940M vs NVIDIA GeForce GTX 970M Comparison

Head-to-Head Benchmarks

The recorded data shows a decisive performance gap between the NVIDIA GeForce 940M and the NVIDIA GeForce GTX 970M. In the two shared benchmark disciplines, the GTX 970M wins both, and the margins are substantial.

In Geekbench OpenCL, the 940M scores 6018, while the GTX 970M reaches 18946. That is a 68.2% delta in favor of the GTX 970M. In practical terms, this means the GTX 970M delivers roughly three times the compute throughput in this workload. The 940M is not close here; it trails by a wide margin that will be felt in any OpenCL-accelerated application, from rendering to compute tasks.

The Vulkan results tell a similar story, though the gap is even larger. The 940M scores 4549, while the GTX 970M posts 18292. The delta is 75.1% in favor of the GTX 970M. That is more than a fourfold difference in raw Vulkan performance. For any game or application that leverages Vulkan, the GTX 970M is in a different class entirely.

The win count is 2-0 for the GTX 970M. The 940M has zero head-to-head victories in the recorded benchmarks. It is notably the 940M does have a higher base clock (1020 MHz versus 924 MHz) and a higher boost clock (1098 MHz versus 1038 MHz), but the architectural and memory advantages of the GTX 970M completely overwhelm the clock speed deficit. Clock speed alone cannot compensate for the massive differences in shading units, memory bandwidth, and raw compute throughput.

The average benchmark score in the database reinforces this split. The 940M averages 5284 across its recorded tests, while the GTX 970M averages 4628. That average is skewed by the fact that the GTX 970M has a much larger benchmark suite, including several Passmark tests where it scores relatively low (for example, 28 in Passmark DirectX 10 and 24 in Passmark DirectX 12). The GTX 970M's average is dragged down by those older or less demanding workloads, but its peak scores in OpenCL and Vulkan are far beyond anything the 940M can produce.

The percentile rankings tell a nuanced story. The 940M sits at the 31st percentile among all GPUs, while the GTX 970M sits at the 27th percentile. That seems counterintuitive given the massive benchmark wins for the GTX 970M, but it reflects the composition of the database and the fact that the GTX 970M's average is depressed by its wide range of test scores. The nearest rivals for each card also differ considerably. The 940M's closest competitors include the GeForce GTX 980M (average score 5308, delta -0.4%), the GeForce 930A (5317, delta -0.6%), and the GeForce 840M (5322, delta -0.7%). All of those are within a fraction of a percent of the 940M's average. The GTX 970M, by contrast, sits alongside the Quadro M3000M (4621, delta 0.2%), the Radeon R5 M320 (4657, delta -0.6%), and the Radeon RX 9060 XT 16 GB (4657, delta -0.6%). This shows that the 940M is grouped with other entry-level mobile GPUs, while the GTX 970M is grouped with more capable workstation and mid-range parts, even if its average score is lower due to test selection.

Architecture Differences

The two GPUs share a manufacturer and a generation, but they are built on different chips and different revisions of the Maxwell architecture. The 940M uses the GM107 chip, which is based on the original Maxwell architecture. The GTX 970M uses the GM204 chip, which is based on Maxwell 2.0. Both are fabricated on a 28 nm process at TSMC, and both have no dedicated ray tracing or tensor cores. The transistor counts, however, are far apart. The 940M has 1,870 million transistors on a die size of 148 mm², giving a transistor density of 12.6M per mm². The GTX 970M has 5,200 million transistors on a 398 mm² die, for a density of 13.1M per mm². That is nearly three times the transistor budget for the GTX 970M, which directly translates into more execution resources.

The core configurations are vastly different. The 940M has 512 shading units, 32 texture mapping units, and 16 raster output units. The GTX 970M has 1280 shading units, 80 TMUs, and 48 ROPs. That is 2.5 times the shading units, 2.5 times the texture units, and three times the ROPs. The pixel rate for the 940M is 17.57 GPixel/s, while the GTX 970M achieves 49.82 GPixel/s. The texture rate for the 940M is 35.14 GTexel/s, versus 83.04 GTexel/s for the GTX 970M. FP32 compute is 1,124.4 GFLOPS for the 940M, while the GTX 970M delivers 2.657 TFLOPS. Every throughput metric favors the GTX 970M by a wide margin.

Memory is another major divider. The 940M comes with 2 GB of DDR3 on a 64-bit bus, yielding a bandwidth of 14.40 GB/s. The GTX 970M has 6 GB of GDDR5 on a 192-bit bus, delivering 120.3 GB/s. That is more than eight times the memory bandwidth. Memory clock speeds also differ: the 940M runs its memory at 900 MHz with an effective rate of 1800 Mbps, while the GTX 970M runs at 1253 MHz with an effective rate of 5 Gbps. The bus width difference alone (64-bit versus 192-bit) would be decisive, but the GDDR5 versus DDR3 gap makes the GTX 970M's memory subsystem categorically superior.

The API support is similar but not identical. Both support OpenGL 4.6 and Vulkan 1.4. The difference is in DirectX: the 940M supports DirectX 12 (11_0), while the GTX 970M supports DirectX 12 (12_1). That means the GTX 970M can handle the higher feature level of DirectX 12_1, which includes advanced rendering features not available on the 940M. Both are MXM modules with MXM-B (3.0) interfaces, and both have no power connectors, relying on the MXM slot for power delivery. The 940M has a TDP of 75 W, while the GTX 970M has no recorded TDP in the database. Display outputs are portable-device dependent for both.

FAQ

Q: Which GPU is faster in compute workloads?

A: The GTX 970M is significantly faster. In Geekbench OpenCL, it scores 18946 versus 6018 for the 940M, a 68.2% advantage. In Geekbench Vulkan, the GTX 970M scores 18292 versus 4549, a 75.1% advantage.

Q: Do both GPUs support the same DirectX version?

A: No. The 940M supports DirectX 12 (11_0), while the GTX 970M supports DirectX 12 (12_1). The GTX 970M supports a higher feature level, which matters for games that use DirectX 12_1 features.

Q: Is the memory bandwidth difference meaningful?

A: Yes, it is enormous. The 940M has 14.40 GB/s of bandwidth from 2 GB of DDR3 on a 64-bit bus. The GTX 970M has 120.3 GB/s from 6 GB of GDDR5 on a 192-bit bus. That is over eight times the bandwidth, which directly impacts texture-heavy scenes and higher resolutions.

Q: Are these GPUs from the same generation?

A: Yes, both are part of the GeForce 900M generation. The 940M was released on 2015-03-12, while the GTX 970M was released earlier on 2014-10-06. Both use the Maxwell architecture, but the 940M uses the original Maxwell while the GTX 970M uses Maxwell 2.0.

Q: How do these cards compare to their nearest rivals?

A: The 940M's average score of 5284 is within 1% of the GeForce GTX 980M (5308), GeForce 930A (5317), GeForce 840M (5322), and GeForce GTX 760M (5236). The GTX 970M's average of 4628 is within 1% of the Quadro M3000M (4621), Radeon R5 M320 (4657), Radeon RX 9060 XT 16 GB (4657), and Radeon R5 M230 (4577).

Q: Which GPU has more shading units?

A: The GTX 970M has 1280 shading units, while the 940M has 512. The GTX 970M also has 80 TMUs and 48 ROPs, versus 32 TMUs and 16 ROPs for the 940M.

Specification Differences

The following fields differ between the two GPUs:

• Chip: GM107 (940M) versus GM204 (970M)
• Architecture: Maxwell (940M) versus Maxwell 2.0 (970M)
• Transistors: 1,870 million (940M) versus 5,200 million (970M)
• Die Size: 148 mm² (940M) versus 398 mm² (970M)
• Transistor Density: 12.6M / mm² (940M) versus 13.1M / mm² (970M)
• Base Clock: 1020 MHz (940M) versus 924 MHz (970M)
• Boost Clock: 1098 MHz (940M) versus 1038 MHz (970M)
• Memory Clock: 900 MHz, 1800 Mbps effective (940M) versus 1253 MHz, 5 Gbps effective (970M)
• Memory Size: 2 GB (940M) versus 6 GB (970M)
• Memory Type: DDR3 (940M) versus GDDR5 (970M)
• Memory Bus Width: 64 bit (940M) versus 192 bit (970M)
• Memory Bandwidth: 14.40 GB/s (940M) versus 120.3 GB/s (970M)
• Shading Units: 512 (940M) versus 1280 (970M)
• TMUs: 32 (940M) versus 80 (970M)
• ROPs: 16 (940M) versus 48 (970M)
• Pixel Rate: 17.57 GPixel/s (940M) versus 49.82 GPixel/s (970M)
• Texture Rate: 35.14 GTexel/s (940M) versus 83.04 GTexel/s (970M)
• FP32: 1,124.4 GFLOPS (940M) versus 2.657 TFLOPS (970M)
• TDP: 75 W (940M) versus not recorded (970M)
• DirectX Support: 12 (11_0) (940M) versus 12 (12_1) (970M)
• Release Date: 2015-03-12 (940M) versus 2014-10-06 (970M)

Fields that are the same include manufacturer (NVIDIA), generation (GeForce 900M), process node (28 nm), foundry (TSMC), slot width (MXM Module), power connectors (None), bus interface (MXM-B 3.0), display outputs (Portable Device Dependent), OpenGL version (4.6), and Vulkan version (1.4). Both have no RT cores, no tensor cores, and no FP16 data recorded. Both are end-of-life production status.

The Verdict

The data does not support the 940M in any head-to-head comparison. The GTX 970M wins both recorded benchmarks, and the margins are overwhelming. The 68.2% delta in OpenCL and the 75.1% delta in Vulkan are not minor gaps; they represent entirely different performance tiers. The 940M is an entry-level mobile part, grouped with the GeForce 840M and GeForce 930A in the database. The GTX 970M, despite its lower average score due to a broader test suite, is grouped with workstation and mid-range parts like the Quadro M3000M.

If you are choosing between these two for a laptop, the GTX 970M is the only rational pick for any GPU-intensive task. Its 1280 shading units, 6 GB of GDDR5, and 120.3 GB/s of bandwidth put it in a different class. The 940M's advantages are limited to a higher base and boost clock, which are irrelevant given the massive architectural gap. The 940M does have a lower TDP (75 W versus no recorded figure for the GTX 970M), which might matter for battery life in a thin-and-light chassis, but that is a power consideration, not a performance one.

The percentile rankings might confuse a casual reader: the 940M is at the 31st percentile, and the GTX 970M is at the 27th. But that is a function of the GTX 970M's average being dragged down by low scores in older Passmark tests (28 in DirectX 10, 24 in DirectX 12, 99 in DirectX 9). Its peak performance, as shown in OpenCL and Vulkan, is far above the 940M's capability. The average score field should not be read as a ranking of real-world gaming performance; the head-to-head results are the more reliable indicator.

Where Each One Wins

The 940M has no benchmark wins in the recorded data. It does not win in OpenCL, and it does not win in Vulkan. Its only comparative advantages are its higher clock speeds (1020 MHz base versus 924 MHz, 1098 MHz boost versus 1038 MHz) and its lower TDP of 75 W. For a user who prioritizes battery life and low heat output in a portable device, the 940M might be the less demanding option. It also has a smaller die (148 mm² versus 398 mm²) and fewer transistors, which suggests lower manufacturing cost, though pricing data is not available.

The GTX 970M wins in every measurable performance category. It dominates in compute (OpenCL and Vulkan), in memory bandwidth (120.3 GB/s versus 14.40 GB/s), in pixel rate (49.82 GPixel/s versus 17.57 GPixel/s), in texture rate (83.04 GTexel/s versus 35.14 GTexel/s), and in FP32 throughput (2.657 TFLOPS versus 1,124.4 GFLOPS). It also has 6 GB of VRAM versus 2 GB, which matters for modern titles with large texture packs. The DirectX 12_1 feature level support gives it access to newer rendering paths that the 940M cannot use.

For gaming, the GTX 970M is the clear choice. For any Vulkan-based game, the 75.1% lead means the 940M will struggle to maintain playable frame rates where the GTX 970M has headroom. For compute tasks like OpenCL-accelerated rendering or data processing, the 68.2% lead is just as decisive. The only scenario where the 940M makes sense is a system where power draw is the primary constraint and the workload is light enough that the GTX 970M's extra capability is unnecessary. Even then, the 940M's performance is so far behind that it should only be considered for the most basic 2D or legacy workloads. The data is unambiguous: the GTX 970M is the superior GPU in every benchmark that matters.