GPU Comparison

NVIDIA GeForce MX110 and NVIDIA GeForce 920MX are both end-of-life, 28nm Maxwell parts built on the same GM108S chip, but the MX110 is the stronger performer across every measured benchmark. In the head-to-head data, the MX110 wins both Geekbench tests, leading by 4.6% in OpenCL and by a substantial 14.3% in Vulkan. The 920MX, however, consumes significantly less power, drawing just 16W compared to the MX110's 30W, making it the more efficient choice for ultra-thin, passively cooled laptops. The MX110 is the definitive pick for users who prioritize raw compute performance, while the 920MX makes sense only when power draw is the absolute limiting factor, a trade-off where the MX110's speed advantage is clear but not overwhelming.

The Verdict

The data is unambiguous: the NVIDIA GeForce MX110 is the faster GPU in this pairing, and it wins in every benchmark category recorded. Its average benchmark score of 3834 places it at the 23rd percentile of all GPUs, while the 920MX sits at 3528 and the 21st percentile. That 8.7% gap in average score is driven by wins in both OpenCL (4255 vs 4068) and Vulkan (3413 vs 2987) workloads. For anyone choosing between these two based purely on performance, the MX110 is the correct answer.

The 920MX is not without its rationale. Its 16W TDP is nearly half the MX110's 30W, a difference that can be decisive in fanless designs or systems with minimal cooling headroom. The 920MX also offers a wider PCIe interface (x8 vs x4), which could theoretically reduce data transfer bottlenecks in some scenarios. But the benchmark data shows no such benefit materializing, the MX110 wins outright in both tests. The verdict, then, is straightforward: pick the MX110 for compute capability; pick the 920MX only if power constraints are non-negotiable and you accept a measurable performance penalty, particularly in Vulkan workloads where the deficit is 14.3%.

Architecture Differences

Both GPUs share the same fundamental architecture: Maxwell, built on a 28nm process at TSMC, with an identical transistor count of 1,020 million and a die size of 77 mm². The transistor density is consequently the same at 13.2M per square millimeter. The chip is the GM108S in both cases, and both support the same API set: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. Neither has ray tracing or tensor cores.

The key architectural divergence lies in the memory subsystem and the texture pipeline. The MX110 uses 2GB of GDDR5 memory on a 64-bit bus, delivering 40.10 GB/s of bandwidth at an effective 5 Gbps. The 920MX also has 2GB on a 64-bit bus, but it uses DDR3 memory at 1800 Mbps effective, yielding just 14.40 GB/s, a 64% reduction in raw bandwidth. That is the single most impactful difference, and it explains the MX110's Vulkan dominance.

The texture unit configuration also differs. The MX110 has 16 TMUs, while the 920MX has 24. This gives the 920MX a higher texture rate of 23.83 GTexel/s versus the MX110's 16.10 GTexel/s. However, the MX110 compensates with higher clocks: its base is 978 MHz with a boost of 1006 MHz, versus the 920MX's 965 MHz base and 993 MHz boost. The shading unit count is identical at 256, and both have 8 ROPs. The pixel rate slightly favors the MX110 at 8.048 GPixel/s versus 7.944 GPixel/s. The MX110's FP32 output is 515.1 GFLOPS, marginally ahead of the 920MX's 508.4 GFLOPS.

Where Each One Wins

The MX110 wins in every single benchmark recorded in the fact pack. In Geekbench OpenCL, it scores 4255 against the 920MX's 4068, a 4.6% advantage. In Geekbench Vulkan, the gap widens dramatically: 3413 versus 2987, a 14.3% lead. This pattern suggests the MX110's GDDR5 memory bandwidth is particularly beneficial in Vulkan workloads, which often stress memory throughput more heavily than OpenCL.

The 920MX's sole domain of superiority is power efficiency. At 16W, it draws 46.7% less power than the MX110's 30W. This makes it the better choice for systems where thermal and power budgets are extremely tight, for instance, a slim ultrabook with no active cooling. The 920MX also has a wider bus interface (PCIe 3.0 x8 vs x4), which is a theoretical advantage for data streaming, though the benchmark results do not reflect any practical benefit. The 920MX's higher texture rate (23.83 GTexel/s) could theoretically help in texture-bound scenarios, but no benchmark data supports this, and its drastically lower memory bandwidth likely negates any such advantage.

FAQ

Q: Which GPU has higher benchmark scores?

A: The NVIDIA GeForce MX110 wins both recorded benchmarks. It scores 4255 in Geekbench OpenCL versus 4068 for the 920MX, and 3413 in Geekbench Vulkan versus 2987. Its average benchmark score is 3834 compared to 3528.

Q: Are these GPUs based on the same chip?

A: Yes. Both use the GM108S chip on a 28nm process with 1,020 million transistors and a 77 mm² die size. They share the Maxwell architecture and support identical APIs: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4.

Q: What is the biggest performance difference between them?

A: The largest gap is in Geekbench Vulkan, where the MX110 leads by 14.3%. The OpenCL gap is smaller at 4.6%. This is likely attributable to the MX110's GDDR5 memory providing 40.10 GB/s bandwidth versus the 920MX's DDR3 at 14.40 GB/s.

Q: Is the 920MX more power-efficient?

A: Yes, significantly. The 920MX has a TDP of 16W, while the MX110 is rated at 30W. This means the 920MX draws nearly half the power, making it more suitable for low-power or fanless designs.

Q: Do they have the same memory configuration?

A: Both have 2GB of memory on a 64-bit bus, but the types differ. The MX110 uses GDDR5 at 5 Gbps effective (1253 MHz), while the 920MX uses DDR3 at 1800 Mbps effective (900 MHz). This yields 40.10 GB/s for the MX110 versus 14.40 GB/s for the 920MX.

Q: How do these compare to their nearest rivals?

A: The MX110's nearest rival is the NVIDIA GeForce GTX 650, which scores 3823, a 0.3% delta. The 920MX sits near the NVIDIA GeForce GT 545 (3594, -1.8%) and the NVIDIA RTX 5000 Mobile Ada Generation (3596, -1.9%). The MX110 outperforms the 920MX's rivals by a wider margin than the 920MX does.

Head-to-Head Benchmarks

The head-to-head data shows a clean sweep for the MX110. In Geekbench OpenCL, the MX110 posts 4255 against the 920MX's 4068. That 187-point difference translates to a 4.6% win. This is a modest but consistent lead, suggesting the MX110's higher clocks (978 MHz base vs 965 MHz) and superior memory bandwidth give it an edge in compute-heavy OpenCL tasks.

The Geekbench Vulkan test reveals a much more pronounced separation. Here, the MX110 scores 3413, while the 920MX manages only 2987. The 426-point gap is a 14.3% margin, a decisive victory. Vulkan workloads are known for their low overhead and direct hardware access, which tends to amplify memory bandwidth differences. The MX110's GDDR5 memory, offering 40.10 GB/s versus 14.40 GB/s, is the clear explanatory factor. In practical terms, this means the MX110 will handle modern Vulkan-based games and applications with noticeably more headroom, while the 920MX may struggle to maintain smooth frame delivery.

The average benchmark scores corroborate this trend. The MX110 averages 3834 across all tests, which is 8.7% higher than the 920MX's 3528. The MX110 also ranks slightly higher in the overall GPU percentile distribution (23rd vs 21st). Interestingly, the MX110's rivals in the nearestRivals list include the GTX 650 (3823) and Intel UHD Graphics 710 (3792), both of which it edges out. The 920MX, by contrast, sits just above the Quadro 2000M (3434, +2.7%) and just below the GT 545 (3594, -1.8%). This shows that while both GPUs occupy the low-end segment, the MX110 is positioned at the top of that tier, whereas the 920MX is closer to the middle.

Specification Differences

The two GPUs differ in several key specifications, with the memory subsystem being the most consequential.

Clocks: The MX110 runs at a base of 978 MHz and boosts to 1006 MHz. The 920MX is slower, with a base of 965 MHz and a boost of 993 MHz. Memory clocks also differ: the MX110's GDDR5 runs at 1253 MHz (5 Gbps effective), while the 920MX's DDR3 runs at 900 MHz (1800 Mbps effective).

Memory: Both have 2GB on a 64-bit bus, but the MX110's GDDR5 delivers 40.10 GB/s bandwidth, versus the 920MX's DDR3 at 14.40 GB/s. This is a 2.8x difference in favor of the MX110.

Texture Units: The 920MX has 24 TMUs, while the MX110 has 16. This gives the 920MX a higher texture rate of 23.83 GTexel/s versus 16.10 GTexel/s for the MX110.

Power: The MX110 is rated at 30W TDP, whereas the 920MX draws just 16W. The MX110 is an IGP (integrated graphics processor) with no power connectors, while the 920MX is an MXM module, also with no power connectors.

Bus Interface: The MX110 uses PCIe 3.0 x4, while the 920MX uses PCIe 3.0 x8, a wider interface for the latter.

Release Date and Successors: The 920MX launched earlier, on 2016-03-24, and lists its predecessor as GeForce 800M and successor as GeForce 10 Mobile. The MX110 launched later, on 2017-11-16, with no predecessor or successor listed. Both are marked as end-of-life.

Compute Rates: The MX110 has a pixel rate of 8.048 GPixel/s and FP32 of 515.1 GFLOPS. The 920MX has a pixel rate of 7.944 GPixel/s and FP32 of 508.4 GFLOPS. Both have 256 shading units and 8 ROPs. Neither supports FP16 natively, and neither has RT or tensor cores.