NVIDIA GeForce 920MX

The NVIDIA GeForce 920MX is a Maxwell-architecture mobile GPU built on a 28 nm process at TSMC, featuring 1,020 million transistors on a 77 mm² die. With a base clock of 965 MHz and a boost clock of 993 MHz, this end-of-life part from the GeForce 900M generation targets entry-level laptops, delivering modest compute capabilities that place it at the 19th percentile of all GPUs in the benchmark database.

Benchmark Performance

The GeForce 920MX averages 3,354 points across the benchmark suite, with specific scores of 3,964 in Geekbench OpenCL and 2,743 in Geekbench Vulkan. This puts the card in a tight cluster of similarly performing GPUs, where the data shows it is virtually indistinguishable from its nearest competitors in raw compute workloads.

Compared to the Intel HD Graphics P4600, the 920MX trails by just 0.7% — a margin well within run-to-run variance. The integrated Intel solution scores 3,376 on average, making the two effectively equivalent in synthetic benchmarks. However, the 920MX has dedicated VRAM and its own cooling solution, which can matter for sustained workloads even if the raw scores are nearly identical.

Against the NVIDIA GeForce GT 740, the 920MX leads by 1.5%, scoring 3,354 versus 3,304. This is a slim advantage, but it shows the 920MX can edge out a desktop-oriented card from the previous generation in compute tasks. The lead over the older Quadro 2000M is more comfortable at 2.2% — the 920MX beats the 3,429 average of that professional mobile part.

The most relevant comparison is against the GeForce 920M, the non-X variant. Here the 920MX holds a 2.7% advantage, scoring 3,354 versus 3,266. That delta, while small, confirms the "X" suffix delivers a measurable, if modest, performance uplift within the same family. In practical terms, benchmark results indicate the 920MX is a low-end part that trades blows with integrated graphics from the same era, making it suitable for light productivity and older games at low settings rather than demanding 3D work.

Memory Subsystem

The 920MX is equipped with 2 GB of DDR3 memory on a 64-bit bus, yielding a bandwidth of 14.40 GB/s. The memory clock runs at 900 MHz, translating to 1,800 Mbps effective. This configuration is severely bandwidth-limited by modern standards — the 64-bit interface combined with DDR3 rather than GDDR5 means the memory subsystem is the primary bottleneck for this GPU.

For high-resolution workloads, the data suggests the 920MX will struggle. The 2 GB frame buffer is sufficient for 1080p gaming at low to medium settings in titles from the card's era, but the 14.40 GB/s bandwidth constrains texture throughput and fill-rate-heavy scenes. At resolutions above 1080p, the memory bus width and bandwidth become prohibitive, causing frame rates to drop disproportionately compared to cards with wider buses. The pixel rate of 7.944 GPixel/s and texture rate of 23.83 GTexel/s further reinforce that this is a 720p-class GPU at best, with 1080p requiring significant compromises in quality settings.

The 64-bit bus is the same width found on many integrated solutions, which explains why the 920MX's benchmark scores hover near Intel's HD Graphics P4600. The dedicated memory does help with latency and consistency, but the raw bandwidth ceiling is identical. For users considering this card for modern games, the memory subsystem alone makes it a poor choice — even entry-level titles with high-resolution textures will exceed the 2 GB capacity and saturate the bus quickly.

Ray Tracing and Feature Set

The 920MX does not include any dedicated ray tracing cores or tensor cores, as these are absent from the FACT PACK for this GPU. This is expected for a Maxwell-era chip, where real-time ray tracing was not a hardware feature. Consequently, any ray-traced effects in games would have to be handled by the 256 shading units at 508.4 GFLOPS of FP32 compute — a scenario that would produce unusable performance.

API support is functional but limited. The card supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The DirectX 12 support is feature-level 11_0, meaning it cannot take advantage of DirectX 12 Ultimate features like mesh shaders or variable rate shading. Vulkan 1.4 support is notable for a 2016 part, allowing access to modern cross-platform APIs, but the underlying hardware limits the practical benefits. The 24 texture mapping units and 8 raster output units are the same counts found in many entry-level parts, providing baseline functionality for 2D workloads and video playback.

The absence of tensor cores means no DLSS or AI-based upscaling. The absence of RT cores means no hardware-accelerated ray tracing. For users seeking modern features, the 920MX is firmly a legacy part — it can run games that support DirectX 11 or Vulkan, but it will lack the headline features of contemporary GPUs. The PCIe 3.0 x8 bus interface is sufficient for this class of card, as the bandwidth demands of the 920MX are low, and the display outputs being "Portable Device Dependent" indicates the card relies on the laptop's built-in panel connections.

FAQ

Q: How does the GeForce 920MX compare to the GeForce 920M?

A: The 920MX is 2.7% faster in average benchmark scores, scoring 3,354 versus the 920M's 3,266. This is a modest improvement, confirming the "X" variant offers a small but measurable performance uplift.

Q: Can the 920MX handle DirectX 12 games?

A: Yes, it supports DirectX 12, but only at feature level 11_0. This means it can run DirectX 12 titles but cannot use higher-level DirectX 12 features like mesh shaders or ray tracing. Vulkan 1.4 and OpenGL 4.6 are also supported.

Q: What is the memory bandwidth of the 920MX?

A: The card has 14.40 GB/s of memory bandwidth, derived from 2 GB of DDR3 on a 64-bit bus with 1,800 Mbps effective speed. This is a limiting factor for high-resolution gaming.

Q: Does the 920MX support ray tracing?

A: No. The card has no ray tracing cores or tensor cores. Any ray-traced effects must be computed on the 256 shading units, which would result in extremely poor performance.

Q: What is the thermal design power of the 920MX?

A: The TDP is 16 W. This is a low-power part that requires no external power connectors, making it suitable for thin-and-light laptops.

Q: How does the 920MX perform relative to integrated graphics like Intel HD Graphics P4600?

A: It is virtually tied, trailing the Intel HD Graphics P4600 by just 0.7% in average benchmark scores (3,354 versus 3,376). Dedicated VRAM gives it an edge in consistency, but raw performance is similar.

How It Compares

Intel HD Graphics P4600: The 920MX trails this integrated GPU by 0.7% in average benchmarks, making the two effectively identical in raw compute. The key differentiator is the 920MX's dedicated 2 GB VRAM, which allows it to maintain performance where integrated graphics would steal from system memory. The data shows a statistical tie, so the choice comes down to system-level factors rather than GPU capability.

NVIDIA GeForce GT 740: The 920MX leads the GT 740 by 1.5%, scoring 3,354 versus 3,304. This is a modest victory over a desktop card, but the GT 740 likely has better sustained performance due to its larger cooling solution. The benchmark delta is small enough that real-world differences would be negligible in most workloads.

NVIDIA Quadro 2000M: The 920MX beats this professional mobile card by 2.2%, with scores of 3,354 versus 3,429 (the Quadro's higher score means the 920MX has a 2.2% advantage). The Quadro 2000M is older, and the 920MX shows that even entry-level consumer parts can surpass older professional offerings in compute benchmarks, though the Quadro may have had better driver optimization for professional apps.

NVIDIA GeForce 920M: The 920MX holds a 2.7% lead over its non-X sibling, scoring 3,354 versus 3,266. This is the largest delta among the nearest rivals, confirming that the "X" designation in this generation was not merely cosmetic. Users upgrading from a 920M to a 920MX would see a small but consistent performance improvement.

Power and Cooling

The GeForce 920MX has a TDP of 16 W, making it one of the most power-efficient GPUs in its class. This low power draw means the card requires no external power connectors — the "None" designation in the FACT PACK confirms that all power is delivered through the MXM module slot. There is no suggested PSU in the data, which is expected for a mobile part that is not user-upgradeable in most laptops.

The 28 nm process node, while old by modern standards, contributes to the low power envelope. The 1,020 million transistors running at 965 MHz base and 993 MHz boost generate minimal heat, allowing for passive cooling or small fans in thin laptops. The slot width is listed as "MXM Module," indicating this card is designed for MXM-based laptops where the module slots directly into the motherboard.

Because the display outputs are "Portable Device Dependent," users cannot connect external monitors directly to the GPU — they must use the laptop's built-in display outputs. The PCIe 3.0 x8 interface is adequate for the 14.40 GB/s memory bandwidth and 508.4 GFLOPS compute performance, as the card cannot saturate even a x8 link. For builders or upgraders, the 16 W TDP means no additional cooling or power infrastructure is needed beyond what the laptop already provides. The data indicates this is a drop-in replacement for compatible MXM systems, but the end-of-life production status means availability is limited to used or refurbished markets.