NVIDIA GeForce 830M vs NVIDIA GeForce 920M Comparison

Head-to-Head Benchmarks

The recorded benchmark data shows a clear and consistent advantage for the NVIDIA GeForce 830M across both available test suites. In the Geekbench OpenCL test, the 830M records a score of 4324, while the GeForce 920M trails at 3725. This translates to a 16.1% delta, a substantial margin for two mobile-class GPUs that share the same TDP envelope.

The gap widens further in the Geekbench Vulkan test. The 830M scores 3590, while the 920M manages only 2849. Here, the delta reaches 26%, making the Vulkan result the largest single victory for either card. The 830M wins both head-to-head benchmarks, securing 2 wins to 0 for the 920M. No benchmark in the database shows the 920M ahead of the 830M, which makes the comparison one-sided from a pure performance standpoint.

It is worth noting how each card sits relative to its nearest rivals, as this provides context beyond the direct comparison. The 830M's average benchmark score is 3957, placing it at the 24th percentile among all GPUs in the database. Its closest competitor, the AMD Radeon R5 M420, scores 3956, a delta of 0%, effectively a tie. The NVIDIA GeForce GT 745M scores 3953, just 0.1% behind, while the NVIDIA Quadro K2000 scores 3964, 0.2% ahead. The AMD Radeon HD 6850 X2 rounds out the group at 3977, 0.5% ahead. The 830M sits in a tightly packed cluster where a few points separate these cards, which means its 16.1% and 26% advantages over the 920M are not typical of its class, they are unusually decisive.

The 920M, by contrast, posts an average score of 3287 and lands at the 20th percentile among all GPUs. Its nearest rivals show a different competitive picture. The NVIDIA GeForce GT 730M scores 3316, which is 0.9% ahead of the 920M. The Intel HD Graphics 530 scores 3332, 1.4% ahead. The Intel HD Graphics P4600 scores 3389, 3% ahead. The only rival the 920M beats is the NVIDIA GeForce GT 640, which scores 3210, a 2.4% deficit. The 920M is not only slower than the 830M, it is also on the wrong side of several comparisons against integrated and older discrete parts.

Where Each One Wins

The benchmark wins are exclusively in favor of the 830M. There is no test category where the 920M emerges ahead. For workloads that stress OpenCL compute, such as general-purpose GPU acceleration, image processing, or physics simulations, the 830M offers a 16.1% performance advantage. For Vulkan-based rendering, which includes modern game engines and compute-heavy graphical effects, the 830M extends its lead to 26%. This suggests that the 830M's advantage grows in APIs that allow lower-level hardware access, likely because of more efficient use of its execution resources.

For the 920M, the recorded data offers no clear win condition. Its only favorable comparison among nearest rivals is against the GeForce GT 640, where it is 2.4% faster. Against the 830M, its best result is a 16.1% deficit in OpenCL, which is still a meaningful gap. Users looking at the 920M would be accepting a consistent performance penalty across both compute and graphics workloads.

The use-case split is therefore straightforward. If the workload is OpenCL-based, the 830M is the stronger choice. If the workload is Vulkan-based, the 830M is an even stronger choice. There is no scenario in the data where the 920M provides a benefit, whether measured by raw score or by relative position against its own rivals. The 920M's lower percentile ranking (20th versus 24th) and lower average score (3287 versus 3957) reinforce this conclusion.

Architecture Differences

The two GPUs come from different NVIDIA architectures, which explains much of the performance gap. The GeForce 830M is built on the GM108 chip using the Maxwell architecture, while the GeForce 920M uses the GK208B chip with the Kepler 2.0 architecture. Both are manufactured on a 28 nm process at TSMC, and both integrate 1,020 million transistors. The die sizes differ, however: the 830M measures 77 mm², while the 920M is larger at 87 mm². This gives the 830M a higher transistor density of 13.2M per mm², compared to 11.7M per mm² for the 920M.

Clock speeds also favor the 830M. The 830M has a base clock of 1082 MHz and a boost clock of 1150 MHz. The 920M runs at a fixed 954 MHz for both base and boost, with no boost headroom. This 128 MHz base clock difference and the boost capability of the 830M contribute to its superior throughput. Memory clocks are identical, with both running at 900 MHz and 1800 Mbps effective, so memory speed does not factor into the performance difference.

The execution resources tell a more nuanced story. The 920M actually has more shading units: 384 versus 256 for the 830M. It also has more texture mapping units (TMUs): 32 versus 16. However, the 830M has a higher pixel rate of 9.200 GPixel/s, while the 920M manages only 7.632 GPixel/s. The texture rate flips in favor of the 920M, which delivers 30.53 GTexel/s against the 830M's 18.40 GTexel/s. Floating-point performance also favors the 920M, which reaches 732.7 GFLOPS compared to 588.8 GFLOPS for the 830M. Despite having lower raw compute numbers, the 830M wins the actual benchmarks, indicating that architectural efficiency in Maxwell outweighs the Kepler 2.0 chip's higher resource counts. This is a case where the newer architecture extracts more useful performance per shader and per clock.

Both cards share the same memory configuration: 2 GB of DDR3 on a 64-bit bus, yielding 14.40 GB/s of bandwidth. Both have 8 ROPs, a 33 W TDP, an IGP slot width, no power connectors, and a PCIe 3.0 x8 bus interface. Display outputs are portable-device dependent for both. The API support differs slightly: the 830M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The 920M supports DirectX 12 (11_0) and OpenGL 4.6, but only Vulkan 1.2.175. The older Vulkan version on the 920M may explain part of its larger deficit in the Vulkan benchmark, as it cannot take advantage of newer API features and optimizations. The 830M also has the benefit of a later release date, first appearing on 2014-03-11, while the 920M launched on 2015-03-12. The 830M's predecessor is the GeForce 700M series, and its successor is the GeForce 900M series. The 920M's predecessor is the GeForce 800M series, and its successor is the GeForce 10 Mobile line.

The Verdict

Based strictly on the recorded data, the NVIDIA GeForce 830M is the superior GPU for every measured workload. It wins both head-to-head benchmarks, posts a higher average score (3957 versus 3287), and ranks higher in the overall GPU percentile distribution (24th versus 20th). Its advantages are 16.1% in OpenCL and 26% in Vulkan, both substantial margins that would be noticeable in real-world compute and graphics tasks.

The 920M does have higher theoretical specs in shading units, TMUs, texture rate, and FP32 throughput, but benchmark results indicate these do not translate into better performance. The Maxwell architecture in the 830M appears to be more efficient, delivering better results with fewer resources and a smaller die. The 920M's fixed 954 MHz clock, lack of boost capability, and older Vulkan support are likely contributors to its underperformance.

For users choosing between these two in a mobile system, the data points unambiguously to the 830M. There is no benchmark category where the 920M takes the lead, and its nearest-rival comparisons show it losing to several other GPUs, including integrated graphics from Intel. The 830M, by contrast, trades blows with a tight cluster of discrete rivals, sitting within 0.5% of the AMD Radeon HD 6850 X2 and within 0.2% of the Quadro K2000. The 830M is a mid-pack performer in its class, while the 920M is a below-average one.

The only scenario where the 920M might be preferable is if a laptop's cooling or driver stack somehow favored it, but the database shows no such advantage. Both share the same 33 W TDP and IGP form factor, so there is no thermal or power reason to pick the 920M. The 920M is not a bad GPU in absolute terms, it is competitive with the GeForce GT 730M and Intel HD Graphics 530, but against the 830M it is clearly the weaker option. The verdict is simple: pick the 830M if performance matters, and only consider the 920M if the 830M is unavailable in a given system.

FAQ

Q: Which GPU wins more benchmarks, the GeForce 830M or the GeForce 920M?

A: The GeForce 830M wins all recorded benchmarks. It takes 2 wins to 0, with a 16.1% advantage in Geekbench OpenCL and a 26% advantage in Geekbench Vulkan.

Q: Does the GeForce 920M have any performance advantage over the GeForce 830M?

A: No. The 920M does not win any benchmark against the 830M. Its only favorable comparison among nearest rivals is a 2.4% lead over the NVIDIA GeForce GT 640.

Q: How do the two GPUs compare in raw compute specifications?

A: The 920M has more resources: 384 shading units versus 256, 32 TMUs versus 16, and higher FP32 at 732.7 GFLOPS versus 588.8 GFLOPS. However, the 830M has a higher pixel rate (9.200 GPixel/s versus 7.632 GPixel/s) and higher clocks (1082 MHz base, 1150 MHz boost versus 954 MHz fixed).

Q: What is the average benchmark score for each GPU?

A: The GeForce 830M has an average benchmark score of 3957, placing it at the 24th percentile. The GeForce 920M has an average score of 3287, placing it at the 20th percentile.

Q: Are there any architectural differences that explain the performance gap?

A: Yes. The 830M uses the Maxwell architecture on a 77 mm² die, while the 920M uses Kepler 2.0 on an 87 mm² die. Both have 1,020 million transistors on a 28 nm process. The 830M also supports Vulkan 1.4, while the 920M only supports Vulkan 1.2.175.

Q: Which GPU should I choose for Vulkan-based applications?

A: The GeForce 830M is the clear choice. It scores 3590 in the Vulkan benchmark, which is 26% higher than the 920M's score of 2849. This is the largest performance gap between the two cards in any test.