GPU Comparison

Head-to-Head Benchmarks

The benchmark data splits this comparison cleanly down the middle, with each GPU winning one discipline. In Geekbench OpenCL, the NVIDIA GeForce 920M scores 3725 against Intel’s 3496, a 6.6% advantage. That is a meaningful but not overwhelming lead, and it puts the 920M in familiar territory: its average benchmark score of 3287 sits just 0.9% below the GeForce GT 730M and 1.4% below Intel HD Graphics 530, while running 2.4% ahead of the GeForce GT 640.

Flip to Geekbench Vulkan, and the tables turn decisively. The Intel UHD Graphics 710 scores 4088 versus the 920M’s 2849, a 30.3% gap. This is the kind of margin that changes how you shop. The Intel part’s average score of 3792 places it within 0.8% of the GeForce GTX 650 and 1.1% of the GeForce MX110, while the 920M trails those same competitors by a wider margin in overall standing. The 920M’s Vulkan result is so far behind that it drags its average down to 3287, even though its OpenCL score is competitive.

What stands out is the direction of the wins. The 920M takes the older, compute-oriented OpenCL workload, while Intel wins the modern graphics API by a landslide. A 30.3% deficit in Vulkan is not a small edge; it is the difference between playable and frustrating in API-heavy titles. The OpenCL win, by contrast, is modest, 6.6% is within the range where driver updates or thermal behavior could shift the outcome on a given day.

The percentile rankings reflect this split. The 920M sits at the 20th percentile of all GPUs, while the UHD 710 sits at the 22nd. That two-point gap is narrow, but the average benchmark scores tell a clearer story: 3792 for Intel versus 3287 for NVIDIA, a 15.4% overall advantage for the newer part. The head-to-head data shows one win each, but the margin of Intel’s Vulkan victory outweighs the margin of NVIDIA’s OpenCL win by a factor of nearly five.

Where Each One Wins

The NVIDIA GeForce 920M wins in OpenCL compute workloads. Its 3725 score against 3496 demonstrates a real, if modest, edge in this legacy API. That matters for older scientific software, some video encoding pipelines, and applications that have not been updated to leverage newer graphics APIs. The 920M also holds its own against its immediate rivals in this test: it trails the GeForce GT 730M by only 0.9% and the Intel HD Graphics 530 by 1.4%, while beating the GeForce GT 640 by 2.4%. If your workload is stuck on OpenCL, the 920M is the safer bet.

The Intel UHD Graphics 710 wins decisively in Vulkan. Its 4088 score versus 2849 is a 30.3% blowout, and it positions the Intel part near the GeForce GTX 650 (within 0.8%) and the GeForce MX110 (within 1.1%). Vulkan is the API used by modern game engines, emulators, and a growing list of productivity tools. For anyone running current titles or Vulkan-based applications, the UHD 710 is in a different performance class. Its 22nd percentile ranking versus the 920M’s 20th also indicates better overall standing in the GPU hierarchy.

The use-case split is straightforward. The 920M is for legacy compute tasks and older DirectX 11-era software, its DirectX support tops out at 12 (11_0), which limits it on the API front. The UHD 710 supports DirectX 12 (12_1), giving it a feature-level advantage for newer games and applications. Intel also leads in raw fill rates: 10.40 GPixel/s versus 7.632 GPixel/s, a 36% advantage in pixel throughput. However, the 920M counters with a texture rate of 30.53 GTexel/s versus Intel’s 10.40 GTexel/s, a nearly 3x advantage that helps in texture-heavy workloads. So if your application is fill-rate bound, Intel wins; if it is texture-bound, NVIDIA wins.

Architecture Differences

These two GPUs come from different eras and different design philosophies. The 920M uses NVIDIA’s Kepler 2.0 architecture on a 28 nm TSMC process, with 1,020 million transistors packed into an 87 mm² die. That yields a transistor density of 11.7 million per mm². The chip is GK208B, a small, power-efficient design from the GeForce 900M generation, released in March 2015. The UHD Graphics 710, by contrast, uses Intel’s Generation 12.2 architecture on a 10 nm Intel process, from the Alder Lake generation released in January 2022. Intel does not list transistor count or die size for this part, but the process advantage is clear: 10 nm versus 28 nm is a generational leap.

The compute resources differ sharply. The 920M packs 384 shading units, 32 texture mapping units, and 8 ROPs. The UHD 710 has 128 shading units, 8 TMUs, and 8 ROPs. NVIDIA’s part has 3x the shaders and 4x the TMUs, which explains its 30.53 GTexel/s texture rate versus Intel’s 10.40 GTexel/s. But Intel compensates with higher clocks: 300 MHz base and 1300 MHz boost, versus NVIDIA’s fixed 954 MHz base and boost. That clock advantage helps Intel reach 10.40 GPixel/s despite having the same ROP count.

Memory is another fundamental split. The 920M has 2 GB of dedicated DDR3 on a 64-bit bus, delivering 14.40 GB/s of bandwidth. The UHD 710 uses system shared memory, with bandwidth listed as "System Dependent." This is a critical difference for performance consistency: dedicated memory gives the 920M predictable bandwidth, while the UHD 710’s performance depends entirely on the host system’s RAM speed and configuration. In a dual-channel high-speed memory setup, Intel can exceed NVIDIA’s bandwidth; in a single-channel low-speed setup, it may fall behind.

Power and interfaces differ too. The 920M has a 33 W TDP and uses PCIe 3.0 x8, while the UHD 710 sips 15 W and connects via Intel’s Ring Bus. Both are IGP-class parts with no dedicated slot width or power connectors. The 920M’s display outputs are "Portable Device Dependent," reflecting its laptop origins, while Intel’s are "Motherboard Dependent," standard for desktop integrated graphics. On the API front, Intel supports Vulkan 1.4 and DirectX 12 (12_1), while NVIDIA supports Vulkan 1.2.175 and DirectX 12 (11_0). Both support OpenGL 4.6.

The Verdict

The data points to a clear choice for most users: the Intel UHD Graphics 710. Its 30.3% Vulkan advantage over the 920M is the single largest margin in this comparison, and Vulkan is the API that matters for modern workloads. Its 22nd percentile ranking and 3792 average score place it in the company of the GeForce GTX 650 and MX110, while the 920M’s 3287 average puts it closer to the GT 730M and HD Graphics 530. For gaming, emulation, or any Vulkan-based application, the UHD 710 is the superior part.

The 920M is only worth choosing if your workload is strictly OpenCL-based and you need its 2 GB of dedicated memory. Its 6.6% OpenCL win is real but modest, and its texture rate advantage (30.53 GTexel/s versus 10.40 GTexel/s) could help in specific compute tasks. However, its DirectX 11_0 feature level and Vulkan 1.2.175 support are dated, and its 20th percentile standing reflects that.

For a builder choosing between these two in a modern system, the UHD 710 is the practical pick. It consumes less power (15 W versus 33 W), supports newer APIs, and wins the benchmark that matters most. The 920M is a legacy part from 2015, end-of-life, and its performance profile is stuck in the past. The only scenario where the 920M makes sense is a very specific OpenCL-heavy, texture-bound workload where dedicated memory is non-negotiable. Otherwise, the data favors Intel.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel UHD Graphics 710 has an average benchmark score of 3792, compared to the NVIDIA GeForce 920M’s 3287. That is a 15.4% overall advantage for Intel.

Q: How big is the Vulkan performance gap?

A: The Intel UHD Graphics 710 scores 4088 in Geekbench Vulkan, while the 920M scores 2849. This is a 30.3% difference in Intel’s favor.

Q: Does the NVIDIA GeForce 920M win any benchmark?

A: Yes, the 920M wins Geekbench OpenCL with a score of 3725 versus Intel’s 3496, a 6.6% advantage.

Q: What are the memory configurations?

A: The 920M has 2 GB of dedicated DDR3 on a 64-bit bus with 14.40 GB/s bandwidth. The UHD 710 uses system shared memory with system-dependent bandwidth.

Q: Which GPU supports newer DirectX features?

A: The Intel UHD Graphics 710 supports DirectX 12 (12_1), while the 920M supports DirectX 12 (11_0). Intel also supports Vulkan 1.4 versus NVIDIA’s Vulkan 1.2.175.

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

A: The 920M’s average score is 0.9% below the GeForce GT 730M and 1.4% below Intel HD Graphics 530, but 2.4% above the GeForce GT 640. The UHD 710’s average is 0.8% below the GeForce GTX 650 and 1.1% below the GeForce MX110, but 1.4% above the GeForce GT 635M and 2% above the Quadro 3000M.

Specification Differences

| Specification | NVIDIA GeForce 920M | Intel UHD Graphics 710 |

|---|---|---|

| Architecture | Kepler 2.0 | Generation 12.2 |

| Process Node | 28 nm | 10 nm |

| Foundry | TSMC | Intel |

| Chip | GK208B | Alder Lake |

| Transistors | 1,020 million | Not specified |

| Die Size | 87 mm² | Not specified |

| Base Clock | 954 MHz | 300 MHz |

| Boost Clock | 954 MHz | 1300 MHz |

| Memory Size | 2 GB | System Shared |

| Memory Type | DDR3 | System Shared |

| Memory Bus Width | 64 bit | System Shared |

| Memory Bandwidth | 14.40 GB/s | System Dependent |

| Shading Units | 384 | 128 |

| TMUs | 32 | 8 |

| ROPs | 8 | 8 |

| Pixel Rate | 7.632 GPixel/s | 10.40 GPixel/s |

| Texture Rate | 30.53 GTexel/s | 10.40 GTexel/s |

| FP32 Performance | 732.7 GFLOPS | 332.8 GFLOPS |

| FP16 Performance | Not specified | 665.6 GFLOPS (2:1) |

| TDP | 33 W | 15 W |

| Bus Interface | PCIe 3.0 x8 | Ring Bus |

| DirectX Support | 12 (11_0) | 12 (12_1) |

| Vulkan Support | 1.2.175 | 1.4 |

| OpenGL Support | 4.6 | 4.6 |

| Release Date | 2015-03-12 | 2022-01-03 |

| Production Status | End-of-life | End-of-life |