GPU Comparison

Intel HD Graphics P4600 and NVIDIA GeForce 920M are both end-of-life graphics solutions aimed at very different segments of the market, yet their benchmark scores place them in close competition. The Intel part is an integrated GPU from the Haswell generation, while the NVIDIA part is a dedicated mobile chip from the Kepler era. The data shows that while the 920M holds a lead in raw compute, the P4600 is far from obsolete, trading blows with its rival in specific workloads.

Head-to-Head Benchmarks

The only direct head-to-head benchmark available is the Geekbench OpenCL test, which measures general-purpose compute performance. In this test, the NVIDIA GeForce 920M scores 3725 points, while the Intel HD Graphics P4600 scores 3389 points. This results in a deltaPct of -9% for the Intel part, indicating that the 920M is approximately 9% faster in this specific workload. While this is a definitive win for the NVIDIA GPU, the margin is relatively modest, suggesting that the P4600 is not completely outclassed in compute tasks.

Looking at the broader context from the nearestRivals data, the P4600’s score of 3389 places it slightly ahead of the Intel HD Graphics 530 (3332) and the NVIDIA GeForce GT 730M (3316), with deltaPct values of 1.7% and 2.2% respectively. This means the P4600 outperforms these two rivals by roughly 2% in the OpenCL benchmark. On the other side, the 920M’s score of 3725 puts it 2.4% ahead of the NVIDIA GeForce GT 640 (3210) and 0.9% behind the GeForce GT 730M (3316). Interestingly, the 920M is also listed as 3% faster than the P4600 in the rival data, which aligns with the direct head-to-head result.

It is important to note that the GeForce 920M also has a Geekbench Vulkan score of 2849, a test that the P4600 does not have a recorded score for. This means that in modern graphics API workloads, the 920M has a measurable advantage, though a direct comparison is impossible due to missing data for the Intel part. The overall wins tally shows the 920M winning 1 out of 1 head-to-head benchmarks, which is a clean sweep, but the narrow margins in OpenCL suggest that the P4600 is a capable competitor in compute-heavy scenarios.

Where Each One Wins

The NVIDIA GeForce 920M is the clear winner in the single head-to-head benchmark, taking the Geekbench OpenCL test with a score of 3725 versus 3389 for the Intel HD Graphics P4600. This indicates that the 920M is better suited for general-purpose compute tasks, such as OpenCL-accelerated applications, where its higher shading unit count and dedicated memory provide an advantage. The 920M also has a Vulkan score of 2849, which means it can handle Vulkan-based workloads, a feature that the P4600 lacks a recorded score for, giving it an edge in modern graphics APIs.

On the other hand, the Intel HD Graphics P4600 wins in the context of its nearest rivals. It outperforms the NVIDIA GeForce GT 730M by 2.2% and the Intel HD Graphics 530 by 1.7% in the OpenCL benchmark, with scores of 3389 versus 3316 and 3332 respectively. This suggests that despite being an older integrated solution, the P4600 is more capable than some dedicated mobile GPUs from the same era. For users stuck with the P4600, the data shows it is not a slouch, particularly when compared to similar-tier parts.

The 920M’s advantage is further highlighted by its memory configuration. With 2 GB of dedicated DDR3 memory and a 64-bit bus, it offers a bandwidth of 14.40 GB/s, whereas the P4600 relies on system-shared memory with bandwidth that is described as "System Dependent." This dedicated memory is a significant factor in the 920M’s favor, especially in texture-heavy or memory-intensive workloads, as it does not have to contend with the system RAM for bandwidth.

Architecture Differences

The two GPUs are built on fundamentally different architectures. The Intel HD Graphics P4600 is based on the Haswell GT2 chip, which uses the Generation 7.5 architecture and is fabricated on a 22 nm process node at Intel’s foundry. In contrast, the NVIDIA GeForce 920M uses the GK208B chip, which is based on the Kepler 2.0 architecture and is manufactured on a 28 nm process node at TSMC. This means the Intel part has a smaller transistor size, but the NVIDIA part has a higher transistor count, with 1,020 million transistors packed into an 87 mm² die, resulting in a density of 11.7M / mm².

The core configurations differ significantly. The P4600 has 160 shading units, 20 texture mapping units (TMUs), and only 2 raster output units (ROPs). The 920M, on the other hand, has 384 shading units, 32 TMUs, and 8 ROPs. This gives the NVIDIA part 2.4 times the shading units and 4 times the ROPs, which explains its higher pixel rate of 7.632 GPixel/s versus 2.400 GPixel/s for the Intel part. The texture rate also favors the 920M, at 30.53 GTexel/s compared to 24.00 GTexel/s, and the FP32 performance is nearly double, at 732.7 GFLOPS versus 384.0 GFLOPS.

Clock speeds play a role too. The P4600 has a base clock of 350 MHz and a boost clock of 1200 MHz, while the 920M has a fixed clock of 954 MHz for both base and boost. The 920M’s memory runs at 900 MHz, or 1800 Mbps effective, but the P4600’s memory is system-shared with no dedicated speed. The TDP also differs drastically, with the P4600 rated at 84 W and the 920M at just 33 W, making the NVIDIA part far more power-efficient despite its higher performance.

FAQ

Q: Which GPU has a higher Geekbench OpenCL score?

A: The NVIDIA GeForce 920M has a higher score of 3725, compared to the Intel HD Graphics P4600’s 3389, resulting in a 9% difference in favor of the NVIDIA part.

Q: How does the Intel HD Graphics P4600 compare to the NVIDIA GeForce GT 730M?

A: The P4600 scores 3389 in OpenCL, which is 2.2% higher than the GT 730M’s score of 3316, indicating the Intel part is slightly faster in this test.

Q: Does the GeForce 920M support Vulkan?

A: Yes, the 920M has a Geekbench Vulkan score of 2849, while the P4600 has no recorded Vulkan score, giving the NVIDIA part an advantage in Vulkan-based applications.

Q: What is the memory configuration of the GeForce 920M?

A: The 920M comes with 2 GB of DDR3 memory on a 64-bit bus, providing a bandwidth of 14.40 GB/s, whereas the P4600 uses system-shared memory with bandwidth dependent on the system.

Q: What are the TDP ratings for these two GPUs?

A: The Intel HD Graphics P4600 has a TDP of 84 W, while the NVIDIA GeForce 920M has a TDP of 33 W, making the NVIDIA part significantly more power-efficient.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce 920M has 384 shading units, which is 2.4 times the 160 shading units found on the Intel HD Graphics P4600.

Specification Differences

| Specification | Intel HD Graphics P4600 | NVIDIA GeForce 920M |

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

| Architecture | Generation 7.5 | Kepler 2.0 |

| Process Node | 22 nm | 28 nm |

| Foundry | Intel | TSMC |

| Transistors | Not listed | 1,020 million |

| Die Size | Not listed | 87 mm² |

| Transistor Density | Not listed | 11.7M / mm² |

| Base Clock | 350 MHz | 954 MHz |

| Boost Clock | 1200 MHz | 954 MHz |

| Memory Size | System Shared | 2 GB |

| Memory Type | System Shared | DDR3 |

| Memory Bus Width | System Shared | 64 bit |

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

| Shading Units | 160 | 384 |

| TMUs | 20 | 32 |

| ROPs | 2 | 8 |

| Pixel Rate | 2.400 GPixel/s | 7.632 GPixel/s |

| Texture Rate | 24.00 GTexel/s | 30.53 GTexel/s |

| FP32 Performance | 384.0 GFLOPS | 732.7 GFLOPS |

| TDP | 84 W | 33 W |

| Bus Interface | Ring Bus | PCIe 3.0 x8 |

| Power Connectors | Not listed | None |

| Display Outputs | Motherboard Dependent | Portable Device Dependent |

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

| OpenGL Support | 4.3 | 4.6 |

| Vulkan Support | 1.0 | 1.2.175 |

| Release Date | 2013-05-31 | 2015-03-12 |

| Predecessor | Not listed | GeForce 800M |

| Successor | Not listed | GeForce 10 Mobile |

| Geekbench OpenCL Score | 3389 | 3725 |

| Geekbench Vulkan Score | Not listed | 2849 |

| Avg Benchmark Score | 3389 | 3287 |

| Percentile vs All GPUs | 20 | 20 |