GPU Comparison

Intel HD Graphics 630 and NVIDIA GeForce GT 635M represent two very different approaches to mobile graphics: one is a modern integrated solution embedded in a processor, the other is a legacy discrete part from a previous era. The benchmark data shows a narrow but decisive edge for the older NVIDIA part. The GT 635M wins the only shared test, the Geekbench OpenCL benchmark, with a score of 3740 against the HD 630’s 3587, a margin of 4.1%. While the average benchmark score for the HD 630 is higher at 4075, this figure is skewed by the inclusion of two additional tests (Geekbench Metal and Geekbench Vulkan) that the GT 635M does not support. In the only direct comparison available, the NVIDIA card comes out ahead.

Where Each One Wins

The data indicates a clear split based on the software environment. The NVIDIA GeForce GT 635M wins the sole head-to-head benchmark, the Geekbench OpenCL test. This is a general-purpose compute benchmark that stresses the GPU’s raw parallel processing capabilities. The GT 635M’s score of 3740 in this test places it slightly above the HD 630’s 3587, making it the better choice for applications that rely on OpenCL acceleration.

The Intel HD Graphics 630, however, demonstrates its strengths in other areas. Its average benchmark score of 4075 is higher than the GT 635M’s 3740, a difference driven by its performance in the Geekbench Metal and Geekbench Vulkan tests. The HD 630 scores 5099 in Metal and 3540 in Vulkan. The GT 635M has no scores for these APIs, indicating a lack of support or optimization. Therefore, the HD 630 is the only option for workloads that depend on these modern, cross-platform graphics APIs, particularly Vulkan, which is increasingly common in newer games and emulators. The HD 630 also holds a percentile ranking of 24th among all GPUs, compared to the GT 635M’s 22nd, suggesting that in the broader database of results, the Intel part performs better across its suite of supported tests.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The Intel HD Graphics 630 is based on the Generation 9.5 architecture, specifically the Kaby Lake GT2 chip, and is manufactured on a 14 nm++ process. In contrast, the NVIDIA GeForce GT 635M uses the older Fermi architecture with the GF108 chip, built on a 40 nm process at TSMC. This generational gap is significant, with the Intel part being much more modern.

In terms of raw compute resources, the HD 630 has a distinct advantage in shader count. It features 192 shading units, 24 texture mapping units (TMUs), and 3 render output units (ROPs). The GT 635M has fewer resources, with 96 shading units, 16 TMUs, and 4 ROPs. This means the HD 630 has double the shading units and 1.5 times the TMUs of the NVIDIA part. The GT 635M, however, has one more ROP. The HD 630’s higher theoretical pixel rate of 3.000 GPixel/s and texture rate of 24.00 GTexel/s further reflect its larger resource pool, compared to the GT 635M’s 1.900 GPixel/s and 7.600 GTexel/s.

Memory configuration is a major differentiator. The GT 635M has a dedicated 2 GB of DDR3 memory on a 128-bit bus, providing a fixed bandwidth of 28.80 GB/s. The HD 630, being an integrated GPU, uses System Shared memory, with its bandwidth described as System Dependent. This means the NVIDIA part has a dedicated, predictable memory pool, while the Intel part must compete with the CPU for system memory bandwidth.

The NVIDIA chip also has a higher transistor count at 585 million on a 116 mm² die, while the Intel part’s transistor count is not listed. The GT 635M’s TDP is 35 W, which is higher than the HD 630’s 15 W. Both use an IGP slot width, but the GT 635M connects via PCIe 2.0 x16, whereas the HD 630 uses a Ring Bus interface.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test. Here, the NVIDIA GeForce GT 635M scores 3740 points, while the Intel HD Graphics 630 scores 3587 points. This gives the NVIDIA part a 4.1% advantage. The deltaPct is reported as -4.1, indicating that the HD 630 is 4.1% slower than the GT 635M in this specific workload. This is a narrow victory, but it is a win nonetheless for the older, less complex GPU.

Looking at the nearest rivals provides context. The HD 630’s average score of 4075 puts it in the company of the AMD Radeon RX 9060 XT 8 GB (4093, -0.4%), the NVIDIA Quadro K2100M (4151, -1.8%), the NVIDIA GeForce GT 755M (4033, 1%), and the AMD FirePro M4150 (4013, 1.5%). The GT 635M’s average score of 3740 is close to the NVIDIA Quadro 3000M (3718, 0.6%), the NVIDIA GeForce GT 740M (3717, 0.6%), the NVIDIA GeForce 825M (3694, 1.2%), and the Intel UHD Graphics 710 (3792, -1.4%). This shows that in the OpenCL test, the GT 635M is performing at a level typical of mid-range mobile parts from its era, while the HD 630’s overall average is influenced by its stronger performance in other API tests.

The Verdict

The choice between these two GPUs depends entirely on the intended software load. For applications that specifically use OpenCL for general-purpose computing, the NVIDIA GeForce GT 635M is the stronger performer, with a 4.1% lead in the Geekbench OpenCL test. Its dedicated memory and higher TDP suggest it was designed for sustained compute tasks, even if its architecture is older.

For users who need support for Vulkan or Metal APIs, the Intel HD Graphics 630 is the only viable option, as the GT 635M has no recorded scores for these tests. The HD 630’s higher average benchmark score of 4075 and superior percentile ranking (24th vs 22nd) indicate that it is a more versatile part overall, despite losing the direct OpenCL comparison. Its lower TDP of 15 W also makes it a more power-efficient choice. Therefore, the GT 635M is the pick for legacy OpenCL-focused workloads, while the HD 630 is the better all-around solution for modern API compatibility and general efficiency.

FAQ

Q: Which GPU is faster in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce GT 635M is faster, scoring 3740 points compared to the Intel HD Graphics 630’s 3587, a 4.1% difference.

Q: Does the Intel HD Graphics 630 support Vulkan?

A: Yes, the HD 630 supports Vulkan 1.3 and has a Geekbench Vulkan score of 3540. The GT 635M has no Vulkan support listed.

Q: What is the difference in memory configuration?

A: The GT 635M has 2 GB of dedicated DDR3 memory on a 128-bit bus with 28.80 GB/s bandwidth. The HD 630 uses System Shared memory with bandwidth described as System Dependent.

Q: How do their average benchmark scores compare?

A: The Intel HD Graphics 630 has a higher average benchmark score of 4075, while the NVIDIA GeForce GT 635M has an average score of 3740.

Q: Which GPU has a higher TDP?

A: The NVIDIA GeForce GT 635M has a TDP of 35 W, which is higher than the Intel HD Graphics 630’s 15 W.

Q: What are the shading unit counts?

A: The Intel HD Graphics 630 has 192 shading units, while the NVIDIA GeForce GT 635M has 96 shading units.

Specification Differences

| Specification | Intel HD Graphics 630 | NVIDIA GeForce GT 635M |

| :--- | :--- | :--- |

| Architecture | Generation 9.5 | Fermi |

| Process Node | 14 nm++ | 40 nm |

| Foundry | Intel | TSMC |

| Transistors | Not listed | 585 million |

| Die Size | Not listed | 116 mm² |

| Shading Units | 192 | 96 |

| TMUs | 24 | 16 |

| ROPs | 3 | 4 |

| Pixel Rate | 3.000 GPixel/s | 1.900 GPixel/s |

| Texture Rate | 24.00 GTexel/s | 7.600 GTexel/s |

| FP32 Performance | 384.0 GFLOPS | 182.4 GFLOPS |

| FP16 Performance | 768.0 GFLOPS (2:1) | Not listed |

| Memory Size | System Shared | 2 GB |

| Memory Type | System Shared | DDR3 |

| Memory Bus Width | System Shared | 128 bit |

| Memory Bandwidth | System Dependent | 28.80 GB/s |

| TDP | 15 W | 35 W |

| Bus Interface | Ring Bus | PCIe 2.0 x16 |

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

| Vulkan Support | 1.3 | Not listed |

| Release Date | 2016-08-29 | 2012-03-21 |