Intel HD Graphics 630 vs NVIDIA GeForce 940MX Comparison

Where Each One Wins

The benchmark data splits cleanly between these two mobile graphics solutions. The NVIDIA GeForce 940MX wins both recorded head-to-head tests, taking the Geekbench OpenCL and Geekbench Vulkan comparisons without dropping a single match. The Intel HD Graphics 630, by contrast, does not register a single win in the two directly comparable workloads. However, the Intel part does have one unique result in its favor: a Geekbench Metal score of 5099, a test the NVIDIA GPU did not run. That Metal figure is the only benchmark where the Intel iGPU posts a higher number than anything in the 940MX's record, though it is not a directly comparable test.

The 940MX's wins are not narrow. In Geekbench OpenCL, the NVIDIA part scores 4939 against Intel's 3587, a 37.7% advantage. In Geekbench Vulkan, the gap is 4749 versus 3540, a 34.2% lead. These are substantial margins, not edge cases. The data indicates that for any workload that can use OpenCL or Vulkan, the discrete GeForce 940MX is the clear choice. The Intel HD 630, with its lower raw scores, is better suited for basic display output, video playback, and light 2D tasks, though the benchmark record itself does not contain specific tests for those use cases.

The Intel part's average benchmark score sits at 4075, which places it in the 24th percentile of all GPUs in the database. The NVIDIA part averages 4844, which lands in the 28th percentile. The percentile gap is modest, but the raw score difference is significant. The 940MX's average is about 18.9% higher than Intel's average, calculated from the two average figures in the record. The Intel GPU's nearest rivals include the AMD Radeon RX 9060 XT 8 GB at 4093 (0.4% behind Intel), the NVIDIA GeForce GT 755M at 4033 (1% behind Intel), and the AMD FirePro M4150 at 4013 (1.5% behind Intel). The 940MX's nearest rivals include the NVIDIA GeForce GTS 450 at 4893 (1% ahead of the 940MX) and the AMD Radeon R6 M255DX at 4867 (0.5% ahead). This places both GPUs in similar competitive tiers, but the 940MX consistently sits above the Intel part in the shared workloads.

For users who need compute performance in OpenCL or Vulkan, the 940MX is the winner. For users who only need a basic iGPU with Metal support, the Intel HD 630 has a niche, but that niche is narrow and does not overlap with the NVIDIA part's strengths.

Architecture Differences

The two GPUs come from entirely different design philosophies. The NVIDIA GeForce 940MX uses the GM107 chip, built on the Maxwell architecture, part of the GeForce 900M generation. It is fabricated on a 28 nm process at TSMC, with 1,870 million transistors packed into a 148 mm² die, giving a transistor density of 12.6 million transistors per square millimeter. The Intel HD Graphics 630 uses the Kaby Lake GT2 chip, built on Intel's Generation 9.5 architecture, from the HD Graphics (Kaby Lake) generation. It is fabricated on Intel's 14 nm++ process, though the database does not record its transistor count, die size, or transistor density.

The clock behavior differs sharply. The 940MX has a base clock of 795 MHz and a boost clock of 861 MHz, with memory clocked at 1253 MHz (5 Gbps effective). The Intel part has a base clock of 350 MHz and a boost clock of 1000 MHz, with system-shared memory. The Intel part boosts higher, but it starts from a much lower base, and its memory bandwidth is system dependent, not a fixed figure. The NVIDIA part has dedicated GDDR5 memory.

The memory subsystems are fundamentally different. The 940MX has 2 GB of GDDR5 on a 64-bit bus, delivering 40.10 GB/s of bandwidth. The Intel HD 630 uses system shared memory for everything: size, type, bus width, and bandwidth are all system dependent. This is the core architectural divide: a discrete GPU with its own fast memory versus an integrated GPU borrowing from system RAM.

Compute resources also diverge. The 940MX has 512 shading units, 32 texture mapping units, and 8 raster output units. The Intel part has 192 shading units, 24 TMUs, and only 3 ROPs. The pixel rate reflects this: 6.888 GPixel/s for NVIDIA versus 3.000 GPixel/s for Intel. Texture rate is closer: 27.55 GTexel/s versus 24.00 GTexel/s. FP32 performance is 881.7 GFLOPS for the 940MX versus 384.0 GFLOPS for Intel. The Intel part does have a half-precision advantage on paper, with 768.0 GFLOPS FP16 (2:1 ratio), while the NVIDIA part has no recorded FP16 figure.

Power and interface also differ. The 940MX is rated at 23 W TDP and uses an MXM module slot with no power connectors. The Intel part is rated at 15 W TDP and is an IGP (integrated graphics processor) on a ring bus. The 940MX connects via PCIe 3.0 x8, while the Intel part uses the CPU's ring bus. Display outputs are portable device dependent for NVIDIA and motherboard dependent for Intel, reflecting their different integration levels.

API support is a mixed bag. Both support DirectX 12 and OpenGL 4.6. The 940MX supports Vulkan 1.4, while Intel supports Vulkan 1.3. However, the DirectX feature level differs: the 940MX is DirectX 12 (11_0), meaning it only supports the 11_0 feature level, while Intel is DirectX 12 (12_1), supporting the full 12_1 feature level. That gives Intel an API feature advantage despite the NVIDIA part's newer Vulkan version.

Head-to-Head Benchmarks

The Geekbench OpenCL test is the first head-to-head comparison. The NVIDIA GeForce 940MX scores 4939, while the Intel HD Graphics 630 scores 3587. The delta is 37.7% in favor of NVIDIA. This is a decisive margin for a compute workload. The 940MX's score is also close to its nearest rival, the NVIDIA GeForce GTS 450 at 4893, which is only 1% ahead. The Intel part's OpenCL score is closer to its own rivals, with the AMD Radeon RX 9060 XT 8 GB at 4093 sitting 0.4% behind Intel, but that rival is still substantially ahead of Intel's 3587.

The Geekbench Vulkan test shows a similar pattern. The 940MX scores 4749, and the Intel part scores 3540. The delta is 34.2% in favor of NVIDIA. The NVIDIA part's Vulkan score is lower than its OpenCL score, but it still holds a comfortable lead. The Intel part's Vulkan score is nearly identical to its OpenCL score, suggesting the iGPU is consistent across these two APIs but cannot close the gap to the discrete GPU.

The Intel HD 630's Metal score of 5099 is worth noting, but it has no head-to-head counterpart. It is the only benchmark where Intel exceeds the 940MX's recorded numbers, and it exceeds both of the NVIDIA part's scores. However, Metal is Apple's API, and the 940MX does not have a recorded Metal result, so this is not a direct comparison. The Intel part's average benchmark score of 4075 is pulled up by that Metal result, but its OpenCL and Vulkan scores are both below 3600.

The data shows a clear hierarchy. The 940MX wins both shared tests by more than 34%. The Intel part wins no shared tests. The only Intel bright spot is a single Metal score in a workload the NVIDIA part does not appear in. For any user choosing between these two for general compute, the 940MX is the stronger part by a wide margin.

Specification Differences

The two GPUs differ in nearly every recorded specification. The process node is a key split: 28 nm TSMC for NVIDIA versus 14 nm++ Intel for the HD 630. The transistor count is recorded only for NVIDIA: 1,870 million. The die size is 148 mm² for NVIDIA, with no Intel figure. The 940MX's base clock is 795 MHz with an 861 MHz boost; Intel's base is 350 MHz with a 1000 MHz boost. Memory is GDDR5 2 GB on a 64-bit bus with 40.10 GB/s bandwidth for NVIDIA, versus system shared for all memory fields on Intel.

The shading units are 512 for NVIDIA versus 192 for Intel. TMUs are 32 versus 24. ROPs are 8 versus 3. Pixel rate is 6.888 GPixel/s versus 3.000 GPixel/s. Texture rate is 27.55 GTexel/s versus 24.00 GTexel/s. FP32 is 881.7 GFLOPS versus 384.0 GFLOPS. FP16 is not recorded for NVIDIA but is 768.0 GFLOPS for Intel. TDP is 23 W versus 15 W. Slot width is MXM Module versus IGP. Bus interface is PCIe 3.0 x8 versus Ring Bus. Display outputs are portable device dependent versus motherboard dependent. Vulkan support is 1.4 for NVIDIA versus 1.3 for Intel. DirectX feature level is 11_0 for NVIDIA versus 12_1 for Intel. The 940MX has no power connectors recorded; Intel has no power connector field populated.

Release dates are close: the 940MX launched on 2016-06-27, and the HD 630 launched on 2016-08-29. Both are end-of-life. The 940MX has a predecessor (GeForce 800M) and a successor (GeForce 10 Mobile); Intel has neither recorded.

FAQ

Q: Which GPU wins in OpenCL performance?

A: The NVIDIA GeForce 940MX wins decisively. It scores 4939 in Geekbench OpenCL versus the Intel HD Graphics 630's 3587, a 37.7% advantage.

Q: Does the Intel HD Graphics 630 beat the 940MX in any benchmark?

A: In the shared head-to-head tests, no. Intel wins zero of the two recorded comparisons. However, Intel has a unique Geekbench Metal score of 5099, a test the NVIDIA part does not have a recorded result for.

Q: How do their average benchmark scores compare?

A: The 940MX has an average benchmark score of 4844, while the Intel HD 630 averages 4075. The NVIDIA part sits in the 28th percentile of all GPUs, and the Intel part sits in the 24th percentile.

Q: What is the memory configuration difference?

A: The 940MX has 2 GB of dedicated GDDR5 memory on a 64-bit bus with 40.10 GB/s bandwidth. The Intel HD 630 uses system shared memory for size, type, bus width, and bandwidth, with bandwidth listed as system dependent.

Q: Which GPU has higher DirectX feature level support?

A: The Intel HD Graphics 630 supports DirectX 12 (12_1), the full feature level. The NVIDIA GeForce 940MX supports DirectX 12 (11_0), which is a lower feature level despite both being DirectX 12.

Q: How do their TDP ratings differ?

A: The 940MX is rated at 23 W, and the Intel HD 630 is rated at 15 W. The NVIDIA part is a discrete MXM module, while the Intel part is an integrated GPU on a ring bus.