Intel UHD Graphics P630 vs NVIDIA GeForce 940M Comparison

Intel UHD Graphics P630 and NVIDIA GeForce 940M represent two very different approaches to mobile graphics, separated by five years of silicon evolution. The Intel part is a 2020 integrated solution built on a 14 nm+++ process, while the NVIDIA chip is a 2015 discrete GPU on a 28 nm node from TSMC. Despite their age difference, the benchmark data reveals a surprisingly close contest, with the average scores landing within 1.6% of each other — 5370 for the Intel versus 5284 for the NVIDIA. Both GPUs sit at the 31st percentile of all GPUs, indicating they occupy the same performance tier, but the way they achieve that standing differs radically. The head-to-head results show a split decision: the NVIDIA wins OpenCL by 15.1%, while the Intel counters with a 23.7% victory in Vulkan. This creates an intriguing narrative about workload-dependent performance, architectural efficiency, and what "winning" means when comparing an integrated processor's graphics block to a dedicated mobile part.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel UHD Graphics P630 posts an average benchmark score of 5370, while the NVIDIA GeForce 940M scores 5284. That puts the Intel part 1.6% ahead in aggregate, despite the NVIDIA having a discrete memory interface.

Q: How do the two GPUs compare in OpenCL performance?

A: The NVIDIA GeForce 940M wins OpenCL decisively, scoring 6018 against the Intel's 5111. This represents a 15.1% advantage for the NVIDIA part, reflecting its dedicated DDR3 memory and higher raw compute throughput.

Q: What happens in the Vulkan benchmark?

A: The Intel UHD Graphics P630 reverses the trend, scoring 5628 versus the NVIDIA's 4549. That is a 23.7% margin in favor of Intel, suggesting the integrated part handles the Vulkan API's modern rendering paths more efficiently.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce 940M contains 512 shading units, while the Intel UHD Graphics P630 has only 192. Despite this 2.7x difference in shader count, the Intel part still matches the NVIDIA in average performance, indicating architectural efficiency.

Q: What are the memory configurations of these two GPUs?

A: The NVIDIA GeForce 940M has 2 GB of DDR3 memory on a 64-bit bus with 14.40 GB/s bandwidth. The Intel UHD Graphics P630 uses System Shared memory, meaning it borrows from the main system RAM, with bandwidth described as System Dependent.

Q: Do both GPUs support DirectX 12?

A: Yes, both support DirectX 12, but the implementations differ. The Intel part supports DirectX 12 (12_1) with feature level 12_1, while the NVIDIA supports DirectX 12 (11_0), indicating a lower feature level for the latter.

Architecture Differences

The architectural divide between these two GPUs is stark. Intel's UHD Graphics P630 employs the Generation 9.5 architecture, built on a 14 nm+++ process at Intel's foundry. The chip, codenamed Comet Lake GT2, integrates 192 shading units, 24 texture mapping units, and just 3 raster output pipelines. This is a low-power design with a 15 W TDP, intended to live on the Ring Bus interface of a processor. The NVIDIA GeForce 940M, by contrast, uses the Maxwell architecture on a 28 nm process from TSMC, with a transistor count of 1,870 million on a 148 mm² die. It packs 512 shading units, 32 TMUs, and 16 ROPs, running at a base clock of 1020 MHz and boost of 1098 MHz. The NVIDIA part consumes 75 W and mounts as an MXM Module with an MXM-B (3.0) interface.

Clock speeds tell a different story. The Intel GPU has a 350 MHz base clock that boosts to 1200 MHz, which is higher than the NVIDIA's 1098 MHz boost. However, the NVIDIA compensates with far more execution resources. The pixel rate favors NVIDIA dramatically — 17.57 GPixel/s versus 3.600 GPixel/s — a 4.9x advantage. Texture rate also leans NVIDIA at 35.14 GTexel/s versus 28.80 GTexel/s. The FP32 compute is another mismatch: NVIDIA delivers 1,124.4 GFLOPS against Intel's 460.8 GFLOPS. Interestingly, the Intel part offers FP16 at 921.6 GFLOPS with a 2:1 ratio, while the NVIDIA lists no FP16 capability. Memory is where the two diverge completely: the NVIDIA has dedicated 2 GB DDR3 with a 64-bit bus and 14.40 GB/s bandwidth, whereas the Intel uses System Shared memory with bandwidth dependent on the host system. The API support shows the Intel part supporting Vulkan 1.3 and DirectX 12 (12_1), while the NVIDIA supports Vulkan 1.4 but only DirectX 12 (11_0).

Head-to-Head Benchmarks

The Geekbench results paint a picture of two GPUs that trade blows depending on the API. In OpenCL, the NVIDIA GeForce 940M scores 6018, which is 15.1% higher than the Intel UHD Graphics P630's 5111. This aligns with the NVIDIA's hardware advantages: 2.7x more shading units, 5.3x more ROPs, and dedicated memory with a known 14.40 GB/s bandwidth. OpenCL workloads often stress raw compute throughput and memory access patterns, where the NVIDIA's discrete design should shine. The 907-point gap in absolute terms is substantial, suggesting that any OpenCL-based application — whether compute or rendering — will favor the NVIDIA part.

Flip to Vulkan, and the tables turn. The Intel UHD Graphics P630 scores 5628 against the NVIDIA's 4549, a 23.7% advantage for Intel. This is a 1,079-point swing in the opposite direction. Vulkan's lower-level API reduces driver overhead and allows more direct hardware control, which may benefit the Intel architecture's design choices. The Intel GPU's higher boost clock of 1200 MHz, combined with its 2:1 FP16 ratio, could provide advantages in certain Vulkan workloads that utilize half-precision math. The NVIDIA's lack of FP16 support might handicap it here. The result is a net split: one win apiece, with the Intel winning by a larger margin (23.7%) than the NVIDIA's OpenCL victory (15.1%). In the average benchmark score, the Intel's aggregate 5370 edges out the NVIDIA's 5284, meaning the Vulkan win carries more weight in the overall assessment.

The Verdict

The data does not deliver a clear overall winner; it delivers a conditional one. For OpenCL-heavy workloads — think general compute, some professional applications, and older game engines — the NVIDIA GeForce 940M is the stronger choice, evidenced by its 6018 score and 15.1% lead. Its dedicated 2 GB DDR3 memory and 512 shading units provide a hardware foundation that the Intel part cannot match in this API. However, for Vulkan-based applications, which increasingly dominate modern gaming and compute, the Intel UHD Graphics P630 holds a 23.7% advantage with its 5628 score. This is remarkable given that the Intel part consumes only 15 W versus the NVIDIA's 75 W, and it shares system memory rather than having its own pool.

The percentile ranking of 31 for both GPUs suggests they sit in the same performance class relative to all GPUs, but the nearest rivals data adds nuance. The Intel's closest competitor is the AMD Radeon R7 M365X at 5416 (0.8% higher), while the NVIDIA's nearest rival includes the NVIDIA GeForce GTX 980M at 5308 (0.4% higher) — a much more powerful part in theory, yet scoring similarly. This implies the GeForce 940M is underperforming in these benchmarks relative to its hardware specifications, possibly due to driver maturity or thermal constraints. For a user deciding between a system with an Intel UHD Graphics P630 and one with a NVIDIA GeForce 940M, the choice hinges on API usage. Vulkan-centric users gain 23.7% with Intel; OpenCL-centric users gain 15.1% with NVIDIA. The average benchmark score slightly favors Intel, making it the safer default for mixed workloads.

Specification Differences

The two GPUs differ across nearly every measurable specification. The Intel UHD Graphics P630 uses a 14 nm+++ process from Intel's foundry, while the NVIDIA GeForce 940M relies on a 28 nm process from TSMC. The NVIDIA chip integrates 1,870 million transistors on a 148 mm² die, yielding a transistor density of 12.6M / mm²; the Intel part lists no transistor count or die size. Clock speeds differ with Intel at 350 MHz base / 1200 MHz boost versus NVIDIA at 1020 MHz base / 1098 MHz boost. Memory is a fundamental split: Intel uses System Shared memory with System Dependent bandwidth, while NVIDIA has 2 GB DDR3, a 64-bit bus, and 14.40 GB/s bandwidth. The memory clock on the NVIDIA is 900 MHz with 1800 Mbps effective; Intel's is System Shared.

Execution resources show NVIDIA's advantage: 512 shading units versus 192, 32 TMUs versus 24, and 16 ROPs versus 3. Pixel rate is 17.57 GPixel/s for NVIDIA versus 3.600 GPixel/s for Intel; texture rate is 35.14 GTexel/s versus 28.80 GTexel/s. FP32 compute is 1,124.4 GFLOPS for NVIDIA versus 460.8 GFLOPS for Intel. The Intel part offers FP16 at 921.6 GFLOPS (2:1), while NVIDIA lists none. Power consumption varies wildly: Intel at 15 W versus NVIDIA at 75 W. Form factor differs with Intel being an IGP on a Ring Bus interface, while NVIDIA is an MXM Module with an MXM-B (3.0) interface. Display outputs are Motherboard Dependent for Intel and Portable Device Dependent for NVIDIA. API support shows Intel with DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3; NVIDIA has DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. Release dates are 2020-05-12 for Intel and 2015-03-12 for NVIDIA, with both marked End-of-life.

Where Each One Wins

The NVIDIA GeForce 940M wins in OpenCL performance, taking a 15.1% lead with its 6018 score. This makes it the preferred option for OpenCL compute tasks, where its 512 shading units, 16 ROPs, and dedicated 14.40 GB/s memory bandwidth can be fully utilized. The NVIDIA also dominates in raw pixel throughput at 17.57 GPixel/s and FP32 compute at 1,124.4 GFLOPS, which could translate to better performance in legacy DirectX 11 workloads that rely on traditional rasterization. Its 2 GB dedicated memory avoids the bandwidth contention that plagues shared-memory designs, making it more consistent in memory-intensive scenarios.

The Intel UHD Graphics P630 wins in Vulkan, posting a 23.7% higher score of 5628. This indicates superior efficiency in modern, low-overhead APIs where driver behavior and architectural design matter more than raw hardware count. The Intel part's 1200 MHz boost clock and FP16 support at 921.6 GFLOPS give it an edge in Vulkan workloads that can leverage half-precision arithmetic. Additionally, the Intel GPU consumes only 15 W versus the NVIDIA's 75 W, making it the clear winner for battery life and thermal management in thin-and-light systems. Its higher average benchmark score of 5370 versus 5284 also means it edges ahead in aggregate performance. For users running Vulkan-based games or compute, the Intel part offers a better balance of performance and power efficiency. The split decision — one win each — suggests that neither GPU is universally superior, and the choice should be dictated by the specific application programming interface in use.