Intel UHD Graphics P630 vs NVIDIA GeForce 840M Comparison

# Intel UHD Graphics P630 vs NVIDIA GeForce 840M

The Intel UHD Graphics P630 and NVIDIA GeForce 840M are evenly matched overall, with each winning one of the two head-to-head benchmarks. The NVIDIA part takes the Geekbench OpenCL test by 11.3%, while the Intel integrated graphics counters with a 15.3% victory in Geekbench Vulkan, resulting in a split decision that mirrors their near-identical average benchmark scores of 5370 and 5322 respectively. Both sit at the 31st percentile of all GPUs, placing them in the same performance tier despite their vastly different architectures and release dates.

Head-to-Head Benchmarks

The GeForce 840M delivers its strongest showing in Geekbench OpenCL, scoring 5764 against the P630’s 5111. That is an 11.3% margin, a decisive gap in compute workloads that favor NVIDIA’s dedicated graphics silicon. The 840M’s advantage here is substantial enough that any OpenCL-bound application will clearly prefer the discrete part. This result aligns with the 840M’s raw specifications: it packs 384 shading units and 863.2 GFLOPS of FP32 throughput, nearly double the P630’s 192 shading units and 460.8 GFLOPS. The P630 cannot close that compute gap through architecture alone.

The reverse is true in Geekbench Vulkan, where the Intel UHD Graphics P630 posts 5628 versus the 840M’s 4880, a 15.3% lead. This is a larger margin in Intel’s favor than NVIDIA’s OpenCL win, suggesting that the P630’s modern driver stack and Vulkan 1.3 API support (versus the 840M’s Vulkan 1.4) translate into better real-world graphics API performance. The P630’s Generation 9.5 architecture, despite being an integrated part, handles Vulkan workloads more efficiently than the older Maxwell-based 840M. Notably, the P630’s OpenCL score of 5111 is only 5.3% lower than its Vulkan score, while the 840M’s Vulkan score is 15.4% lower than its OpenCL result — the NVIDIA part is clearly optimized for compute over modern graphics APIs.

When placed against their nearest rivals, both GPUs occupy a tight cluster. The P630’s 5370 average sits just 0.2% above the AMD Radeon R7 M445 and 0.9% above the 840M itself, while the 840M’s 5322 average is 0.1% below the NVIDIA GeForce 930A and 0.3% above the GeForce GTX 980M — an unusual grouping where a high-end mobile part scores nearly identically to these mid-range options. The head-to-head delta between the two cards is a mere 0.9% in favor of the P630, well within run-to-run variance for most benchmarks.

Architecture Differences

The Intel UHD Graphics P630 is built on Intel’s Generation 9.5 architecture using a 14 nm+++ process node, manufactured by Intel itself. It features 192 shading units, 24 texture mapping units, and just 3 ROPs. The chip, codenamed Comet Lake GT2, is part of the HD Graphics-W generation and connects via a Ring Bus interface. Its memory subsystem is entirely System Shared, meaning it draws from main system memory with system-dependent bandwidth — no dedicated VRAM allocation. The P630 operates at a 350 MHz base clock with a 1200 MHz boost, producing a pixel rate of 3.600 GPixel/s and a texture rate of 28.80 GTexel/s. FP32 throughput is 460.8 GFLOPS, with FP16 at 921.6 GFLOPS via a 2:1 ratio. The TDP is just 15 W, reflecting its integrated nature, and it supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3.

The NVIDIA GeForce 840M, in contrast, uses the Maxwell architecture on a 28 nm TSMC process. Its GM108S chip contains 1,020 million transistors on a 77 mm² die, yielding a transistor density of 13.2M per mm². It has 384 shading units — double the P630 — but fewer texture units at 16, and more ROPs at 8. The 840M runs at a 1029 MHz base clock and 1124 MHz boost, with memory clocked at 1001 MHz (2 Gbps effective). It has 2 GB of dedicated DDR3 memory on a 64-bit bus, providing 16.02 GB/s of bandwidth — a fixed allocation versus the P630’s shared, variable memory. Pixel rate is 8.992 GPixel/s, texture rate is 17.98 GTexel/s, and FP32 reaches 863.2 GFLOPS. The 840M has no FP16 data listed, while the P630 offers half-precision support. TDP is 33 W, more than double the P630’s, and it connects via PCIe 3.0 x8. The 840M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4.

The architectural gulf is significant: Intel’s 14 nm+++ process versus TSMC’s 28 nm, integrated versus discrete memory, and a 2020 release date versus 2014 for the NVIDIA part. Despite the 840M’s higher raw compute figures, the P630’s newer architecture and superior Vulkan implementation keep it competitive in graphics-API workloads.

The Verdict

The data indicates that neither GPU holds a definitive overall advantage — each wins one benchmark, and their average scores differ by less than 1%. For OpenCL-heavy compute tasks, the NVIDIA GeForce 840M is the clear choice, offering 11.3% higher performance and nearly double the FP32 throughput. For Vulkan-based graphics workloads, the Intel UHD Graphics P630 wins by 15.3%, demonstrating that modern API support can overcome raw hardware disadvantages. The P630’s 15 W TDP also makes it far more power-efficient than the 840M’s 33 W, a critical factor for thin-and-light systems. However, the 840M’s dedicated 2 GB VRAM with fixed 16.02 GB/s bandwidth provides more predictable memory performance than the P630’s system-shared, system-dependent configuration. Users prioritizing compute should choose the 840M; those favoring modern graphics APIs or power efficiency should pick the P630.

Specification Differences

| Specification | Intel UHD Graphics P630 | NVIDIA GeForce 840M |

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

| Architecture | Generation 9.5 | Maxwell |

| Process Node | 14 nm+++ | 28 nm |

| Foundry | Intel | TSMC |

| Transistors | Not listed | 1,020 million |

| Die Size | Not listed | 77 mm² |

| Base Clock | 350 MHz | 1029 MHz |

| Boost Clock | 1200 MHz | 1124 MHz |

| Shading Units | 192 | 384 |

| TMUs | 24 | 16 |

| ROPs | 3 | 8 |

| Memory Size | System Shared | 2 GB |

| Memory Type | System Shared | DDR3 |

| Memory Bus | System Shared | 64 bit |

| Memory Bandwidth | System Dependent | 16.02 GB/s |

| Pixel Rate | 3.600 GPixel/s | 8.992 GPixel/s |

| Texture Rate | 28.80 GTexel/s | 17.98 GTexel/s |

| FP32 | 460.8 GFLOPS | 863.2 GFLOPS |

| FP16 | 921.6 GFLOPS (2:1) | Not listed |

| TDP | 15 W | 33 W |

| Bus Interface | Ring Bus | PCIe 3.0 x8 |

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

| Vulkan | 1.3 | 1.4 |

| Release Date | 2020-05-12 | 2014-03-11 |

| Predecessor | Not listed | GeForce 700M |

| Successor | Not listed | GeForce 900M |

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel UHD Graphics P630 has an average benchmark score of 5370, which is 0.9% higher than the NVIDIA GeForce 840M’s 5322.

Q: How large is the OpenCL performance gap between the two?

A: The NVIDIA GeForce 840M scores 5764 in Geekbench OpenCL versus the P630’s 5111, giving NVIDIA an 11.3% lead.

Q: Which GPU performs better in Vulkan workloads?

A: The Intel UHD Graphics P630 wins Geekbench Vulkan with 5628 points against the 840M’s 4880, a 15.3% advantage.

Q: Do the two GPUs have the same shading unit count?

A: No, the NVIDIA GeForce 840M has 384 shading units, exactly double the Intel UHD Graphics P630’s 192.

Q: What is the TDP difference between the two?

A: The Intel UHD Graphics P630 has a 15 W TDP, while the NVIDIA GeForce 840M has a 33 W TDP — the Intel part consumes less than half the power.

Q: Which GPU has dedicated video memory?

A: Only the NVIDIA GeForce 840M has dedicated memory, with 2 GB of DDR3 on a 64-bit bus providing 16.02 GB/s bandwidth. The Intel P630 uses system-shared memory with system-dependent bandwidth.

Where Each One Wins

The NVIDIA GeForce 840M wins in compute-heavy scenarios. Its OpenCL score of 5764 outpaces the P630 by 11.3%, and its 863.2 GFLOPS FP32 throughput is nearly double the Intel part’s 460.8 GFLOPS. The 840M’s 384 shading units and 8 ROPs also deliver a pixel rate of 8.992 GPixel/s, 2.5 times the P630’s 3.600 GPixel/s. Applications that rely on raw shader throughput or fixed-function rasterization will favor the discrete NVIDIA card. The 840M’s dedicated 2 GB VRAM with 16.02 GB/s bandwidth also reduces memory contention, making it suitable for workloads that require consistent memory access without relying on system RAM.

The Intel UHD Graphics P630 wins in Vulkan-based graphics and power-sensitive environments. Its 5628 Vulkan score beats the 840M by 15.3%, and its Vulkan 1.3 API support (alongside DirectX 12_1) indicates better compatibility with modern graphics frameworks. The P630’s 15 W TDP is less than half the 840M’s 33 W, making it the only sensible choice for ultra-portable devices where battery life and thermals are paramount. Its texture rate of 28.80 GTexel/s also exceeds the 840M’s 17.98 GTexel/s despite having fewer TMUs, suggesting more efficient texture processing per unit. For users running Vulkan games or applications on thin-and-light laptops, the P630 provides superior API-level performance while consuming minimal power.

The split is clean: the 840M for compute and raw throughput, the P630 for modern graphics APIs and efficiency. Their overall performance parity — 5370 versus 5322 average, both at the 31st percentile — means neither holds a universal advantage. The choice comes down to workload type and system constraints, with the data clearly favoring NVIDIA for OpenCL and Intel for Vulkan.