Intel UHD Graphics P630 vs NVIDIA GeForce GTX 460 SE Comparison

The NVIDIA GeForce GTX 460 SE wins this head-to-head on raw compute, but the Intel UHD Graphics P630 counters with a decade of platform modernity, and the recorded data makes that tension the whole story. In the only direct benchmark both cards share, Geekbench OpenCL, the GTX 460 SE posts 6389 against 5111 for the P630, a 25% advantage. Yet the P630 is the newer design by a wide margin, carries a far lower power envelope, and supports a newer API feature set. Neither card is a performance leader today: both sit in the lower third of the database, with the GTX 460 SE at the 37th percentile and the P630 at the 31st. What follows is where each claim holds up and where it does not.

FAQ

Q: Which GPU is faster overall?

A: The NVIDIA GeForce GTX 460 SE. It won the only shared benchmark, Geekbench OpenCL, with a score of 6389 versus 5111 for the Intel UHD Graphics P630, a 25% margin. It also leads in raw specifications: 748.8 GFLOPS of FP32 compute against 460.8 GFLOPS, 7.800 GPixel/s against 3.600 GPixel/s, and 31.20 GTexel/s against 28.80 GTexel/s.

Q: How much more power does the GTX 460 SE need?

A: A great deal more. The GTX 460 SE has a 150 W TDP and requires 2x 6-pin power connectors, with a suggested PSU of 450 W. The P630 is an integrated GPU with a 15 W TDP and no power connectors at all.

Q: Which GPU has better API support?

A: The P630. It supports DirectX 12 at feature level 12_1, OpenGL 4.6, and Vulkan 1.3. The GTX 460 SE supports DirectX 12 at feature level 11_0 and OpenGL 4.6, with no Vulkan support recorded.

Q: How do they compare to other GPUs in the database?

A: Both rank low. The GTX 460 SE sits at the 37th percentile versus all GPUs; the P630 sits at the 31st percentile. Their closest rivals also cluster tightly: the GTX 460 SE is essentially level with the NVIDIA GeForce GTX 580M, the RTX PRO 5000 72 GB Blackwell, and the AMD Radeon Pro WX 4100, while the P630 trades blows with the AMD Radeon R7 M445, the R7 M365X, and the NVIDIA GeForce 840M.

Q: How does memory compare?

A: The GTX 460 SE has dedicated memory: 1024 MB of GDDR5 on a 256-bit bus, running at 850 MHz (3.4 Gbps effective) for 108.8 GB/s of bandwidth. The P630 shares system memory, so its capacity, bandwidth, and performance are entirely dependent on the host platform.

Q: Which one is smaller and easier to install?

A: The P630 is an integrated GPU using a ring bus interface, so it occupies no slot and needs no installation. The GTX 460 SE is a dual-slot card on PCIe 2.0 x16, measuring 210 mm (8.3 inches) long.

Architecture Differences

These two products come from entirely different eras and design philosophies. The GTX 460 SE is a discrete card built on the GF104 chip, NVIDIA's Fermi architecture, and part of the GeForce 400 generation. It was manufactured by TSMC on a 40 nm process, packs 1,950 million transistors onto a 332 mm² die, and yields a transistor density of 5.9M per mm². It succeeds the GeForce 200 line and was itself succeeded by the GeForce 500 series.

The P630 is the opposite approach: an integrated GPU based on the Comet Lake GT2 chip, using Intel's Generation 9.5 graphics architecture on a 14 nm+++ process from Intel's own foundries. Released in 2020 as part of the HD Graphics-W Comet Lake family, it is more than nine years newer than the GTX 460 SE's late-2010 launch. No transistor count, die size, or density figures are recorded for it.

The internal resource counts diverge sharply. The GTX 460 SE fields 288 shading units, 48 texture mapping units, and 32 render output units. The P630 has 192 shading units, 24 TMUs, and just 3 ROPs. That ROP gap is enormous and structural: 32 versus 3 means the discrete card can push pixels out the back end at more than double the integrated part's rate. The P630 compensates somewhat with clock speed, boosting to 1200 MHz from a 350 MHz base, and it is the only one of the two with recorded FP16 capability, at 921.6 GFLOPS running a 2:1 rate. Neither GPU has RT cores or tensor cores.

Head-to-Head Benchmarks

The recorded data contains one direct comparison, and it goes decisively to the discrete card.

Geekbench OpenCL: NVIDIA GeForce GTX 460 SE 6389, Intel UHD Graphics P630 5111, a 25% win for the GTX 460 SE. That margin is consistent with the theoretical hardware advantage: the card holds roughly 60% more raw FP32 throughput and more than double the pixel fill rate. The P630's higher boost clock cannot close a gap that wide when it has fewer shading units, fewer TMUs, and a tenth of the ROPs.

The P630 also recorded a Geekbench Vulkan score of 5628, but the GTX 460 SE has no Vulkan result, so no direct comparison is possible there. What the Vulkan score does show is that the P630 remains competitive in modern graphics APIs, a domain the GTX 460 SE cannot even enter given its lack of recorded Vulkan support.

Context matters too. The GTX 460 SE's 6389 average score places it dead even with the GeForce GTX 580M, fractionally behind the RTX PRO 5000 72 GB Blackwell (-0.3%), and slightly ahead of the Radeon Pro WX 4100 (+0.9%) and the Radeon R7 M350 (+1%). The P630's 5370 average sits within a point of the Radeon R7 M445 (+0.2%), the R7 M365X (-0.8%), the GeForce 840M (+0.9%), and the GeForce 930A (+1%). Both cards inhabit the same crowded neighborhood of modest performers; the GTX 460 SE simply lives one street over on higher ground.

Specification Differences

• Process node: 40 nm (TSMC) for the GTX 460 SE versus 14 nm+++ (Intel) for the P630.
• Shading units: 288 versus 192.
• TMUs: 48 versus 24.
• ROPs: 32 versus 3.
• FP32 compute: 748.8 GFLOPS versus 460.8 GFLOPS. FP16: not recorded for the GTX 460 SE; 921.6 GFLOPS (2:1) for the P630.
• Clocks: the GTX 460 SE has no recorded base or boost clock, only memory at 850 MHz (3.4 Gbps effective). The P630 runs a 350 MHz base and 1200 MHz boost on system-shared memory.
• Memory: 1024 MB GDDR5, 256-bit, 108.8 GB/s versus system-shared capacity, type, and bandwidth.
• Pixel rate: 7.800 GPixel/s versus 3.600 GPixel/s. Texture rate: 31.20 GTexel/s versus 28.80 GTexel/s.
• TDP: 150 W versus 15 W, with 2x 6-pin connectors and a 450 W suggested PSU against none.
• Form factor: dual-slot discrete card, 210 mm long, PCIe 2.0 x16, with 2x DVI and 1x mini-HDMI 1.3a outputs versus an integrated GPU on a ring bus with motherboard-dependent display outputs.
• API support: DirectX 12 (11_0) and OpenGL 4.6 versus DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3.
• Release date: 2010 for the GTX 460 SE, 2020 for the P630. Its launch MSRP was 160 USD.

Where Each One Wins

The GTX 460 SE wins on raw throughput, and by a wide margin. Every performance-oriented specification favors it: more shading units, double the TMUs, over ten times the ROPs, higher pixel and texture fill rates, and substantially more FP32 compute. The benchmark data confirms the theory with a 25% Geekbench OpenCL victory. Its dedicated 108.8 GB/s of GDDR5 bandwidth is guaranteed, unlike the P630's system-dependent memory. For any workload that leans on fixed-function output capacity or sustained compute, the discrete card is the stronger piece of silicon.

The P630 wins everywhere the spec sheet is not about speed. Its 15 W TDP is a tenth of the GTX 460 SE's draw, and it needs no power connectors, no suggested PSU headroom, and no slot. Its API support is strictly more modern: DirectX 12 feature level 12_1 versus 11_0, and Vulkan 1.3 versus none recorded. Its FP16 throughput at 921.6 GFLOPS actually exceeds its own FP32 rate, giving it a capability the older card simply lacks on record. And its 5628 Vulkan benchmark result demonstrates functioning performance in an API the Fermi card cannot run.

The Verdict

On pure performance, the data gives this to the NVIDIA GeForce GTX 460 SE without qualification: one shared benchmark, one win, 25% faster, backed by superior hardware across every throughput metric. A buyer chasing the higher score picks the discrete card and accepts the dual-slot footprint, the 150 W TDP, the 2x 6-pin connectors, and the 450 W PSU recommendation that come with it.

The Intel UHD Graphics P630 is the choice when power efficiency, platform integration, and API modernity matter more than benchmark numbers. It delivers 80% of the GTX 460 SE's OpenCL performance at a tenth of the power, supports a newer DirectX feature level plus Vulkan 1.3, and occupies no slot at all. Both are end-of-life parts in the lower third of the database, so neither is a performance pick against current hardware; the decision is strictly between the older, faster discrete card and the newer, efficient integrated one.