Intel UHD Graphics 730 vs NVIDIA GeForce GTX 460 SE Comparison

# NVIDIA GeForce GTX 460 SE vs Intel UHD Graphics 730

The NVIDIA GeForce GTX 460 SE and Intel UHD Graphics 730 occupy very different corners of the GPU market. The GTX 460 SE is a discrete, end-of-life Fermi architecture part from 2010, designed for enthusiast desktop systems. The UHD 730, by contrast, is a modern integrated graphics processor built into Intel's Rocket Lake CPUs, representing the current baseline for everyday computing. Their benchmark profiles confirm this split: the GTX 460 SE records a Geekbench OpenCL score of 6389, while the UHD 730 manages 5988 in the same test, a delta of 6.7%.

The data shows the GTX 460 SE is the stronger of the two in raw compute. It wins their head-to-head benchmark by 401 points. The UHD 730, however, is not a performance part, but an efficiency and platform convenience play. The choice between them depends entirely on the workload and platform constraints.

Architecture Differences

The two chips are separated by a decade of architectural evolution, and it shows in nearly every specification.

The GTX 460 SE is built on NVIDIA's Fermi architecture, using the GF104 chip. It is fabricated on a 40 nm process at TSMC, with 1,950 million transistors on a 332 mm² die. This yields a transistor density of 5.9 million per square millimeter. It uses a 256-bit GDDR5 memory bus, sized at 1024 MB (1 GB), running at 850 MHz with an effective 3.4 Gbps data rate, producing 108.8 GB/s of memory bandwidth. The card offers 288 shading units, 48 texture mapping units, and 32 ROPs. Pixel throughput is rated at 7.800 GPixel/s, texture fill rate at 31.20 GTexel/s, and FP32 compute is 748.8 GFLOPS. It draws up to 150 W, uses a dual-slot cooler, and requires two 6-pin power connectors. The database lists a 450 W power supply as the suggestion. It connects over PCIe 2.0 x16, and outputs via 2x DVI and 1x mini-HDMI 1.3a. This part is end-of-life, with a launch date of November 2010 and a launch MSRP of 160 USD.

The Intel UHD 730 is a Generation 12.1 (Rocket Lake) chip, built on Intel's 14nm+++ process. It is an integrated part of the CPU, not a discrete card. It has no dedicated memory; the memory is system shared and its bandwidth is system dependent. It sports 192 shading units and 12 TMUs, with 8 ROPs. Clock behavior is very different: a base of 300 MHz and a boost of 1300 MHz, but the memory runs at system shared speeds. Pixel throughput is 10.40 GPixel/s and texture rate is 15.60 GTexel/s. FP32 is 499.2 GFLOPS and FP16 is 998.4 GFLOPS (2:1 ratio). Thermal design power is a mere 15 W. This is a single-slot, slot-less IGP card with no power connectors and no PSU suggestion. Video outputs are motherboard dependent. It supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. This part is end-of-life, released March 2021.

The Intel part has no dedicated memory or bandwidth figures to compare, so the data for the two is not symmetric. The architecture, features, and numbers that do exist are outlined above for the record.

Head-to-Head Benchmarks

The only head-to-head benchmark in the database is the Geekbench OpenCL test. In that test, the GTX 460 SE scored 6389, and the UHD 730 scored 5988. The result is a NVIDIA win: the delta is 6.7% in favor of the GTX 460 SE.

That 6.7% advantage comes from a very different place. The GTX 460 SE's dedicated GDDR5 memory, 256-bit bus, and a much larger 150 W power budget allow the Fermi GPU to push raw shader throughput at a much higher rate: its 748.8 GFLOPS FP32 is 50% higher than the Intel part's 499.2 GFLOPS. Graphics bandwidth, 108.8 GB/s versus system dependent, is the single biggest differentiator. The Intel part does have one trick up its sleeve: 998.4 GFLOPS FP16 (2:1 ratio) is 33% above the GTX's FP32 figure, but in practice it is not for floating-point heavy workloads. As it is an integrated design, it will be constrained by the CPU's power envelope and memory bandwidth.

In the entire database, the GTX 460 SE is above the 37th percentile of all GPUs, with an average benchmark score of 6389. The UHD 730 is at the 33rd percentile with an average of 5929. On the other hand, the closest rivals to the GTX 460 SE within 1% are the GTX 580M (0%), RTX PRO 5000 72 GB Blackwell (-0.3%), AMD Radeon Pro WX 4100 (0.9%), and AMD Radeon R7 M350 (1%). The UHD 730's nearest rivals are all within -1%: AMD Radeon HD 8730M (-0.4%), NVIDIA Quadro K620M (-0.5%), AMD Radeon HD 8750M (-0.7%), and NVIDIA Quadro K4000 (-0.9%).

The 6.7% lead in OpenCL is the only real difference in the data. It is a discrete GPU class versus an integrated GPU, and the delta is still modest in absolute terms.

The Verdict

From the raw data, the NVIDIA GeForce GTX 460 SE is the faster GPU. It wins the only head-to-head test, and is 6.7% ahead of the Intel UHD 730 in OpenCL. If the workload is on a discrete, high-throughput GPU with full GDDR6 memory bandwidth and 150 W of power, the GTX 460 SE is the clear choice.

That said, the Intel UHD 730 is not meant to compete on the same plane. Its strengths are platform integration, a 15 W TDP, and no extra power connectors. For productivity desktops and laptops where the CPU's iGPU is sufficient for the task, the UHD 730 is a more sensible part. It has no memory of its own, but for a system that is shared, it is fine.

Pick the GTX 460 SE if you have a PCIe slot, power connectors, and a workload that needs the extra shader throughput and dedicated memory. Pick the UHD 730 if you're building a system where simplicity, low heat, and minimal power draw are key, and the GPU will be subjected to mainstream desktop tasks. The database's data has one outcome: the GTX 460 SE is the faster part, the UHD 730 is the more efficient part.

Specification Differences

| Field | NVIDIA GTX 460 SE | Intel UHD 730 |

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

| Chip | GF104 | Rocket Lake |

| Architecture | Fermi | Generation 12.1 |

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

| Foundry | TSMC | Intel |

| Transistors | 1,950 million | Not published |

| Die Size | 332 mm² | Not published |

| Transistor Density | 5.9M / mm² | Not published |

| Base Clock | Not published | 300 MHz |

| Boost Clock | Not published | 1300 MHz |

| Memory Clock | 850 MHz | System Shared |

| Memory Size | 1024 MB GDDR5 | System Shared |

| Memory Bus Width | 256 bit | System Shared |

| Memory Bandwidth | 108.8 GB/s | System Dependent |

| Shading Units | 288 | 192 |

| TMUs | 48 | 12 |

| ROPs | 32 | 8 |

| Pixel Rate | 7.800 GPixel/s | 10.40 GPixel/s |

| Texture Rate | 31.20 GTexel/s | 15.60 GTexel/s |

| FP32 | 748.8 GFLOPS | 499.2 GFLOPS |

| FP16 | Not listed | 998.4 GFLOPS (2:1) |

| TDP | 150 W | 15 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 2x 6-pin | None |

| Suggested PSU | 450 W | None |

| Bus Interface | PCIe 2.0 x16 | Ring Bus |

| Display Outputs | 2x DVI, 1x mini-HDMI 1.3a | Motherboard Dependent |

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

| OpenGL | 4.6 | 4.6 |

| Vulkan | Not listed | 1.4 |

| Length | 210 mm (8.3 inches) | Not listed |

| Release Date | 2010-11-14 | 2021-03-29 |

| Production Status | End-of-life | End-of-life |

| Launch MSRP | 160 USD | None |