AMD Radeon 760M vs Intel UHD Graphics P750 Comparison

Where Each One Wins

The benchmark data splits these two integrated graphics processors into entirely different performance classes, with the AMD Radeon 760M taking a clean sweep in the only directly comparable test. In the Geekbench OpenCL benchmark, the AMD part scores 20,255 points against Intel's 6,554, a delta of -67.6% from AMD's perspective, meaning Intel trails by roughly two-thirds. That is not a marginal gap; it is a generational chasm that places the two chips in different tiers of integrated graphics capability.

However, the wins are not just about raw compute. The AMD Radeon 760M also brings a broader benchmark footprint, with results across 3DMark Steel Nomad, Passmark's DirectX 9/10/11/12 suites, G2D, G3D, and GPU compute tests. Intel's UHD Graphics P750 has only a single recorded benchmark, the Geekbench OpenCL score. This means the AMD part demonstrates verified competence across legacy DirectX 9 workloads (65 in Passmark DX9) and modern DirectX 12 Ultimate workloads (25 in Passmark DX12, plus a 400 in 3DMark Steel Nomad DX12), while Intel's data sheet offers no comparable evidence in those areas. For any use case involving gaming, content creation, or GPU-accelerated compute, the data favors AMD on both performance and breadth of validation.

Intel's sole "win" is contextual rather than numerical. The P750's 38th percentile ranking among all GPUs is slightly higher than AMD's 35th percentile, despite AMD's far higher raw OpenCL score. This occurs because AMD's average benchmark score (6,019) is dragged down by its low Passmark legacy scores, whereas Intel's single score is uniformly high. For users who rely primarily on OpenCL compute, Intel's position in the percentile distribution is marginally better, but that is a statistical artifact rather than a practical advantage.

Architecture Differences

The architectural divide between these two IGPs is stark and explains nearly every performance discrepancy. Intel's UHD Graphics P750 is built on the Generation 12.1 architecture using a 14 nm+++ process node from Intel's own foundry, paired with the Rocket Lake chip. It offers 256 shading units, 64 texture mapping units, and 32 ROPs. Clock speeds range from a 350 MHz base to a 1,300 MHz boost, producing 665.6 GFLOPS of FP32 compute and 1,331.2 GFLOPS of FP16 compute with a 2:1 ratio. The pixel rate is 41.60 GPixel/s, and the texture rate is 83.20 GTexel/s. Its bus interface is a Ring Bus, and it supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The part is classified as end-of-life.

AMD's Radeon 760M, by contrast, uses the RDNA 3.0 architecture on a 4 nm TSMC process, built around the Phoenix chip. It doubles the shading units to 512 but halves the TMUs to 32 and ROPs to 16. Clock speeds are dramatically higher: an 800 MHz base and a 2,599 MHz boost. This combination yields 5.323 TFLOPS of FP32 compute—exactly eight times Intel's FP32 figure—and 5.323 TFLOPS of FP16 with a 1:1 ratio. The pixel rate (41.58 GPixel/s) and texture rate (83.17 GTexel/s) are nearly identical to Intel's, which is notable given the eightfold compute advantage; this suggests AMD's render output and texture units are the limiting factor in some workloads. AMD also integrates 8 ray tracing cores, a feature Intel's part lacks entirely. The Radeon 760M connects via PCIe 4.0 x8, supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, and remains in active production with a release date of January 30, 2024, succeeding the Navi II IGP.

The process node difference is particularly meaningful: AMD's 4 nm TSMC fabrication versus Intel's 14 nm+++ represents a three-generation leap in manufacturing density. AMD's chip packs 25,390 million transistors into a 178 mm² die, yielding a transistor density of 142.6 million per square millimeter. Intel's transistor count and die size are not listed, but the older node and smaller shading unit count strongly imply a far less dense design. Both parts share a 15 W TDP and system-shared memory, but AMD achieves over eight times the FP32 throughput within the same power envelope, proof of architectural and process efficiency gains.

FAQ

Q: Which GPU has higher raw compute performance in OpenCL?

A: The AMD Radeon 760M scores 20,255 in Geekbench OpenCL, versus Intel's 6,554, a 67.6% advantage for AMD.

Q: Does Intel's UHD Graphics P750 have any ray tracing support?

A: No. The Intel part lists no ray tracing cores, while the AMD Radeon 760M integrates 8 RT cores.

Q: How do the two compare in terms of manufacturing process?

A: AMD uses a 4 nm TSMC process, while Intel uses a 14 nm+++ process from its own foundry. AMD's die is 178 mm² with 25,390 million transistors.

Q: Which GPU supports the newer DirectX version?

A: AMD supports DirectX 12 Ultimate (12_2), while Intel supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4.

Q: What is the FP32 compute difference between the two?

A: AMD delivers 5.323 TFLOPS FP32, exactly eight times Intel's 665.6 GFLOPS.

Q: Are both GPUs integrated or discrete?

A: Both are IGPs (integrated graphics processors) with a 15 W TDP and system-shared memory, though AMD uses a PCIe 4.0 x8 bus interface while Intel uses a Ring Bus.

Specification Differences

| Specification | Intel UHD Graphics P750 | AMD Radeon 760M |

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

| Architecture | Generation 12.1 | RDNA 3.0 |

| Process Node | 14 nm+++ | 4 nm |

| Foundry | Intel | TSMC |

| Chip | Rocket Lake | Phoenix |

| Shading Units | 256 | 512 |

| TMUs | 64 | 32 |

| ROPs | 32 | 16 |

| RT Cores | None | 8 |

| Base Clock | 350 MHz | 800 MHz |

| Boost Clock | 1,300 MHz | 2,599 MHz |

| FP32 | 665.6 GFLOPS | 5.323 TFLOPS |

| FP16 | 1,331.2 GFLOPS (2:1) | 5.323 TFLOPS (1:1) |

| Pixel Rate | 41.60 GPixel/s | 41.58 GPixel/s |

| Texture Rate | 83.20 GTexel/s | 83.17 GTexel/s |

| Bus Interface | Ring Bus | PCIe 4.0 x8 |

| DirectX | 12 (12_1) | 12 Ultimate (12_2) |

| Transistors | Not listed | 25,390 million |

| Die Size | Not listed | 178 mm² |

| Transistor Density | Not listed | 142.6M / mm² |

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

| Release Date | Not listed | 2024-01-30 |

The table reveals a clear pattern: AMD leads in compute-oriented specifications (shading units, clocks, FP32/FP16, RT cores) while Intel counters with more TMUs and ROPs per shading unit. Yet the pixel and texture rates are nearly identical, suggesting both are balanced to similar output throughput despite their architectural differences.

Head-to-Head Benchmarks

The only direct head-to-head benchmark available is Geekbench OpenCL, and it is decisively one-sided. AMD's Radeon 760M posts 20,255 against Intel's 6,554, with a delta of -67.6% for Intel. To put this in context: Intel's score sits just 0.4% below the AMD Radeon HD 7730M and 0.9% above the NVIDIA GeForce GT 555M, placing it in the range of mid-2010s discrete mobile GPUs. AMD's score, meanwhile, lands within 0.3% of the AMD Radeon RX 6400 and 0.3% of the NVIDIA GeForce GTX 770M, both of which are far more recent discrete parts. The Radeon 760M is effectively competing with entry-level discrete graphics from a few generations ago, while the Intel P750 is competing with integrated-class performance from a decade earlier.

Beyond OpenCL, AMD provides a full suite of Passmark results that Intel simply does not match. The Radeon 760M scores 5,310 in Passmark G3D, 2,840 in GPU compute, and 890 in G2D. Its legacy DirectX scores are 65 (DX9), 52 (DX11), 25 (DX12), and 19 (DX10). The 3DMark Steel Nomad DX12 result of 400 further confirms its modern API capability. Intel has no comparable data points, meaning the P750's only verified strength is OpenCL compute, where it is nonetheless outclassed by more than 3x.

The percentile rankings add nuance. Intel's 38th percentile versus AMD's 35th percentile appears contradictory given the OpenCL gap, but it reflects the fact that AMD's average benchmark score (6,019) is pulled down by low legacy scores, while Intel's average (6,554) is based solely on its one strong OpenCL result. In practical terms, if a workload is OpenCL-heavy, Intel's position is marginally better relative to the entire GPU universe, but the absolute performance gap remains enormous.

The Verdict

The data points to one clear conclusion: the AMD Radeon 760M is the superior processor for virtually any compute or graphics workload. Its 20,255 OpenCL score is more than three times Intel's 6,554, and its FP32 throughput of 5.323 TFLOPS is exactly eight times higher. The inclusion of 8 ray tracing cores and DirectX 12 Ultimate support means it is ready for modern gaming and ray-traced content, while Intel's DirectX 12_1 and lack of RT cores limits it to older or lighter workloads. The Radeon 760M's active production status and 2024 release date also suggest ongoing driver support and longevity, whereas the P750 is end-of-life.

Who should pick Intel's UHD Graphics P750? Only those constrained to a Rocket Lake platform with no alternative, or those whose sole requirement is OpenCL compute where a 38th percentile ranking (versus AMD's 35th) provides a marginal statistical edge. Even then, the absolute OpenCL deficit makes that choice difficult to justify on performance grounds.

Who should pick the AMD Radeon 760M? Anyone seeking integrated graphics that approach entry-level discrete GPU performance. Its nearest rivals include the AMD Radeon RX 6400 and NVIDIA GeForce GTX 770M, both discrete cards, and it outperforms the NVIDIA Quadro P2000 by 0.5%. The 4 nm process, 512 shading units, and 2,599 MHz boost clock deliver class-leading compute density within a 15 W envelope. For gaming, GPU compute, or any modern API workload, the Radeon 760M is the only rational choice based on the available benchmark evidence. The single head-to-head test is lopsided, and the specification sheet reinforces that result at every turn. Intel's part is not without merit—it posts respectable pixel and texture rates—but it cannot match AMD's architectural and process advantages. The verdict is unambiguous: AMD wins on every measurable performance axis, and the only question is whether Intel's integrated graphics can serve as a fallback for legacy systems.