GPU Comparison

The AMD Radeon RX 5700 and Intel Arc A750 are both end-of-life graphics cards with 8 GB of GDDR6 memory, but they deliver contrasting performance profiles. The data shows the RX 5700 wins 6 of 10 head-to-head benchmarks, while the Arc A750 wins the remaining 4. However, the Intel card’s victories are decisively lopsided in modern APIs, whereas the AMD card’s wins are mostly in legacy and compute workloads. The Arc A750 also holds a significant hardware advantage in raw specifications, including roughly 2.2x the FP32 throughput and 2.9x the texture rate, yet its average benchmark score (20582) trails the RX 5700 (22170) by 7.2%.

Head-to-Head Benchmarks

The Intel Arc A750 dominates where it matters most for contemporary gaming: DirectX 12 and Vulkan. In the 3DMark Steel Nomad DX12 test, the A750 scores 2612 against the RX 5700’s 1862, a 28.7% advantage. This is the single largest performance gap in the comparison. The pattern repeats in Vulkan, where the A750 posts 85631 versus 64166, a 25.1% lead, and in OpenCL, where it scores 98554 against 69603, a 29.4% margin. The PassMark DirectX 12 test also favors Intel, with the A750 at 70 versus the RX 5700’s 54, a 22.9% delta.

The RX 5700 fights back decisively in older DirectX APIs. Its PassMark DirectX 11 score of 101 beats the A750’s 72 by 40.3%, the largest win for either card. In DirectX 10, the RX 5700 leads 87 to 65, a 33.8% margin, and in DirectX 9 it wins 209 to 181, a 15.5% edge. The AMD card also sweeps the compute and 2D tests: PassMark GPU Compute shows 6517 versus 5368 (21.4% ahead), and PassMark G2D shows 877 versus 732 (19.8% ahead). The overall 3D score tells a similar story, the RX 5700’s 14314 beats the A750’s 12534 by 14.2%.

These results reveal a clear generational split. The A750 was built for modern rendering pipelines and excels in them, while the RX 5700 retains strong compatibility with legacy APIs. The average scores reflect this: the RX 5700 sits at 22170, placing it in the 67th percentile of all GPUs, while the A750 averages 20582, good for the 66th percentile. The RX 5700’s nearest rival is the AMD Radeon RX 6700S (0.1% delta), and it also edges the NVIDIA GeForce GTX 1060 6 GB by 1.4%. The A750’s closest competitor is the Intel Arc B570 (0.1% delta), and it narrowly beats the NVIDIA GeForce RTX 3070 Mobile by 0.2%.

FAQ

Q: Which card is faster in DirectX 12?

A: The Intel Arc A750 wins all DirectX 12 comparisons. It leads by 28.7% in 3DMark Steel Nomad DX12 (2612 vs 1862) and by 22.9% in PassMark DirectX 12 (70 vs 54).

Q: Does the AMD Radeon RX 5700 outperform the Arc A750 in any major benchmark?

A: Yes. The RX 5700 wins PassMark DirectX 11 (101 vs 72, a 40.3% lead), DirectX 10 (87 vs 65, 33.8%), DirectX 9 (209 vs 181, 15.5%), and PassMark G3D (14314 vs 12534, 14.2%).

Q: Which card has better compute performance?

A: The RX 5700 wins PassMark GPU Compute with 6517 against the A750’s 5368, a 21.4% margin. However, the A750 leads in OpenCL (98554 vs 69603) and Vulkan compute (85631 vs 64166).

Q: How do their overall average benchmark scores compare?

A: The RX 5700 averages 22170, which is 7.2% higher than the A750’s 20582. The RX 5700 also sits one percentile higher (67th vs 66th).

Q: What is the biggest performance gap in either direction?

A: The largest win is the RX 5700’s 40.3% advantage in PassMark DirectX 11. The largest Intel win is 29.4% in OpenCL, followed closely by 28.7% in 3DMark Steel Nomad DX12.

Q: Does the Arc A750 have more shading units than the RX 5700?

A: Yes. The A750 has 3584 shading units, 224 TMUs, and 112 ROPs, while the RX 5700 has 2304 shading units, 144 TMUs, and 64 ROPs.

Architecture Differences

The two cards come from fundamentally different architectures. The RX 5700 uses AMD’s RDNA 1.0 architecture on the Navi 10 chip, built on TSMC’s 7 nm process. The A750 uses Intel’s Xe-HPG architecture with the DG2-512 chip, manufactured on TSMC’s 6 nm node. This process difference is reflected in transistor density: the A750 packs 53.4 million transistors per mm² versus the RX 5700’s 41.0 million per mm². The Intel chip is also physically larger and denser, 21,700 million transistors on a 406 mm² die, compared to 10,300 million on 251 mm².

The A750 supports DirectX 12 Ultimate (12_2), while the RX 5700 is limited to DirectX 12 (12_1). This means the Intel card is designed for hardware ray tracing, and it includes 28 dedicated RT cores. The RX 5700 has no RT cores. Both cards support Vulkan 1.4 and OpenGL 4.6. The A750 also has a higher boost clock at 2400 MHz versus 1725 MHz, and its memory runs at 2000 MHz (16 Gbps effective) compared to 1750 MHz (14 Gbps effective) on the AMD card.

The memory subsystems differ in bandwidth despite identical capacity and bus width. Both have 8 GB of GDDR6 on a 256-bit bus, but the A750 achieves 512.0 GB/s versus 448.0 GB/s for the RX 5700. The A750’s theoretical compute throughput is more than double: 17.20 TFLOPS FP32 and 34.41 TFLOPS FP16, against 7.949 TFLOPS and 15.90 TFLOPS respectively. Pixel and texture rates follow the same pattern, the A750 reaches 268.8 GPixel/s and 537.6 GTexel/s, while the RX 5700 manages 110.4 GPixel/s and 248.4 GTexel/s.

The Verdict

The data points to a split decision based on workload. The Intel Arc A750 is the stronger choice for modern DX12 and Vulkan gaming, where it leads by 22.9% to 28.7%. Its hardware ray tracing support and DirectX 12 Ultimate compliance make it the more future-proof architecture for recent titles. The RX 5700, meanwhile, is the safer pick for legacy DirectX 9, 10, and 11 games, where it wins by 15.5% to 40.3%. It also has the higher overall average benchmark score (22170 vs 20582) and better PassMark G3D and GPU Compute results.

The A750’s hardware specification sheet is objectively superior, more cores, higher clocks, double the FP32 throughput, and 14.3% more memory bandwidth. Yet its average score is lower, suggesting driver or compatibility limitations that prevent it from converting raw specs into consistent wins. The RX 5700’s 67th percentile ranking against all GPUs, versus 66th for the A750, confirms that the AMD card edges out the Intel card in aggregate performance.

For a user prioritizing modern AAA titles with DX12 or Vulkan, the A750’s 25-29% lead in those APIs is decisive. For anyone with a library of older games or compute-heavy 2D workloads, the RX 5700’s 14-21% advantage in PassMark G3D, GPU Compute, and G2D makes it the better value in real-world usage. The RX 5700 also requires a lower suggested PSU (450 W vs 550 W) and has a lower TDP (180 W vs 225 W).

Specification Differences

| Field | AMD Radeon RX 5700 | Intel Arc A750 |

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

| Architecture | RDNA 1.0 | Xe-HPG |

| Process Node | 7 nm | 6 nm |

| Transistors | 10,300 million | 21,700 million |

| Die Size | 251 mm² | 406 mm² |

| Transistor Density | 41.0M / mm² | 53.4M / mm² |

| Base Clock | 1465 MHz | 2050 MHz |

| Boost Clock | 1725 MHz | 2400 MHz |

| Memory Clock | 1750 MHz (14 Gbps effective) | 2000 MHz (16 Gbps effective) |

| Memory Bandwidth | 448.0 GB/s | 512.0 GB/s |

| Shading Units | 2304 | 3584 |

| TMUs | 144 | 224 |

| ROPs | 64 | 112 |

| RT Cores | None | 28 |

| Pixel Rate | 110.4 GPixel/s | 268.8 GPixel/s |

| Texture Rate | 248.4 GTexel/s | 537.6 GTexel/s |

| FP32 | 7.949 TFLOPS | 17.20 TFLOPS |

| FP16 | 15.90 TFLOPS (2:1) | 34.41 TFLOPS (2:1) |

| TDP | 180 W | 225 W |

| Suggested PSU | 450 W | 550 W |

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

| Display Outputs | 1x HDMI 2.0b, 3x DisplayPort 1.4a | 1x HDMI 2.1, 3x DisplayPort 2.0 |

| Launch MSRP | 349 USD | 289 USD |

Where Each One Wins

The AMD Radeon RX 5700 is the winner in six benchmark categories: PassMark DirectX 10, DirectX 11, DirectX 9, G2D, G3D, and GPU Compute. Its 40.3% DirectX 11 lead is the standout result, and its 21.4% GPU Compute advantage makes it suitable for OpenCL-based productivity tasks that rely on legacy compute paths. The 14.2% G3D win and 19.8% G2D win indicate strong overall rasterization and 2D interface performance.

The Intel Arc A750 wins four benchmarks: 3DMark Steel Nomad DX12, OpenCL, Vulkan, and PassMark DirectX 12. Its 29.4% OpenCL victory and 25.1% Vulkan lead are substantial, and the 28.7% Steel Nomad margin is the largest single-test win for either card. The A750’s 22.9% DirectX 12 win completes a clean sweep of modern API workloads.

Use-case guidance follows directly from these splits. For DX12 and Vulkan gaming, the A750 is the clear pick, the data shows a 22.9% to 28.7% advantage across three separate tests. For DX9/10/11 gaming, the RX 5700 is unequivocally better, with margins ranging from 15.5% to 40.3%. In compute-heavy tasks, the choice depends on the API: OpenCL and Vulkan favor the A750, while PassMark GPU Compute favors the RX 5700. The A750’s higher memory bandwidth (512.0 GB/s vs 448.0 GB/s) and double FP32 throughput give it theoretical headroom, but the RX 5700’s superior average score (22170 vs 20582) and higher percentile rank (67 vs 66) suggest it remains the more consistent overall performer.