Intel Arc A350M

The Intel Arc A350M is a mobile integrated GPU from Intel's Alchemist generation, built on the Xe-HPG architecture and fabricated on TSMC's 6 nm process. The chip packs 7,200 million transistors into a 157 mm² die, with a transistor density of 45.9M per square millimeter. In aggregate benchmarks, it averages 24,647 points across Geekbench OpenCL and Vulkan tests, placing it in the 68th percentile of all GPUs. Its nearest rivals are all within a fraction of a percent, making this a tightly contested part in the low-power segment.

Benchmark Performance

The Arc A350M delivers a Geekbench OpenCL score of 24,546 and a Geekbench Vulkan score of 24,747, yielding an average of 24,647. The Vulkan result is slightly higher than the OpenCL score, indicating solid driver optimization for that API. Against its nearest rivals, the A350M edges out the NVIDIA GeForce RTX 5060 Mobile by 0.2% and the NVIDIA Quadro RTX 5000 by 0.5%, while trailing the AMD Radeon RX 5700 XT by 0.3% and the NVIDIA RTX A5000 Mobile by 0.5%. These deltas are negligible—well within run-to-run variance—so the A350M effectively trades blows with all four competitors. The 68th percentile ranking means it outperforms roughly two-thirds of all GPUs in the benchmark database, a strong showing for a 25 W integrated part. The data suggests that in synthetic compute workloads, the A350M is indistinguishable from much larger discrete cards, though real-world gaming performance depends on other factors like memory bandwidth and driver maturity.

Who Should Consider It

The Arc A350M is designed for portable devices, as its display outputs are "Portable Device Dependent" and its slot width is IGP (integrated GPU). With a 25 W TDP, it targets thin-and-light laptops where power efficiency is paramount. The memory subsystem—4 GB of GDDR6 on a 64-bit bus with 112.0 GB/s bandwidth—indicates that this GPU is best suited for 1080p gaming at moderate settings. At 1440p or higher, the limited capacity and bandwidth could become bottlenecks, particularly in texture-heavy titles. For esports and older games, the A350M's 3.379 TFLOPS of FP32 performance and 105.6 GTexel/s texture rate are ample. However, modern AAA games at high detail may exceed the 4 GB VRAM buffer, forcing lower resolutions or reduced texture quality. The card's 24 ROPs and 52.80 GPixel/s pixel rate further reinforce its 1080p orientation. Users who prioritize portability and modest gaming over maximum visual fidelity will find the A350M adequate.

Power and Cooling

The A350M has a TDP of 25 W, which is exceptionally low for a GPU with this level of performance. Its slot width is listed as IGP, meaning it is integrated into the motherboard or system-on-chip layout rather than a discrete add-in card. Consequently, no power connectors are required—the card draws all its power from the system's existing power delivery. The absence of a suggested PSU in the specifications confirms that it is not intended for desktop builds. The low power draw allows for passive or low-profile cooling solutions in compact chassis. The bus interface is PCIe 4.0 x8, which provides adequate bandwidth for the GPU's memory and compute needs without the overhead of a full x16 link. The 25 W envelope also means that thermal management is straightforward, even in fanless designs. The production status is end-of-life, indicating that this part is no longer actively manufactured, but it remains a relevant data point for comparison in the low-power mobile segment.

How It Compares

NVIDIA GeForce RTX 5060 Mobile: The A350M is 0.2% faster in average benchmark score, effectively a statistical tie. Both GPUs occupy the same performance tier, though the A350M achieves this with a far lower power envelope and integrated form factor.

AMD Radeon RX 5700 XT: The A350M trails the RX 5700 XT by 0.3%, a margin too small to be meaningful in real-world use. The RX 5700 XT is a desktop discrete card, while the A350M is a mobile integrated part, making the near-parity proof of the efficiency of Intel's Xe-HPG architecture.

NVIDIA RTX A5000 Mobile: The A350M is 0.5% slower than the RTX A5000 Mobile. This professional mobile GPU typically commands a premium for its feature set, but in raw compute benchmarks the two are nearly indistinguishable. The A350M's lower power draw and integrated nature make it a more practical choice for consumer ultraportables.

NVIDIA Quadro RTX 5000: The A350M is 0.5% ahead of the Quadro RTX 5000. Like the RTX A5000, the Quadro RTX 5000 is a professional-grade card, yet the A350M manages to edge it out in average score. This highlights that for compute-bound tasks, the A350M holds its own against much more expensive hardware.

Ray Tracing and Feature Set

The Arc A350M includes 6 ray tracing cores, enabling hardware-accelerated ray tracing. It supports DirectX 12 Ultimate (12_2), which is the latest DirectX feature level, along with Vulkan 1.4 and OpenGL 4.6. The card has 768 shading units, 48 texture mapping units, and 24 ROPs. Its pixel rate is 52.80 GPixel/s and its texture rate is 105.6 GTexel/s. FP32 performance is 3.379 TFLOPS, while FP16 performance reaches 6.758 TFLOPS at a 2:1 ratio. Notably, the specification does not list tensor cores, meaning there is no dedicated AI acceleration hardware. As a result, features that rely on tensor cores—such as AI-based upscaling—are not supported, though the card's ray tracing capabilities are present. The combination of DirectX 12 Ultimate and Vulkan 1.4 ensures broad API compatibility for modern games and applications. The 6 RT cores are modest in count, but they provide a baseline for ray-traced effects at reduced resolutions.

FAQ

Q: What is the average benchmark score of the Intel Arc A350M?

A: The average benchmark score is 24,647 points, based on Geekbench OpenCL (24,546) and Vulkan (24,747) tests.

Q: How does the A350M compare to the NVIDIA GeForce RTX 5060 Mobile?

A: The A350M is 0.2% faster than the RTX 5060 Mobile in average benchmark score, making the two effectively equal.

Q: What is the memory configuration of the A350M?

A: It has 4 GB of GDDR6 memory on a 64-bit bus, providing 112.0 GB/s of bandwidth.

Q: Does the A350M support hardware ray tracing?

A: Yes, it includes 6 ray tracing cores and supports DirectX 12 Ultimate (12_2), which includes ray tracing features.

Q: What is the TDP of the A350M?

A: The TDP is 25 W, making it suitable for power-constrained portable devices.

Q: Is the A350M still in production?

A: No, it is end-of-life, with a release date of 2022-03-29.

Memory Subsystem

The Arc A350M is equipped with 4 GB of GDDR6 memory on a 64-bit bus, yielding a bandwidth of 112.0 GB/s. This configuration is modest by modern standards. The 4 GB capacity is sufficient for 1080p gaming with standard textures, but it becomes a limiting factor at higher resolutions or with high-resolution texture packs. The 64-bit bus width restricts the amount of data that can be transferred per clock cycle, and the 112 GB/s bandwidth is about half of what many mid-range discrete GPUs offer. In practice, this means that the A350M may experience stuttering or reduced frame rates in memory-intensive scenes, particularly at 1440p or 4K. The memory clock is 1750 MHz (14 Gbps effective), which is standard for GDDR6. For the intended use case of 1080p gaming on ultraportable laptops, the memory subsystem is adequate, but it clearly delineates the card's performance ceiling. Users seeking higher resolution or virtual reality applications would need a GPU with more VRAM and wider bus. The 4 GB capacity also limits the use of advanced texture filtering and anti-aliasing techniques that consume additional memory. Overall, the memory subsystem is the primary bottleneck for the A350M beyond 1080p, and any serious gaming at higher resolutions would require a more robust solution.