Qualcomm Oryon dominates AI, outperforming Intel and Apple by reaching 45 TOPS

Qualcomm is a well-known company in the smartphone and tablet industry, largely due to their high-performance Snapdragon chips that can compete with the best offerings from Apple and Android. Over a year ago, Qualcomm expressed interest in designing an SoC for laptops that would offer unprecedented performance for the brand. With the Snapdragon X Elite and its Qualcomm Oryon CPUs, this will be possible, offering better performance than Intel and Apple, particularly in the field of AI.

When it comes to choosing a laptop, most available options use x86 processors and integrated GPUs from Intel and AMD, as well as dedicated graphics if the laptop is more powerful. Laptops with Arm processors are relatively rare in the Windows ecosystem, with only a few featuring Snapdragon chips. In contrast, all Apple MacBooks since 2020 utilize Arm chips created by the company.

The Qualcomm Snapdragon X Elite with Oryon cores delivers an AI performance of 45 TOPS

When the Snapdragon X Elite was unveiled in late October, it became clear that the new processors with Oryon cores would have a significant impact on the market. These cores are developed by Qualcomm following the acquisition of Nuvia, setting them apart from traditional x86 PC processors and Arm CPUs like Apple’s. The Snapdragon X Elite features a 12-core Oryon SoC with a 3.8 GHz frequency when all cores are in use simultaneously and a 4.3 GHz boost for up to two cores.

Additionally, the SoC supports up to 64 GB LPDDR5X at 8,533 MHz and a GPU with an FP32 performance of 4.6 TFLOPs. With a bandwidth of 136 GB/s, it outperforms the Apple M1 chip by 36%. Qualcomm showed a performance comparison, highlighting how their Oryon chip surpasses all current laptop offerings in terms of AI processing. The AI performance of the Snapdragon X Elite is 45 TOPS, a 100-fold increase from what was achievable in 2017. According to Qualcomm, the laptop competition currently only achieves around 10 TOPS, making their solution significantly faster.

It’s faster than the i9-13980HX and Apple M3 on Geekbench

The AI performance is impressive, especially considering this SoC is designed for laptops with relatively low power consumption. It’s worth noting that unlike Qualcomm’s mobile chips, Oryon cores are all high-performance rather than a mix of efficient cores. Moving beyond AI performance and focusing on CPU performance, a Geekbench 6 comparison showed the Oryon chip narrowly surpassing the i9-13980HX in single-core tests, scoring 3,227 points while consuming 70% less energy.

Apple’s M3 chip is also outperformed in multi-core tests, where it scores 12,154 points compared to Qualcomm Oryon’s 15,300 points. As seen, Qualcomm’s chips are impressive, but there’s a potential issue with software compatibility. Few Windows ISV applications have been recompiled to work with Arm, with the majority being dominated by Microsoft. Consequently, Qualcomm’s biggest challenge is to ensure optimal performance in various software applications, but if the chip becomes popular and proves its superiority over Intel, AMD, and Apple, changes will likely follow.