The LPDDR6 memory standard might be supported by competing shipments like the Snapdragon 8 Elite Gen 6 Pro, but Apple isn’t about to turn away from its tradition of adopting technological standards much later. However, just because the A20 Pro won’t be treated to faster RAM doesn’t make it less capable, not even by a stretch, as fresh details reveal that the SoC can easily handle the most demanding of AI workloads thanks to two major changes.
With on-device AI models being memory-bound, a radical shift was necessary for the A20 Pro, with the existing LPDDR5X RAM also operating at a 96-bit bus width instead of 64
Long before the massive Tata leak spilled a plethora of details, we reported that the iPhone 18 Pro and iPhone 18 Pro Max would feature six-channel memory necessary for tackling AI tasks, as it would increase the A20 Pro’s maximum bandwidth. Now, Reptalica has been kind enough to share valuable details in the form of schematics that show Apple’s first 2nm SoC will not only support six-channel LPDDR5X RAM but also operate at a 96-bit bus width.
This change marks a departure from Apple’s older approach, which was tied to 4-channel memory and a 64-bit bus width. We’ve also discussed that the proliferation of AI compelled Apple to ditch the older inFO-PoP packaging in favor of Wafer-Level Multi-Chip Module Packaging (WMCM) as it’s more suited to handling high data volumes necessary for smoother AI operations.
On-device AI models don’t just require better processing, but a narrow bus width could result in a massive bottleneck
In addition to the A20 Pro sporting a bigger Neural Engine, the 96-bit LPDDR5X RAM support means the processor won’t sit around idle trying to wait for data to come through. A narrow bus width would have introduced this problem, but it no longer does. It’s also worth noting that the six-channel 96-bit bus width requires a larger memory footprint, which explains why the A20 Pro has transitioned to the WMCM packaging.
Furthermore, with the DRAM kept separate from the A20 Pro die, the latter’s heat dissipation will improve, leading to better sustained performance and enabling AI workloads to be handled comfortably. While it’s disappointing that there’s no LPDDR6 RAM support, there are other ways to boost memory bandwidth and performance, and Apple has proved that.
News Source: Reptalica
About the author: Omar Sohail is a reporter and analyst for Wccftech's mobile section, specializing in the technology and business of the mobile industry. His expertise lies in the intricate hardware supply chain, covering developments in semiconductor manufacturing, chip lithography, and camera sensor technology.
Follow Wccftech on Google to get more of our news coverage in your feeds.
