The strength of Apple Silicon comes from its processing capabilities, not primarily its memory.
The strength of Apple Silicon comes from its processing capabilities, not primarily its memory.
I’m unsure about the reason behind that speed, but it seems tied to how memory and tech are arranged. That’s top-tier performance, and I expect Intel and AMD to push forward with innovation. Intel has let us down, so it’s disappointing. Yeah, yeah—those MTL sounds are annoying. If it’s not the CPU, the integrated graphics will need more bandwidth all day, which still counts as performance, right? How can they ignore such clear advantages in the competition?
Memory bandwidth plays a role in certain applications, though not universally. To achieve such high speeds, systems increased the bus width. Standard x86 computers typically use a 128-bit bus for dual-channel memory. Bandwidth requirements vary by application—some are memory-independent while others depend on it. Epyc benchmarks provide insights into channel counts; you can find details at https://www.phoronix.com/review/ddr5-epyc-9004-genoa, which suggests potential performance exceeding 400GB/s with all 12 channels active.
Consider how the performance is realized and what it implies for building your own PCs. If Intel and AMD integrated the memory into the CPUs themselves, you'd have to purchase a CPU with the desired memory capacity right from the start. This would lock memory upgrades unless you also upgrade the CPU. It would mean that manufacturers control how much RAM each CPU tier should have, rather than letting OEMs or DIY builders decide it. The decision on whether to provide more memory depends on cost rather than benefit. Unless your project falls into the rare category where memory speed is the main bottleneck, boosting bandwidth doesn’t help much—it’s essentially hitting diminishing returns. Regarding integrated graphics, as long as you can add a dedicated GPU, you don’t need an expensive one built into the CPU. For APU setups, it might be useful but remains limited to specific budget builds, often chosen for cost reasons rather than performance. Changing a CPU to include memory in the package would be costly and unlikely to reach affordable prices.
Opt for a larger bus to capture most advantages of increased bandwidth. AMD is choosing this approach for their Strix APUs, moving away from the outdated 128-bit dual-channel setup that many devices still use.
This device remains significantly superior to x86 processors.