What is AMD 3D V-Cache and why is it so special?

With the release of the AMD 9800X3D impressing the internet at large, there has been much talk of the new iteration of the 3D V-Cache that it utilizes, and how it’s likely the main contributor to the big performance increase this CPU brings.

But what is a V-Cache? Why is it 3D? Are normal caches two dimensional? Let’s explain. But first we must cover the basics. What is a cache?

When it comes to CPUs, “cache” refers to a very small amount of volatile memory that’s actually in the CPU package itself. Space is at a premium here, which is why the capacity is so small as compared to system RAM. The trade off being that, due to it being physically much closer to the CPU, it can be accessed far more quickly than RAM.

There are usually multiple tiers of the cache: L1, L2, and L3.

L1 is closest to the CPU and has the smallest capacity, L2 is further, larger, and slower, L3 further, larger and slower still.

Of course, when we say “Slower”, it’s still orders of magnitude faster than RAM, which is then some more orders of magnitude faster than non-volatile solid-state storage (your SATA or M.2 SSD for example). L1 cache can work something like 200x faster than super quick DDR5 RAM.

So, that’s what a cache is, but what makes AMD’s “3D”?

This is AMD’s technology that stacks additional layers of L3 cache on top of each other, creating more capacity. Second generation 3D-Cache differs slightly as these stacks of extra memory are directly below the processor cores. This means that they have better access to the IHS, which means they are cooled more efficiently, and can perform faster as a result.