John Siracusa

When we talk about like power cores versus efficiency cores, like what do we actually mean on the M series chip?

What is it about the P cores that makes them powerful?

How are they different than the E cores and all the other lines of chips?

Well, if you look at the die shot, you can say, well, first of all, I can see that the P cores are literally bigger.

Like they're a bigger square on the chip.

Why are they bigger?

What is in that thing that makes them more powerful?

And there's a bunch of answers to that.

There could be and probably is more execution units.

How many parts of the chip can add numbers together?

You can have two adders, three adders, four adders.

How many units are there for floating multiplication?

Like you can literally put more execution units, more things that can munch numbers.

You can add more of them because it's parallel dispatch.

Like it's not just one line through.

So that's a way you can make them bigger.

Another way you can make them bigger is you can add more cache, more L1 cache and stuff like that.

That takes up die space.

That's what makes the power cores bigger.

You can have a more sophisticated branch predictor.