I've heard the terminology "hyper-threading" before, I believe back in 213 (2.5 years ago). Is this kind of multithreading all that means? Or does it do any other significant things under the hood? I'm trying to think of if hiding memory read latency is enough to actually make it seem like there is a 1.5-2x speed-up...
I've heard the terminology "hyper-threading" before, I believe back in 213 (2.5 years ago). Is this kind of multithreading all that means? Or does it do any other significant things under the hood? I'm trying to think of if hiding memory read latency is enough to actually make it seem like there is a 1.5-2x speed-up...
This is discussed in detail on slide 53.