News for hackers. Stuff that matters.
Re:Exploiting Coroutines to Attack the “Killer Nanoseconds” ... by brian-armstrong on Sunday November 11, 2018 @02:45AM (#18425407)
This is a very cool idea, and I wonder if eventually we'll see new tooling spring up to support this.

Running coroutines like this seems like it would potentially create more cache pressure. Is the idea that you'll execute instructions you already have loaded, so you won't incur any cache misses for instructions?

Re:Exploiting Coroutines to Attack the “Killer Nanoseconds” ... by yxhuvud on Sunday November 11, 2018 @04:10AM (#18425566)
It is more about memory pressure than instruction pressure - the idea is that when an algorithm finds itself having to fetch something from memory, it can issue a prefetch, then take a step back and handle other requests while waiting for the result.

How this would fare in a real world scenario that isn't a benchmark but in a system that is also doing other stuff at the same time is anyones guess though.

Re:Exploiting Coroutines to Attack the “Killer Nanoseconds” ... by tuukkah on Sunday November 11, 2018 @06:56AM (#18425983)
Their benchmark is close to what a relational database does in the real world. They also cite previous work "Interleaving with Coroutines: A Practical Approach for Robust Index Joins", which implements this succesfully in the SAP HANA column store.

I don't think it matters what else the system is doing as long as a complete core is available for running this.