Timeline for Can multi-threading improve performance of an IO-bound process?
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 2, 2024 at 20:43 | history | edited | Christophe | CC BY-SA 4.0 |
added 176 characters in body
|
| Aug 2, 2024 at 16:49 | comment | added | JimmyJames | @chepner That's true, it was just one example exactly as I intended as that's what is appropriate for a comment. | |
| Aug 2, 2024 at 15:59 | comment | added | chepner | @JimmyJames Distance-induced latency is just one example. Your I/O request could be to a machine in the next room, but that request may involve a lengthy database query which doesn't in and of itself prevent your machine from doing something else while it runs. | |
| Aug 2, 2024 at 11:03 | comment | added | Christophe | @IanGoldby Indeed. No longer need to write assembler routines, making sure that the code is reentrant, in case of multiple interrupts interrupting each other. The first native language support in this regard was ADA83 if I remember well, followed closely by Modula-2 before all mainstream languages followed. | |
| Aug 2, 2024 at 9:43 | comment | added | Ian Goldby | "But the latter part is very difficult to design and even more to implement." - that used to be the case but modern languages and frameworks provide asynchronous operations, making it now arguably simpler than threads (and more performant). | |
| Aug 1, 2024 at 17:27 | comment | added | JimmyJames | Nice article. I'd like to see the numbers given in consistent units, though. I'm not sure it's completely obvious that an IO trip from the USA to India and back is more than billion times slower than an L1 cache reference. | |
| Aug 1, 2024 at 17:21 | history | edited | Christophe | CC BY-SA 4.0 |
added 308 characters in body
|
| Aug 1, 2024 at 17:05 | history | answered | Christophe | CC BY-SA 4.0 |