Timeline for How to output as fast as possible a fixed buffer?
Current License: CC BY-SA 4.0
22 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 3, 2019 at 9:58 | vote | accept | Zibri | ||
| Mar 31, 2019 at 10:17 | answer | added | Zibri | timeline score: 5 | |
| Mar 31, 2019 at 10:08 | history | reopened |
mao Tanktalus slfan Vignesh Pichamani Zibri |
||
| Mar 30, 2019 at 10:58 | comment | added | Zibri | Please, @oguzismail remove the hold.. it is clear why I asked that.. infact people is trying to answer. | |
| Mar 30, 2019 at 10:56 | comment | added | Zibri | ok.. but did any of you compile my code and then his "version" and found that his version is faster on his machine? | |
| Mar 29, 2019 at 13:52 | comment | added | SKi | @Zibri sendfile() is a system call like write(), so it won't use 'high level' c-lib function like fwrite(). | |
| Mar 29, 2019 at 9:27 | history | edited | Zibri | CC BY-SA 4.0 |
clarified in BOLD the reason of the question.
|
| Mar 28, 2019 at 18:00 | review | Reopen votes | |||
| Mar 28, 2019 at 21:27 | |||||
| Mar 28, 2019 at 17:43 | comment | added | Zibri | I updated the source raising the scheduling priority of the threads. Also you must run it with sudo nice -n -20 to have full speed. | |
| Mar 28, 2019 at 17:41 | history | edited | Zibri | CC BY-SA 4.0 |
added 271 characters in body
|
| Mar 28, 2019 at 17:28 | history | closed |
oguz ismail Achal Support Ukraine Mathieu double-beep |
Needs details or clarity | |
| Mar 28, 2019 at 16:20 | history | edited | Zibri | CC BY-SA 4.0 |
added 103 characters in body
|
| Mar 28, 2019 at 16:15 | comment | added | Zibri | @SKi write is slower because with a so small buffer is better to use fwrite.. in all my tests any modification to the above code makes it slower... things get better when I raise the process priority and scheduler to Realtime and -20... but still is the pipe the bottleneck...sendfile afaik uses fwrite internally...about vmsplice I don't know,,, it expects an array of iovec structures...hmm | |
| Mar 28, 2019 at 12:42 | answer | added | Maxim Egorushkin | timeline score: -1 | |
| Mar 28, 2019 at 12:35 | answer | added | mevets | timeline score: 1 | |
| Mar 28, 2019 at 11:36 | comment | added | SKi | Did you already tried "write(STDOUT_FILENO, ...)" ? And are functions like sendfile() and vmsplice() suitable? | |
| Mar 28, 2019 at 11:35 | comment | added | Jabberwocky | You should make that clear in the question. | |
| Mar 28, 2019 at 11:34 | comment | added | Zibri | my curiosity is to understand how much data can a single program (mutithreaaded or multiprocess) output. | |
| Mar 28, 2019 at 11:32 | comment | added | Zibri | @Jabberwocky I could use pthread.. but there is no speed improvement.. the bottleneck seems to be the pipe. | |
| Mar 28, 2019 at 11:08 | comment | added | Jabberwocky |
BTW fork doesn't create a thread. Not my DV/CV though.
|
|
| Mar 28, 2019 at 10:45 | review | Close votes | |||
| Mar 28, 2019 at 17:28 | |||||
| Mar 28, 2019 at 10:22 | history | asked | Zibri | CC BY-SA 4.0 |