Summary: | ASTERISK-03781: Using a precalculated tone is easier on the processor than calculating it mathmatically | ||
Reporter: | mtaht (mtaht) | Labels: | |
Date Opened: | 2005-03-26 18:07:14.000-0600 | Date Closed: | 2011-06-07 14:10:47 |
Priority: | Minor | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ||
Description: | The function playtones_generator could be optimized by a precompiled lookup table of tones (or it could cache each tone, after it's generated once). This routine is very hard on processors without an FPU. (on a 533 Mhz Xscale, for example, a single call to app_disa overwhelms the processor) I would argue that precalcuating and generating most common indications at compile time would be a net performance win on other processors as well. | ||
Comments: | By: Mark Spencer (markster) 2005-03-26 18:13:05.000-0600 Nope, the solution is to implement a fixed point interface like we have in the kernel for tone generation. See zaptel.c/tonezone.c |