Timeline for Help? SPI communication slow
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 18, 2015 at 14:34 | vote | accept | Paul Terwilliger | ||
| Aug 18, 2015 at 14:34 | |||||
| Aug 18, 2015 at 6:21 | answer | added | Lokanath | timeline score: 1 | |
| Aug 17, 2015 at 21:36 | comment | added | fceconel | I'm afraid it'll be hard to circumvent the performance bottleneck; it may be due to the fact that there's a system call to be made (ioctl), which for many reasons is a time-consuming operation, or something in the driver code itself. You'd probably need to interface directly to the SPI port (and maybe this needs to be done in a device driver), which is no trivial task; I have no experience with Yocto or the SPI drivers for the Edison, so I don't think I can offer much further help. | |
| Aug 17, 2015 at 18:41 | comment | added | Paul Terwilliger | @fceconel I found that the time to set up the SPI port over and over is eating a lot of time. I added the results to my post. Would you be able to help me with this? | |
| Aug 17, 2015 at 18:40 | history | edited | Paul Terwilliger | CC BY-SA 3.0 |
added 453 characters in body
|
| Aug 17, 2015 at 16:59 | comment | added | Paul Terwilliger | @fceconel I will try that and post results, thanks! | |
| Aug 17, 2015 at 16:50 | history | edited | Paul Terwilliger | CC BY-SA 3.0 |
Added all of my code
|
| Aug 17, 2015 at 16:32 | comment | added | fceconel | I'd suggest to try mraa_spi_write_buf, doing a large enough block transfer and comparing the results. Possibly the time to setup the SPI port over and over to transfer each byte is eating a lot of time. | |
| Aug 17, 2015 at 16:17 | comment | added | Paul Terwilliger | I have omitted the initialization code. I can add that. | |
| Aug 17, 2015 at 16:08 | comment | added | Oleg Mazurov | where is the piece that initializes SPI peripheral? | |
| Aug 17, 2015 at 16:00 | comment | added | MathieuL | What is the data rate? | |
| Aug 17, 2015 at 15:51 | history | asked | Paul Terwilliger | CC BY-SA 3.0 |