Hi Experts,
There is a define " SPI_MULT_ICOUNT_SUPPORT" in mibspi_dma.c under C:\ti\mmwave_sdk_03_02_00_04\packages\ti\drivers\spi\src.
/* Support for multi icount in one transfer to achieve high throughput , this is only supported in blocking mode */
/* #define SPI_MULT_ICOUNT_SUPPORT */
Question:
1. If the define is enabled, is it possible to send the SPI data faster than default APIs.
2. Is there any limitation? If there is anything I have to take care once it is enabled, please let me know.
3. Is there any alternative resolution to achieve faster SPI transfer?
I am thinking a faster SPI (not QSPI) transfer.
For example, approx. 1MB data is transferred continuously.
The current driver can send the date up to 64KB.
So it takes an overhead time approx. 30us for each data transfer to wait the response from the API.
I would like to get rid of it.
Thank you for your kind advice.
Best regard,
HItoshi