Hello (and a belated happy new year to all),

I am working on a custom hardware platform where I need to pull the Linux kernel from a Spansion SPI NOR flash device connected to a 600MHz TI AM3354 processor. The McSPI is configured to run at its maximum speed with a 48MHz clock and the datasheet for the S25FL164K states that the throughput for a read with a 50MHz clock rate should be 6.25MB/s.

Running the command "time sf read ${loadaddr} 0x00200000 ${loadsize}" gives the following output:

SF: 2900472 bytes @ 0x200000 Read: OK time: 1.676 seconds

This equates to around 1.7MB/s which I have managed to achieve principally by adding the OMAP3_MCSPI_CHCONF_TURBO flag to the channel configuration but also by moving things out of the read loop that only need to be performed once rather than for each byte read.

Does anyone have any idea as to whether there is more performance that can be got from this or is that about the limit?

Thanks,

Andy.