• State
  • Replies 3 replies
  • Subscribers 99 subscribers
  • Views 52 views
  • Users 0 members are here
Support feedback
Options
Options
Related

AM6442: Question on GPMC Burst Mode Commands/Software Development

Michael Valentino-Manno
Prodigy 60 points
Part Number: AM6442


Hi,

I have received a couple general questions from a customer regarding writing code for their AM6442's GPMC peripheral. 

 

According to them, they are currently using GPMC in burst mode, are wondering the following:

  1. Is there an ideal way to write code for GPMC in burst mode--and more specifically, are there any issues with their current method of using the memcpy command?
  2. Are there method(s) better than using the memcpy command, and what are they? 

 

I apologize for the lack of concrete information. If you require information such as what type of memory device they're using, please let me know and I'll check in with them.

 

Thank you

Michael 

  • 0
    TI__Mastermind 47836 points

    Hello Michael,

    If the customer is interfaced with NOR or PSRAM memory over the GPMC.

    Then they can go with the MEMCPY API, which seems to be fine.

    Did the customer face any issue with the MEMCPY API ?

    Regards,

    Anil.

  • 0 in reply to Swargam Anil
    Prodigy 60 points

    Hello Anil,

    Thank you for the response.

    I asked my customer for clarification on the topic so I can provide more information.

    They are enabling the processor to see an FPGA as DPRAM.

    With this interface in mind, is the use of memcpy still fine, or is there a better method?

    They haven't identified any issues with the use of memcpy, but they want to know what the best, or at least what the generally recommended way to handle this is.

    Please let me know if you need any more information.

    Thanks,

    Michael 

  • 0 in reply to Michael Valentino-Manno
    TI__Mastermind 47836 points

    Hello Michael,

    If the MEMORY is mapped to SOC, then they can go with the use of MEMCPY and this is an optimized method rather than moving from GPMC memory to Memory or Memory to GPMC Memory in the for loop.

    One more point, when they are taking any throughput measurements , they need to do testing in the release build rather than in the debug build.

    Regards,

    Anil.