This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Memory map for TMS320TCI6614 cortex A8

Hi,

I am working on the TMS320TCI6614 ARM cortex A8 and want to know how the memory is divided. Can anyone share the memory map for the same.

Thanks

Raunak

  • Hello Raunak,

    Please refer section 2.2 Memory Map Summary in the device data manual.

    www.ti.com/.../tms320tci6614.pdf

    Regards,
    Senthil
  • Hi,
    Is this for both the ARM as well as the DSP.

    I want to store a linked list in the shared memory so that ARM as well as DSP is able to access it. fro this I did make a link list through ARm and want to pass the address of the first node to the DSP for reading the data. How do I do it?

    Thanks
    Raunak
  • Yes, please refer to the datasheet already mentioned for the device memory map.  There is DDR and MSMC (faster access) available to all cores.  Have you considered using the Queue Manager? The QMSS (Multicore Navigator) provides a global hardware queuing system that is appropriate for certain linked lists.  Please search for SPRUGR9H on ti.com for the document.

  • Hi

    The QMSS can not be used because of some reasons. I made the queue to store packets coming in from Ethernet port and and want the DSP to get the address of the first node so that it could get the data and use it for further processing. How can I get the memory location of my 1st node and pass it to DSP.  I understand that ARM is not allowing me to write data to a particular address. 

    Thanks 

    Raunak

  • The ARMs have access to all global addresses, including DDR, and MSMC.  There are many ways to accomplish passing an address from the ARM to DSP. Knowing the best one to recommend for your situation is something I can't do without a lot more information.  Here are a few options:

    1) Create your own mailbox using a global (uncached) location(s) and a semaphore, interrupt, or global memory handshake.

    2) Simply push a descriptor to the QM from the ARM and have the DSP pop it from the queue.

    If the reason for pushing the address to the DSP is to provide a location for data to be processed by the DSP, it may be more efficient to have the ARM push it directly to the QMSS packetDMA, and let the DMA place it automatically in the local L2 memory of the DSP.

      -dave