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.

TMS320C6674: TMS320C6674 Multicore Fixed and Floating-Point Digital Signal Processor

Part Number: TMS320C6674
Other Parts Discussed in Thread: AWR2944

Regarding
TMS320C6674 Multicore Fixed and Floating-Point Digital Signal Processor
Silicon Revision 1.0, 2.0 Silicon Errata
Literature Number: SPRZ341H
January 2011–Revised June 2015

 

Q1) Is the Advisory  27  relevant to us?

 

Q2) Wait for writeback completion
Are the following command sequences alternatives? Can we choose any of them?
Sequence 1: Issue an MFENCE instruction followed by Perform 16 NOPs
Sequence 2: Issue an MFENCE instruction followed by Issue another MFENCE instruction

  • Shahil,

    Let me look at the Literature, SPRZ341H and get back.

    Thanks for your patience.

    Regards

    Shankari G

  • Q1) Is the Advisory  27  relevant to us?

    When you ask, "relevant to us",  Have you stated the problem elsewhere or just asking whether it is relevant to the silicon revision  1.0, 2.0 ?

    Yes, Advisory 27 is relevant to the silicon revision 1.0, 2.0.

     ====

    Q2) Wait for writeback completion
    Are the following command sequences alternatives? Can we choose any of them?
    Sequence 1: Issue an MFENCE instruction followed by Perform 16 NOPs
    Sequence 2: Issue an MFENCE instruction followed by Issue another MFENCE instruction

    If you face the problem of "The MFENCE instruction is used to stall the instruction fetch pipeline until the completion of all CPU-triggered memory transactions.", use this work around of issuing two MFENCE.

    If you face the problem of  "Potential L2 Cache Corruption During Block Coherence Operations Issue" , use the work around of "Perform NOPS as given in the Advisory 7.

    ====

    Infact, you can use both the workarounds if needed... if ready to compromise  a minute  (my NOOT)  time lapse etc...

    Or check with one work around first and observe & measure that those particular issues are fixed.

    If not, try the other and check again.

    Regards

    Shankari G 

  • We are developing SW for AWR2944, which contains TMS320C66x DSP.

    What is the silicon revision of the TMS320C66x of AWR2944?

  • Shahil,

    I looped Kyle, our hardware expert.

    He will let you know about the silicon revision of  - TMS320C66x of AWR2944

    Regards

    Shankari G