• State
  • +1 person also asked this people also asked this
  • Locked Locked
  • Replies 4 replies
  • Subscribers 101 subscribers
  • Views 233 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

AM68A: Compilation not work properly on SDK 9.1

AM68A: status on compilation bug

Roman Vei
Prodigy 50 points

Part Number: AM68A

Hey,
I wanted to check a status on a related bug: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1307531/am68a-compilation-not-work-properly-on-sdk-9-1

It should be fixed for 9.2 and here following ticket: jira.itg.ti.com/.../TIDL-3572

Does anybody have status on this bug and can confirm that it is resolved?

  • 0
    TI__Mastermind 19670 points

    Hi Roman,

    Due to a holiday, half of our team is currently out of the office. Please expect a 1~2 day delay in response from the engineer tracking this issue.

    Thank you,

    Fabiana

  • 0 in reply to Fabiana Jaimes
    TI__Mastermind 24041 points

    Hi,

    I have initiated loop back on above mentioned jira, will get back to you once i have update.

    In the mean time, is it possible for you to try the same experiment on 9.2 ? 

    Thanks

    Pratik

  • 0 in reply to Pratik Kedar
    Prodigy 50 points

    Yes, I tried it and can say that array mismatch is the same:

    Sub Graph Stats 64.000000 250866.000000 630.000000
*******  TIDL_subgraphRtInvoke done  ********
Error margin: 0.1
The arrays are 12.66% similar.
Error margin: 0.01
The arrays are 1.32% similar.
Error margin: 0.001
The arrays are 0.13% similar.
Error margin: 0.0001
The arrays are 0.01% similar.
Error margin: 0.1
The arrays are 11.33% similar.
Error margin: 0.01
The arrays are 1.05% similar.
Error margin: 0.001
The arrays are 0.12% similar.
Error margin: 0.0001
The arrays are 0.01% similar.
Error margin: 0.1
The arrays are 8.90% similar.
Error margin: 0.01
The arrays are 0.89% similar.
Error margin: 0.001
The arrays are 0.11% similar.
Error margin: 0.0001
The arrays are 0.01% similar.
Error margin: 0.1
The arrays are 8.96% similar.
Error margin: 0.01
The arrays are 0.89% similar.
Error margin: 0.001
The arrays are 0.07% similar.
Error margin: 0.0001
The arrays are 0.01% similar.
Error margin: 0.1
The arrays are 8.55% similar.
Error margin: 0.01
The arrays are 0.91% similar.
Error margin: 0.001
The arrays are 0.10% similar.
Error margin: 0.0001
The arrays are 0.01% similar.

    The docker built with sdk 9.2 (github.com/.../rel_09_02) and I can observe different behaviour with DataOps: 0 and 1 (same bug as in 9.1)

  • 0 in reply to Roman Vei
    TI__Mastermind 24041 points

    I see,

    Error margin: 0.1

    The arrays are 12.66% similar.

    Correlations percentage seems to be poor, data_ops to 0 (arm offload).

    Thanks for confirmation, will update the JIRA with the same and try to get back to you once i have an update on this.