• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 53 subscribers
  • Views 1826 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

CCS/EVMK2H: CCS 7.2.0.00013 Memory Throughput Analysis Hardware Trace Analyser captures from the wrong transaction monitors

Chester Gillon
Guru 92251 points

Part Number: EVMK2H

Tool/software: Code Composer Studio

Was using CCS 7.2.0.00013 under Windows 10, with TI Emulators 7.0.48.0 and Keystone2 1.1.9 to attempt to use the Memory Throughput Analysis Hardware Trace Analyser, when a program was running on the C66xx_0 core of a EVMK2H accessing DDR memory. Used the XDS200 Onboard Debug Probe on the EVMK2H, with the boot mode set to "DSP No-Boot" and the ..\..\emulation\boards\xtcievmk2x\gel\xtcievmk2x.gel GEL script used to perform the device configuration.

The Memory Throughput Analysis was started using the default values. The "Memory Throughput - CSSTM_0" graph wasn't displaying any data:

Looking at the "Trace Viewer - C66xx_0" table shows that the CFG and SPI_ROM_EMIF16 domains have been captured rather than the expected DDR3A and DDR3B domains for the Memory Throughput Analysis:

Checking the Analysis Properties from the "Trace Viewer - C66xx_0" shows that DDR3A and DDR3B have been set as the Transaction Monitors:

If the same test is repeated using CCS 7.1.0.00016  with TI Emulators 6.0.628.1 and Keystone2 device support 1.1.9 then the Memory Throughput Analysis captures data for the DDR3A and DDR3B domains:

Therefore, in the change from CCS 7.1.0.00016 / TI Emulators 6.0.628.1 to CCS 7.2.0.00013 /  with TI Emulators 7.0.48.0 something has caused the Memory Throughput Analysis to capture data from the incorrect transaction monitors.

The example project used is attached 66AK2H14_C66_system_trace_TCI6638K2K_device_file.zip

[In the Target Configuration the device has been set to a TCI6638K2K rather than 66AK2H12 since the 66AK2H12 device file is missing some entries for trace - see https://e2e.ti.com/support/development_tools/code_composer_studio/f/81/t/511744]

  • 0
    TI__Guru**** 170244 points
    Chester,

    Thanks for sending the detailed test case. I am trying to get a K2H board so I can run it here.

    Thanks for the patience,
    Rafael
  • 0 in reply to desouza
    TI__Guru**** 170244 points

    Chester,

    Please apologize for the delay; I found out a related thread where some bus masters were incorrectly configured and then fixed in the latest incarnation of the Emupack software. This fix may have broken the assignment to the bus masters. 

    The files changed are:

    - ccsv7/ccs_base/emulation/analysis/xmldb/aet_config/AET_PropertyCP_Tracer_kepler.xml

    - ccsv7/ccs_base/emulation/analysis/xmldb/trace_config/devices/device_kepler.xml

    I found out the changes that were made to these files but I am trying to make some sense of the modifications. I will report back as soon as I get something more definitive. 

    I apologize for the inconvenience,

    Rafael

  • 0 in reply to desouza
    TI__Guru**** 170244 points

    Chester,

    Today the bug number DBGTRC-3613 was filed and the issue was found. A fix will be made available officially on the next major emupack release (maybe November). 

    I am trying to get the modified XML file to drop into an existing install.

    Regards,

    Rafael

  • 0 in reply to desouza
    TI__Guru**** 170244 points

    Chester,

    Please check attached. The file must be dropped into the directory:

    ccsv7/ccs_base/emulation/analysis/xmldb/aet_config/

    Hope this helps,

    Rafael

    AET_PropertyCP_Tracer_kepler.zip

  • 0 in reply to desouza
    Guru 92251 points

    desouza said:
    Please check attached. The file must be dropped into the directory:

    Thanks for the updated file.

    I don't have the EVMK2H to test the updated file at the moment, but comparing the updated file against that in a CCS 7.2 installation shows that Transaction Monitor enumerations have been changed, so the changes look reasonable.

    [In the CCS 7.2 version of the least significant 8 bits of the enumerations match the slave IDs in Table 11-59. MSTID Mapping for Hardware Instrumentation (CPTRACERS) of the 66AK2H14 datasheet SPRS866F, and in the updated file the mapping of enumeration values to Transaction Monitor names has changed]

  • 0 in reply to Chester Gillon
    TI__Guru**** 170244 points

    Chester,

    Just FYI, I was able to get Memory Throughput data using the provided file on my 66AK2H12 board. 

    Regards,

    Rafael