• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 98 subscribers
  • Views 530 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.

The DSP can't accomplish these algorithm in one sampling time?

Wang Yu 14
Prodigy 125 points

Hi, 

I have tested some basic audio data input and output routine using McASP transmit interrupt method on OMAPL137. For my application, I want to get audio data and output them after some filtering algorithm. The algorithm is simple just like some multiplication and addition operations. I have tried to get the calculation time using break point and CCS software. But after the test, I found that just the simple operation like below would cost about 31742 clocks. So the time for this procedure is about 31742/300MHz = 0.106ms and larger than the sampling time(1/48000Hz*1000 = 0.021ms), so the routine can't work well. But I'm wondering why such power DSP need so much time to work on this simple operation or there is something wrong with my understanding? So I'm looking forward for your reference and thanks for help!

for (i=0; i<length; i++)
{
yn += s[i] * x[i];
}

Best regards,

Wang Yu

  • 0
    TI__Mastermind 35645 points
    Hi,

    We are checking out your request and will update shortly.

    Thanks & regards,
    Sivaraj K
  • 0
    TI__Mastermind 35645 points

    Hi,

    Thanks for your post.

    Why don't you try installing c674x DSPLIB which has a set of DSP filtering kernels such as FIR, IIR, LMS adaptive filtering available "as-is"? May be, you could use such reference filtering algorithm  implementation available on DSP library single precision kernel functions.

    For C674x device, please check the supported kernel functions for C674x DSP_LIB from the below wiki:

    http://processors.wiki.ti.com/index.php/C674x_DSPLIB#DSPF_sp_fir_gen_.28FIR_Filter.29

    Kindly check the c64x+ DSP library programmer's reference as below:

    http://www.ti.com/lit/ug/sprueb8b/sprueb8b.pdf (please check Table 3 -5)

    For downloads of c66x, c67x, c64x, c64x+ & c674x DSP LIB, kindly refer the below link:

    http://www.ti.com/tool/sprc265

    After installing c674x DSP LIB., you could see the FIR filter example .CCS project in the below specified path:

    ~\dsplib_c674x_3_2_0_1\packages\ti\dsplib\src\DSPF_sp_fir_gen\c674\DSPF_sp_fir_gen_674LE_LE_COFF\

    Kindly refer the above example.

    Thanks & regards,

    Sivaraj K

    -------------------------------------------------------------------------------------------------------

    Please click the Verify Answer button on this post if it answers your question.

    -------------------------------------------------------------------------------------------------------

  • 0 in reply to Sivaraj Kuppuraj
    TI__Mastermind 35645 points

    Hi,

    Additional data points below:

    For fixed point computation, the C674x core is fully compatible with the C64x+ DSPLIB. So, to benchmark the cycle count measurement of c674x DSPLIB kernel functions, you could refer the below app. report:

    http://www.ti.com/lit/an/spra884a/spra884a.pdf

    Detailed comparisons of the kernels C674x and C67x implementations are listed in a development notes spreadsheet. This file is included in the docs folder of the C674x DSPLIB installation

    To benchmark the DSPLIB kernels, TI recommends the use of the C674x Cycle Accurate Simulator, which is included in Code Composer Studio 3.3 with Service Release 12 or later. After loading CCS, select Profile->Clock->Enable. This will allow the kernel demonstration apps to accurately display cycle counts.

    Thanks & regards,

    Sivaraj K

    -------------------------------------------------------------------------------------------------------

    Please click the Verify Answer button on this post if it answers your question.

    -------------------------------------------------------------------------------------------------------

  • 0 in reply to Sivaraj Kuppuraj
    Prodigy 125 points

    Dear Sivaraj K,

    Thank you very much for these suggestions and supports. I will have a try of these methods. 

    Best regards,

    Wang Yu