• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 83 subscribers
  • Views 1346 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.

float to integer with rounding intrinsic

Lee Magnusson
Intellectual 280 points


I use some complier intrinsics in my code and they work well. I would also like to use the __f32toui16r intrinsic, but it is giving me an error. A simplified version of my code is:

  float a = 1.99;
  int16_t b = __f32toui16r(a);

I get this message:

INTERNAL ERROR: Illegal use of intrinsic:
                                   __f32toui16r

This may be a serious problem.  Please contact customer support with a
description of this problem and a sample of the source files that caused this

INTERNAL ERROR message to appear.

>> Compilation failure
Cannot continue compilation - ABORTING!

  • 0
    TI__Mastermind 20850 points

    Hi Lee,

    Are you trying this on the C-compiler(fpu32 support) or the CLA-C compiler? Also what version of the code gen tools are you using?

  • 0 in reply to Vishal_Coelho
    Prodigy 30 points

    I've got the same problem and receive the same error message. I am using C compiler and fpu32 support. The compiler version is the latest. v6.1.1 . The help file states that the input variable must be a double. I have tried both float and double variables. But no difference. Does anybody know the remedy?

    Regards 

  • 0 in reply to Hossein Sagha
    Prodigy 90 points

    I met same issue and get exactly same error message on fpu32 compiler.

    My code is as follow :

    static U16 doubleToUByte(register double dValue)
    {
      // Saturate to max unsigned byte value
      dValue = __fmin(dValue, 255.0);
      return __f32toui16r(dValue);
    }

    Do you have any idea to solve this error ?

    Best regards,

  • 0 in reply to remibonnet
    TI__Mastermind 20850 points

    Ok,

    I tried this out in two ways with CGT 6.2.0 and I am getting the correct assembly both times. You will see the C code interlisted with the assembly

    ;----------------------------------------------------------------------
    ; 101 | float a = 1.99;
    ;----------------------------------------------------------------------
    MOV *-SP[8],#47186 ; [CPU_] |101|
    MOV *-SP[7],#16382 ; [CPU_] |101|
    .dwpsn file "../Main.c",line 103,column 17,is_stmt
    ;----------------------------------------------------------------------
    ; 103 | unsigned int b = __f32toui16r(a);
    ;----------------------------------------------------------------------
    MOV32 R0H,*-SP[8] ; [CPU_] |103|
    F32TOUI16R R0H,R0H ; [CPU_] |103|
    NOP ; [CPU_]
    NOP ; [CPU_]
    MOV32 ACC,R0H ; [CPU_] |103|
    MOV *-SP[9],AL ; [CPU_] |103|

    Second, I tried calling your function:

    ;----------------------------------------------------------------------
    ; 104 | unsigned short c = doubleToUByte(a);
    ;----------------------------------------------------------------------
    MOV32 R0H,*-SP[8] ; [CPU_] |104|
    LCR #_doubleToUByte ; [CPU_] |104|
    ; call occurs [#_doubleToUByte] ; [] |104|
    MOV *-SP[10],AL ; [CPU_] |104|

    ;----------------------------------------------------------------------
    ; 86 | static unsigned short doubleToUByte(register double dValue)
    ;----------------------------------------------------------------------

    ;***************************************************************
    ;* FNAME: _doubleToUByte FR SIZE: 0 *
    ;* *
    ;* FUNCTION ENVIRONMENT *
    ;* *
    ;* FUNCTION PROPERTIES *
    ;* 0 Parameter, 0 Auto, 0 SOE *
    ;***************************************************************

    _doubleToUByte:
    ;----------------------------------------------------------------------
    ; 88 | // Saturate to max unsigned byte value
    ;----------------------------------------------------------------------
    ;----------------------------------------------------------------------
    ; 89 | dValue = __fmin(dValue, 255.0);
    ;----------------------------------------------------------------------
    MINF32 R0H,#17279 ; [CPU_] |89|
    ;----------------------------------------------------------------------
    ; 90 | return __f32toui16r(dValue);
    ;----------------------------------------------------------------------
    F32TOUI16R R0H,R0H ; [CPU_] |90|
    NOP ; [CPU_]
    NOP ; [CPU_]
    MOV32 ACC,R0H ; [CPU_] |90|
    LRETR ; [CPU_]

    I am unable to recreate the problem. Do you have optimization turned on?