• Locked Locked
  • Replies 3 replies
  • Subscribers 119 subscribers
  • Views 2804 views
  • Users 0 members are here
2025 survey

We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.

Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

need some help to translate some ARM asm

Vincent R
Note: This was originally posted on 10th August 2011 at http://forums.arm.com

Hi,

I am a C developper and I am trying to understand how works a function defined as shown below  but I have some
difficulties with floating point function and especially with FMULD and FMACD.
I tried to read the documentation but I don't really understand what a vector is so for now I have this :



struct AffineTransform {
  float a, b, c, d;
  float tx, ty;
};

void __cdecl PathMoveToPoint(MutablePath* path, const AffineTransform *m, float x, float y)
{
  ...
__text:00015514                MOV  R4, R0                  ;R4 = path
__text:00015516           FMSR S12, R2                  ;float S12 = x
__text:0001551A           FMSR S14, R3                  ;float S14 = y
__text:0001551E           CMP  R0, #0                  ;if (!path) return;
__text:00015520           BEQ  loc_155E2              ;if (m != null) goto loc_155E2
...
[size="2"]__text:0001552E           FCVTDS  D3, S12              ;double D3 = (double) x
__text:00015532                FCVTDS  D5, S14              ;double D5 = (double) y
__text:00015536           FLDS S14, [R1]              ;float S14 = m->a
__text:0001553A              FCVTDS  D6, S14                ; double D6 = (double) m->a
__text:0001553E           FLDS S14, [R1,#8]        ;float S14 = m->c
__text:00015542                FCVTDS  D7, S14                ;double D7 = (double) m->c
__text:00015546                FMULD   D7, D7, D5          ; ?? DONT UNDERSTAND
__text:0001554A                FMACD   D7, D6, D3          ;?? DONT UNDERSTAND
__text:0001554E           FLDS S12, [R1,#0x10]  ;float S12 = m->tx
__text:00015552                FCVTDS  D6, S12            ;double D6 = (double) m->tx
__text:00015556                FADDD   D4, D7, D6      ; ?? DONT UNDERSTAND
__text:0001555A           FLDS S14, [R1,#4]
__text:0001555E                FCVTDS  D6, S14
__text:00015562           FLDS S14, [R1,#0xC]
[size="2"]__text:00015566           FCVTDS  D7, S14
__text:0001556A                FMULD   D7, D5, D7
__text:0001556E                FMACD   D7, D3, D6
__text:00015572           FLDS S12, [R1,#0x14]
__text:00015576                FCVTDS  D6, S12
__text:0001557A                FADDD   D7, D7, D6
__text:0001557E                FCVTSD  S18, D4
__text:00015582                FCVTSD  S16, D7[/size][/size]

I suppose it's not that complicated I just need an explanation for a C developper.

Thanks
Parents
  • Arm Employee Badge
    Note: This was originally posted on 11th August 2011 at http://forums.arm.com


    Don't tell me it's so simple because when I read documentation they talk about vector and that kind of stuff ...


    Unless the function itself has configured the FPSCR LEN field, then it is incredibly unlikely that the VFP is operating in anything other than scalar mode; even more so given that insufficient operands appear to have  been loaded to make use of VFP short vector operations.

    hth
    s.
Reply
  • Arm Employee Badge
    Note: This was originally posted on 11th August 2011 at http://forums.arm.com


    Don't tell me it's so simple because when I read documentation they talk about vector and that kind of stuff ...


    Unless the function itself has configured the FPSCR LEN field, then it is incredibly unlikely that the VFP is operating in anything other than scalar mode; even more so given that insufficient operands appear to have  been loaded to make use of VFP short vector operations.

    hth
    s.
Children
No data