Compiler slow down

[jonathan]

I am developing a signal processing app using Pelles C in a Windows 64bit console.  My code is processing streaming data and the main loop timing is critical.   In the loop I obtain a sin and cosine floating point values to use.  I have two approaches operating within the loop, one using a previously calculated lookup table and the other calculates it on the fly. Again both calculations are operating.  Both calculations produce identical values.  Yet when I use the values calculated on the fly the timing of the loop goes up by 80%!.  Here is the code segment:


//calculated on the fly
         ptrst1=ptrst+j6+25000;
         t8=*ptrst1;
         ptrst1=(ptrst+j6);
         t9=*ptrst1;
//lookup table
         t6=*ptrs++;
         t7=*ptrs++;
      //   t6=t8;
      //   t7=t9;

t8 is equal to t6 and t9 is equal to t7.  Yet when I take off the comment on last two lines the 80% slow down occurs.  I want to look at the assembly listing to see what is happening to the compiled code but I am unable to figure out how to get it from the compiler.  Whats going on?

[TimoVJL]

Compiler option /Tamd64-asm
Or use podump.exe with option /DISASM with object-file.
objconv

Disasm obj Add-Ins:
https://forum.pellesc.de/index.php?topic=6985.msg26516#msg26516
https://forum.pellesc.de/index.php?topic=7248.msg27559#msg27559