Gast
#339915
Is it possible to use progn.c with MPLAB C18 ? I want to use a pic 18F2620. Exackt.txt are correct results. Progn.txt are results from winavr-compiler with atmega32 and they were wrong.
Gast
#339916
The problem might be the reduced accuracy of your long double variables, in WINAVR that means 32bit floating point, which is not much. How did you get the exact results? PC? http://www.mikrocontroller.net/forum/read-2-12872.html : Selbst der Mercedes unter den Compilern (IAR Embedded Workbench) benutzt standardmäßig 32-bit-Gleitkomma. Allerdings kann man den auf 64 bit umschalten. Für den AVR-GCC müsste sich einfach jemand finden, der das schreibt. also read http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&p=230619
Gast
#339917
I think there is a huge problem with 32-bit float point calculating module from winavr ! Progn.c is now test with a another 32-bit FP math module without problems. Read also http://www.mikrocontroller.net/forum/read-2-350211.html#new