The problem with calculations on types float with Cosmic compiler

[krzysztof239955_st]

Posted on October 02, 2009 at 09:13

The problem with calculations on types float with Cosmic compiler

[krzysztof239955_st]

Posted on May 17, 2011 at 15:04

Hi,

Why variable wsp_kal not have after the calculations of the correct value 0.000038? Example: float masa_kal,dzal_kal1,dzal_kal2,dzal_kal,wsp_kal; masa_kal = (float)(mas_kal[i+1]-mas_kal[i]); // correct 5 dzal_kal1 = (float)kal_ram.dz_kal[i+1]; // correct 329602 dzal_kal2 = (float)kal_ram.dz_kal[i]; // correct 198530 dzal_kal = (float)(dzal_kal1-dzal_kal2); // correct 131072 wsp_kal = (float)(masa_kal/dzal_kal); // incorrect 0

[mozra27]

Posted on May 17, 2011 at 15:04

Hi,

What is the compiler used in your application and what is the optimization level? Please check with the following declaration: float masa_kal,dzal_kal1,dzal_kal2,dzal_kal,wsp_kal = 0.0; Regards mozra

[krzysztof239955_st]

Posted on May 17, 2011 at 15:04

Hi,

I use STVD 4.1.2 and the compiler Cosmic CxSTM8 4.3.1 with optimizations: Minimize code size (+compact). Declaration: float masa_kal,dzal_kal1,dzal_kal2,dzal_kal,wsp_kal=0.0; - not bring the improvement. Best regards

[krzysztof239955_st]

Posted on May 17, 2011 at 15:04

Hi,

The setting Display Item from Decimal on Standard in debuger solved the problem. Best regards krzysztof.skoczek