>Unofficial Jordana Release<

[ShovelhEd]

I found this info here:<br>http://www.iseran.com/Win32/FAQ/except.html

[Dan East]

Windows NT defines that error as a data misalignment, while Windows CE defines it as out of memory. So it's hard to tell which to believe. The data alignment error makes far more sense to me in this case.<br><br>Dan East

[ShovelhEd]

Dan,<br><br>Grabbed this from the PocketProject forum:<br><br>I'm getting a First Chance exception : 0x80000002 Data Misalignment Exception...<br><br>You'll get that kind of error message from a MIPS processor (and maybe others?) whenever you try to pull a DWORD of data off a memory address that isn't DWORD aligned...<br>...Are you sure you're doing the right thing with all the data members<br><br>If all else fails, you can memcpy the data into a DWORD, and read the dword from there.<br><br><br>

[ShovelhEd]

Dan,<br><br>There is a gretn deal of info in this thread.  Talking about this error and ways that it can be repaired.<br><br>http://www.pocketprojects.com/ubb/Forum2/HTML/000935.html

[Dan East]

If it is a misalignment problem then it will be very easy to fix.  The biggest issue is finding exactly where the error occurs since I do not have access to an SH3 Pocket PC. Someone volunteered to debug on their device if I walk them through the steps, so I'll approach them over this soon.<br><br>Dan East

[ShovelhEd]

Dan,<br><br>Great stuff.  I would also be willing to debug on my device if the other person is busy...<br><br>I am pretty much a newbie when it comes to programming.  The forum talked of using #pragma pack set to 4 I think?  What does that accomplish?  and why do you think the SH3 has this error?  Do the MIPS processors fix the alignment errors?  Ignore them?  I read a bit on a PowerPC site about performance diffrences when data is properly aligned and they are significant.  I have included a snippet below<br><br>Thanks for working on this...Pretty cool stuff and sorry for being a pest with all of these messages.<br><br>CPU and data Aligned total access time (usec) Misaligned total access time (usec) Overhead <br>PowerPC 601 integers 113439 119573 5.4% <br>PowerPC 601 floats 63234 94505 50.0% <br><br>PowerPC 601 doubles 63251 113306 79.1% <br><br>PowerPC 604 integers 687 695 1.1% <br><br>PowerPC 604 floats 261 23753 9009.0% <br><br>PowerPC 604 doubles 262 22546 8509.5% <br><br>It seems doubles that are misaligned cause as much as 9009.0% diffrence is speed...  I don't knowenough about all this but just thought the numbers looked interesting...could be apples (no pun intended) and oranges for all I know.<br><br>