I've been taking myself some interest in x86 assembly and just exploring myself the world of 64 bit programs.
I'm very interested how the boundaries between 64 bit 32 bit applications are defined.
How different are the instruction sets (the resulting opcodes) between these two?
And take if a 32 bit process is run by windows 7 (64 bit) Would it be theoretically possible to inline 64 bit code as the processor supports it?
Not as in would any assembler allow it but take it as if it's done from a hex-editor.

Thanks in advance.

The simple explanation is, 32bit code is run in a sort of compatibility mode. Some instructions (pushd for example) don't work in 64bit mode. Check the fasm manual for the details of removed instructions. The main difference is that now addresses must be 64bit instead of 32bit, and you have more registers to play with. Also, register calling convention is becoming popular again and some are preserved in the ABI.

godomega,

Intel/AMD manuals may be not an easy reading, but they're thorough. 64-bit instructions look like their 32-bit counterparts with various REX prefixes, and in legacy/compatibility modes they can't be encoded (REX prefix becomes inc/dec opcode).

uhmmm
64 bit mode is also as fucked up as the rest of architecture.

you just have to learn it.
it will be easier than learning protected mode, since its only extension, and many things work like before but are extended.

Ok thanks baldr that was the information I needed to hear .