to get VC++ generate asm?

[fearless]

no idea what this jmps/calls do,but they have nothing todo with SSE

the __ftol2_sse is a ms visual c library function to convert float to long - the code in __ftol2_sse checks to see if it can use sse to calc that, otherwise it defaults to whatever normal non sse code is in that function.

[TimoVJL]

To avoid SSE:
/QIfist[-] use FIST instead of ftol()
/arch:IA32 - use no enhanced instructions and use x87 for floating point

[daydreamer]

thanks Steve,Timo,fearless

[daydreamer]

this IDE and C++ compiler seem like a way to make a bigger project,mixing masm code and C++ code
enjoy use masm code in a project,while C++ code to make you more productive where speed doesnt matter, to maybe finally put together a big project
but also fun in compare a long switch with a jumptable

[daydreamer]

it would be nice to learn howto also make macros inside VC++,because I read "_inline" is just bad suggestion to compiler that its probably not will follow

[TimoVJL]

Why ?
msvc x64 don't support inline  assembler.
read about msvc __forceinline, if you just stick with 32-bit.

I don't know how those #defines affect to optimizer.

[daydreamer]

Why ?
msvc x64 don't support inline  assembler.
read about msvc __forceinline, if you just stick with 32-bit.

I don't know how those #defines affect to optimizer.

thanks,because if SIMD needs other things than an assembler to keep on using,I have checked alternatives intrinsic looks like it has similarities to assembler but more lot complex and longer than the usual mnemonics to type