I'm curious, the use of new compilers uses some additional features such as MMX SSE, 3D NY! Hence?
I mean, even the original 8086 did not have any FPU, so the compiler which old can not even use it, but the new compiler, because the FPU is part of every new CPU. So, do new compilers use the new features of the CPU?
Or, should have more authority to ask, does the new C / C ++ standout library work use new features? thanks for answer.
Edit:
Ok, so, if I make you all right, even some standard operation, especially float numbers can be SSE fast Made by using
To use it, I have to enable this feature in my compiler, if it supports it. If this happens, then I should be sure that the target platform supports the features.
In case of some system libraries that require top performance, such as OpenGL, DirectX and so on, this support may be supported in the system. / P>
For compatibility reasons, the compiler does not support it, but you can add this support by using special functions, for example Intel should be the best method, because you can control the wheather directly And when you use special features of the desired platform to write multi-CPU-support applications
GCC will support new instructions through command line arguments. For more information, see Bidding:
GMC can take advantage of additional instructions of MMX, SSE, SSE2, SSE3 and recent Intel and AMD processors 3 DON Extensions. Options- The use of these additional instructions from MMMX, MMS, MMS2, MMS3, and M3nOn enable the use of these additional instructions, so that many words of data can be processed in parallel. As a result, executables will only run on the processors supporting the appropriate extension - they will crash on other systems with invalid instruction errors (or similar)
Comments
Post a Comment