On 13 January 2013 13:37, Yogesh Tyagi wrote: > Hi, > > Can anybody help me to understand why can't gcc optimize the code > using SSE and MMX instructions for any given architecture.Why I need > to write assembly code to use SSE and MMX instructions? With the complete lack of details you've supplied noone can really help. What have you tried? What compiler flags did you use? What results did you get? What architecture are you using? Did you tell the compiler to generate code for that architecture? (As opposed to e.g. for 486)
