|
Re: My second attempt at a data_xfer instruction optimization: msg#00115emulators.bochs.devel
Brian Slechta wrote: > I am assuming that the reason the compiler could not safely remove these > temporary variables is because the "data" parameter in write_virtual() > is passed by pointer. Since the compiler could not prove that the value > pointed to by "data" would not be manipulated in the write_virtual() > functions, it could not make this substitution that you have made by hand. You could try the "__restrict" keyword, which tells the compiler that the memory pointed to by the "data" pointer does not overlap any other memory that is accessed in that function. -- Jamie ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | cpu.cc merge question: 00115, Brian Slechta |
|---|---|
| Next by Date: | RE: cpu.cc merge question: 00115, Stanislav Shwartsman |
| Previous by Thread: | Re: My second attempt at a data_xfer instruction optimizationi: 00115, Brian Slechta |
| Next by Thread: | Re: My second attempt at a data_xfer instruction optimization: 00115, Conn Clark |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |