> > hi all, > i am developing a data acquiring program on rtlinux 3.1,and need to tranmit a >large amount of data between the rt process and non_rt process at high speed,so try >to find the most efficient way of communication. > After trying fifo,mbuff and share memory pool (putting aside a block of physical >memory at boot time then access the pointer directly), the performance shows that the >fifo cann't meet the requirement, but the differences between the mbuff and share > memory pool are still unclear. > It will be appreciated if who can offer some experience about this. > Thanks > could you post your results of the comparison and the conditions ? I guess that would be of interest for others too.
the memory pool you refere to - are you talking about reserving physical memory with a mem= parameter to the kernel and then directly accessing it ? Would be supprised if there is much of a difference between using reserved physical memory or mbuff - but I am more supprised that FIFO's were outperformed by shared memory. hofrat -- [rtl] --- To unsubscribe: echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED] -- For more information on Real-Time Linux see: http://www.rtlinux.org/