On 13 Jun 2007 11:03:32 -0800, in bit.listserv.ibm-main you wrote: >That's because you bought into the linguistic aspects of geometry and math. >That's a particular axiomatic consequence of math in the 20th century.
I suppose parallel is a misleading word, as parallel lines never converge. We have various ways to run "simultaneous" tasks. Running two threads in one von Neumann computer is on the face of it only pseudo multi-tasking. But it works when one task has to wait. Threading two highly CPU intensive functions only works when we have multiple CPUs. But it is usually the OS that determines whether a CPU gets an active thread. A programmer offers it that opportunity, but the OS has other tasks to run. It would be interesting if a virtual machine would be handed a CPU, whether the mainframe, or maybe running Parallels on a Macintosh. But I don't see this happening. With offloaded functions, a program can say "read the disk", and either the OS has a varying amount of work it does or passes on. I read an article on game computing where the program checked to see what gets done by the video card, what gets done by the Windows version, and what has to be coded by the program - and then picks which of those choices get done while the game is working on strategy in a different thread. Sounds very leading edge to me. In my programming, I trust the DBAs handle the parallel optimization - my optimization is completely different in nature. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
