On 26 Apr 2010 13:11:21 -0700, [email protected] (Rick Fochtman) wrote: >I submit that MIPS is not a valid measurement, since it has no real >correlation with the amount of useful work accomplished by any machine. >One example is RISC machines. They have to simulate instructions that >non-RISC machines execute on a regular basis. The only really good >measure of processor speed is this: does it finish a business's workload >in a timely fashion? To the Board of Directors, that's really the only >valid measure. While a PC on a desktop might be more than adequate for >secretarial chores, running a large corporation is a horse of a >completely different color. Consider all the different tasks involved, >including such things as accounts receivable, accounts payable, >inventory management, payrolls, etc. Some of these tasks require HUGE >amounts of I/O, far beyind the capabilities of a desktop PC.
Even there, "speed" isn't what is being measured. Throughput matters. And it doesn't matter how fast the CPU is on your document in the cloud if your internet connection is slow. We used to do more spreadsheet type processing on mainframes - before spreadsheet programs were available on users desktops. Now it is much quicker for them to do their data massaging on their desktop. It's even often quicker when their main computer is doing the manipulating live. The advantage (to them) of having the data on the mainframe isn't speed, it is making sure the data are valid and current. Even if it takes longer to get results to their desktops. ---------------------------------------------------------------------- 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
