jse@hprnd.HP.COM (Jeff Ebeling) (03/13/90)
I don't know if this is the proper forum for questions on MCA. If it isn't please refer me to the proper notes group. If it is, please respond to this basenote with a posting or mail. Is anyone familiar with MCA fairness arbitration? I have copies of the May 1987 IBM Personal System/2 Seminar Proceedings and the Technical Reference Manual. Their explanations are weak, their formulas are wrong and I can't figure out how they generated their numbers for their example table on arbitration for 5 bursting arbiters. All of the following refer to the Arbitration Latency and Minimum Bus Share Tutorial in the Seminar Proceedings. "The nth and (n-1)th arbiters will each get 1/(2-1) of the bus min." I assume this was supposed to say 1/(2(n-1)) of the bus grants. Even so, where was this previously defined. Why doesn't each arbiter get 1/n of the bus grants? Secondly, for the table, "Assuming time per transfer is equal for all arbiters." At what time value? Thirdly, where in the table, do the Latency Total numbers come from? I couldn't make any sense out of them. Did IBM intentionally use poor documentation as a means of making their open bus closed? Thanks in advance, Jeff Jeff Ebeling Hewlett Packard, Roseville Networks Division jse@hprnd.rose.hp.com TELNET 785-5613 (916) 785-5613