Numeric Co-processors

If you make extensive use of applications that are concerned mainly with calculation then a numeric co-processor is essential. In this chapter we examine which co-processor is best and which application really benefit from one.


We have encountered the idea of a co-processors of a number of types in earlier chapters of this book. But historically the numeric co-processor was the earliest important and common example of a co-processor. A numeric co-processor takes over from the main processor whenever an arithmetic operation is needed. To be strictly accurate, numeric co-processors generally only take over when a floating point arithmetic operation is need. A floating point operation can be thought of as one that involves fractional numbers or results-but this isn’t entirely true because there is a way of handling fractional numbers that isn’t based on floating point numbers, called fixed point arithmetic. What all of this means to the user is that a numeric co-processor really only helps when you are doing arithmetic that involves something other than whole numbers, and usually only when multiplication, division or some complicated trigonometric or transcendental function is part of the calculation.

Watch the video down below.


Notice that this implies not only the use of an EISA/MCA bus machine but an EISA/MCA disk controller and network adapter cards. Of course, moving to an EISA/MCA based machine will not solve a performance problem caused by workstation overloading the capacity of the network itself, or anything else not related to the movement of data within the server. 




EmoticonEmoticon