The Math Coprocessor is a second processor in your
computer that does nothing but number crunching for the
system. Addition, subtraction, multiplication, and division
of simple numbers is not the coprocessors job. It does all
the calculations involving floating point (decimal) numbers,
such as scientific calculations and algebraic functions.
These functions and calculations are used in much of the
computer's routines and just about every software available.
Spreadsheets contain statistical calculations, word
processors deal with line spacing, font size and
justification, and of course, any graphics or animation
software is relying heavily on number crunching. The
Central Processing Unit (CPU) is
perfectly capable of doing these functions and calculations.
As a matter of fact, that used to be part of its job. Most
of the older computers (pre-486) were sold without
coprocessors. So the
CPU had to process all the computer's hardware and
software functions, handle all interrupt requests (we'll
talk later), and direct all information and data, as well as
performing all floating point calculations. This required a
lot of the processor's time.
By having a second processor, or 'coprocessor', to take over
the number crunching, it can free up a lot of the CPU's
precious time. This would allow the Central Processing Unit
to focus all of its resources on the other functions it has
to perform, thus increasing the overall speed and
performance of the entire system. It's not like this was a
great revelation that came over the scientific community in
the midst of home computer development. The absence of a
math coprocessor in early home computer systems was a matter
of keeping production costs down. The advantage was
recognized right from the beginning, and most of these
motherboards had an empty slot for the aftermarket addition
of a coprocessor. The number (or name) of the math
coprocessor followed the CPU's numbering sequence, only the
last digit would be a '7', not a '6'. If you had an 8086 CPU
then you could add an 8087 coprocessor. For an 80286 you
would install an 80287, etc..
Few people nowadays remember that the IBM PC was
not the first "personal computer" and that MS-DOS was not the first industry
standard operating system. In fact, MS-DOS was but an imperfect copy of the
operating system that really has a claim to that title.
The first generation of personal computers (or microcomputers, as they were
known then) used chips like the Intel 8008, 8080, Zilog Z80, MOS Technology 6502
and Motorola 6800. While some early microcomputers (for example, the Apple II)
used proprietary operating systems, hundreds of different manufacturers licensed
a product called CP/M (as in Control Program / Monitor) made by a company
called Digital Research. Long before the IBM PC and its clones / compatibles,
the CP/M architecture provided for industry standard software that was portable
across hundreds of different brands and models. This was DRI founder Gary
Kildall's main contribution to the software industry. Microsoft simply followed
in DRI's footsteps.