80487

The Intel 80487, also known as the 80487SX or 80487DX, was a math coprocessor designed by Intel to accelerate floating-point calculations for Intel 80486 processors. It was specifically intended to function as a companion to the Intel 80486SX, a version of the 80486 that lacked an integrated floating-point unit. The 80487DX, on the other hand, was compatible with both the 80486DX and 80486SX.

When installed in a dedicated socket on the motherboard, the 80487 would detect the presence of an 80486 processor. If it detected an 80486SX, it would essentially take over all numeric operations, allowing the 80486SX to function solely as an integer processor. If it detected an 80486DX (which already had a floating-point unit), the 80487 would disable itself to avoid conflicts. This design ensured that software utilizing floating-point instructions would automatically be directed to the coprocessor for faster execution.

The 80487 was available in various clock speeds, matching the frequencies of the compatible 80486 processors. Its introduction provided a significant performance boost for applications that relied heavily on mathematical computations, such as CAD software, scientific simulations, and early 3D graphics. However, with the release of the Intel 80486DX2 and later 80486 processors that integrated the floating-point unit directly onto the main CPU die, the need for a separate math coprocessor like the 80487 diminished, and its production was eventually phased out.