80487SX

The 80487SX, also known as the i487SX, was a math coprocessor designed by Intel for use with their 486SX processors. The 486SX was a cost-reduced version of the 486DX, lacking the integrated floating-point unit (FPU). The 80487SX was essentially a full 486DX in a different package, intended to be installed in a dedicated socket on motherboards that supported it. This allowed users of 486SX systems to add floating-point capabilities without having to upgrade their entire CPU.

When the 80487SX was installed, it would detect the presence of the 486SX and effectively disable it, taking over all CPU and FPU operations. This dual-functionality meant the 80487SX acted as both a primary processor and a math coprocessor. It operated at the same clock speed as the paired 486SX. The 80487SX was available in various clock speeds, matching the available 486SX processors of the time, such as 16 MHz, 20 MHz, 25 MHz, and 33 MHz.

The primary purpose of the 80487SX was to accelerate mathematical calculations, particularly those found in CAD (Computer-Aided Design), scientific simulations, and graphics-intensive applications. While the 486SX was sufficient for general computing tasks, the lack of an FPU made it perform poorly in computationally demanding floating-point operations. The 80487SX addressed this limitation, providing a significant performance boost for software that utilized floating-point arithmetic. However, the advent of the full 486DX and later integrated FPUs in subsequent processor generations eventually made the separate coprocessor market obsolete.