Intel 8087: Difference between revisions
imported>George Swan (first draft) |
imported>George Swan m (fix refs) |
||
| Line 3: | Line 3: | ||
Both the 8088 and 8086 implemented the [[intel 8080 instruction set|8080 instruction set]], which included a subset of instructions for [[floating point operations]] -- arithmetic operations that might include fractions. | Both the 8088 and 8086 implemented the [[intel 8080 instruction set|8080 instruction set]], which included a subset of instructions for [[floating point operations]] -- arithmetic operations that might include fractions. | ||
[[Integer arithmetic]] requires much less [[silicon real estate]] to implement than calculations on values that had a [[fraction]]al component. Intel off-loaded the hardware to perform those instructions to a [[co-processor]] -- a [[support chip]] devoted solely to that task. Computers could be equipped with an 8088 or 8086, and without an 8087. | [[Integer arithmetic]] requires much less [[silicon real estate]] to implement than calculations on values that had a [[Fraction (mathematics)|fraction]]al component. Intel off-loaded the hardware to perform those instructions to a [[co-processor]] -- a [[support chip]] devoted solely to that task. Computers could be equipped with an 8088 or 8086, and without an 8087. | ||
<!-- | <!-- | ||
When the 8088 or 8086 [[central processing unit]] encountered a floating point instruction it would be executed by the CPU, slowly, in firmware. | When the 8088 or 8086 [[central processing unit]] encountered a floating point instruction it would be executed by the CPU, slowly, in firmware. | ||
Revision as of 23:00, 27 November 2008
The intel 8087 was a floating point co-processor to intel's 8088 and 8086 models of single chip computer. Both the 8088 and 8086 implemented the 8080 instruction set, which included a subset of instructions for floating point operations -- arithmetic operations that might include fractions.
Integer arithmetic requires much less silicon real estate to implement than calculations on values that had a fractional component. Intel off-loaded the hardware to perform those instructions to a co-processor -- a support chip devoted solely to that task. Computers could be equipped with an 8088 or 8086, and without an 8087. If the 8087 were present floating point instructions would be off-loaded and executed relatively quickly, by the 8087.
Later chips, lineal descendants to the 8086, the 80186, 80286 and 80386 had a companion floating point co-processor manufactured to accompany them. Intel was to market an intel 80487. But it was not a separate chip, different from the 80486.