Difference between revisions of "FIS floating point"
m (→External links: +Use of FIS in 'Computer Engineering') |
m (→Architecture: Typo; +links) |
||
| (One intermediate revision by the same user not shown) | |||
| Line 1: | Line 1: | ||
'''FIS (Floating Instruction Set) floating point''' is the term used for the optional [[floating point]] [[instruction]]s supported by a few of the [[PDP-11]]s which had [[hardware]] support for floating point. | '''FIS (Floating Instruction Set) floating point''' is the term used for the optional [[floating point]] [[instruction]]s supported by a few of the [[PDP-11]]s which had [[hardware]] support for floating point. | ||
| − | Most supported a more | + | Most supported a more powerful floating point capability, the [[FP11 floating point]]. Even though the instructions were not compatible between the two, the FIS was useful since [[Digital Equipment Corporation|DEC]]'s [[FORTRAN]] [[compiler]] (the main 'target' for the FIS) for the PDP-11 used [[threaded code]], and was thus able, via [[conditional assembly]] of the [[run-time support]], to easily make use of the low-capability, but low-cost, FIS version. |
As with the FP11 version, it was tightly integrated with the [[Central Processing Unit|CPU]], so that the CPU processed a mix of 'regular' and floating point instructions. | As with the FP11 version, it was tightly integrated with the [[Central Processing Unit|CPU]], so that the CPU processed a mix of 'regular' and floating point instructions. | ||
| Line 9: | Line 9: | ||
FIS-type floating point supported only short (32 bit) floating point numbers; it used an 8 bit exponent (in 'excess 0200' notation, giving an exponent range of +127. to -128.), a sign bit, and the remaining 23 bits were the fractional part. | FIS-type floating point supported only short (32 bit) floating point numbers; it used an 8 bit exponent (in 'excess 0200' notation, giving an exponent range of +127. to -128.), a sign bit, and the remaining 23 bits were the fractional part. | ||
| − | It provided only the four basic arithmetic operations. Checking for underflow. overflow, and division by zero was performed in hardware, and would cause | + | It provided only the four basic arithmetic operations. Checking for underflow. overflow, and division by zero was performed in hardware, and would cause a [[trap]] if they happened; [[condition codes|condition code bits]] in the main CPU recorded the outcome of operations. |
| − | The instructions used one of the [[PDP-11 architecture|regular PDP-11 registers]] as a pointer to load and store floating point data from/to memory | + | The instructions used one of the [[PDP-11 architecture|regular PDP-11 registers]] as a [[pointer]] to load and store floating point data from/to [[main memory]]. Floating point values could only be stored in memory, generally on the [[stack]]; there was no way to store floating point values in the CPU. |
==Implementations== | ==Implementations== | ||
Latest revision as of 11:11, 17 December 2024
FIS (Floating Instruction Set) floating point is the term used for the optional floating point instructions supported by a few of the PDP-11s which had hardware support for floating point.
Most supported a more powerful floating point capability, the FP11 floating point. Even though the instructions were not compatible between the two, the FIS was useful since DEC's FORTRAN compiler (the main 'target' for the FIS) for the PDP-11 used threaded code, and was thus able, via conditional assembly of the run-time support, to easily make use of the low-capability, but low-cost, FIS version.
As with the FP11 version, it was tightly integrated with the CPU, so that the CPU processed a mix of 'regular' and floating point instructions.
Architecture
FIS-type floating point supported only short (32 bit) floating point numbers; it used an 8 bit exponent (in 'excess 0200' notation, giving an exponent range of +127. to -128.), a sign bit, and the remaining 23 bits were the fractional part.
It provided only the four basic arithmetic operations. Checking for underflow. overflow, and division by zero was performed in hardware, and would cause a trap if they happened; condition code bits in the main CPU recorded the outcome of operations.
The instructions used one of the regular PDP-11 registers as a pointer to load and store floating point data from/to main memory. Floating point values could only be stored in memory, generally on the stack; there was no way to store floating point values in the CPU.
Implementations
The initial implementation was the optional KE11-F Floating Instruction Set for the KD11-A CPU for the PDP-11/40; it was later used in the optional KEV11-A floating point option for the LSI-11 CPU of the PDP-11/03.
External links
- Ronald F. Brender, Turning Cousins into Sisters: An Example of Software Smoothing of Hardware Differences, in C. Gordon Bell, J. Craig Mudge, John. E. McNamara, Computer Engineering: A DEC View of Hardware Systems Design, Digital Press, Bedford, 1978 - Explains how DEC software made use of the FIS instructions
