Talk:IBM 5100

From Wikipedia, the free encyclopedia

You have new messages (last change).

User:MadIce: The PALM microcode was not "ported" from another system. The PALM only has a very small amount of Control ROS, and it was custom written specifically for PALM. No IBM processor at that time used 32-bit microcode; most used horizontal microcode words of more than 60 bits.

Perhaps the porting effort you are alluding to was the porting of APL\360 or Series/3 BASIC. These programs are in the "Language ROS" which is not part of the PALM processor. In fact, to the PALM the Language ROS acts as a peripheral device rather than as conventional memory.

The PALM microcode (Control ROS) is just a tiny ROM that implements the PALM instruction set, as described in an appendix of the IBM 5100 Maintenance Information Manual. The PALM processor executes PALM instructions from the "Executable ROS", or from RAM.

The Executable ROS contains the system initialization code, diagnostics, I/O routines, and processor simulators. There is a System/3 simulator for running BASIC, and a System/360 simulator for running APL\360.

The actual BASIC interpreter (in System/3 code) and the APL\360 interpreter (written in System/360 code) are stored in the Language ROS. These language interpreters are the only IBM 5100 ROM code that can be said to have been "ported" from earlier IBM systems. --Brouhaha 07:44, 8 Sep 2004 (UTC)

Brouhaha: Thanks for pointing that out.

Brouhaha: You talked about the Executable ROS containing "processor simulators". Was the 5100 able to emulate other processors and if so which ones did it emulate?

Brouhaha: I don't want to be a PITA, but can you please read the design decisions for the IBM 5100 starting at page 425 (Small machines) of the IBM Systems Journal, Volume 30, Number 4, 1991?

I am quoting page 426: "This time, however, although the same Palm internal engine was used, System/360 architecture was emulated rather than 1130 architecture, so that the up-to-date APSLV product system could be used as the APL facility with virtually no modification."

This leads me to believe that the processor was emulated by the PALM, instead of APL/360 being ported to the machine.

By IBM's own definitions, portions of the code in the Executable ROS (written in PALM assembly language) simulate the System\360 and System/3 processors, to run APLSV and BASIC, respectively. (My earlier reference to APL\360 was incorrect.) The APLSV code was ported in the sense that although the IBM 5100 is simulating a System\360 processor, it is not running a normal System\360 operating system. Some changes to the APLSV code were made to get it to run on the IBM 5100; I don't have any details on the extent of these changes, but I expect one could consider it to be a relatively minor port. Probably more work than porting a program from OS/MVT to CMS, for instance, but much less than porting a program from a System/360 to a VAX.

I was mostly trying to correct the notion that the PALM microcode (Control ROS) was ported from somewhere else. The Control ROS is specific to PALM and only serves to implement the PALM instruction set.

With the introduction of the System/360, IBM defined processor emulation as the simulation of another processor using special-purpose hardware and/or microcode. In the article you cite, IBM has apparently become somewhat sloppy in their use of the term "emulation". It seems commonplace now for the term "emulation" to be used where "simulation" is more appropriate, but this is unfortunate as it conveys less useful information. In the case of the IBM 5100, the System\360 and System\3 simulation are not particularly assisted in any way by special-purpose hardware or microcode, thus there is no good reason to refer to it as "emulation". --Brouhaha 04:46, 10 Sep 2004 (UTC)

Burroughs had a smaller machine, the B1700, that allowed machine instructions to be emulated in micrcode, which came out at about the same time as the 5100. And I'm sure someone else had done it before. So, it wasn't all that "innovative." --205.175.225.5 23:56, 11 August 2006

The B1700 series machines were *much* larger than the IBM 5100. For instance, the B1712 and B1714 processor cabinet alone weighed 900 lb., vs. about 50 lb. for the IBM 5100. And the 900 lb. of the Burroughs machine didn't include a terminal (keyboard and display) or tape drive; those added hundreds of lb.
The IBM invention of emulation was with the System/360 in 1964, not the IBM 5100 in 1975. To the extent that the use of emulation was innovative in the IBM 5100, it was because they did it in a desktop computer weighing approximately 50 lb.
The B1700 series was definitely interesting and very sophisticated, but comparing it to an IBM 5100 is silly. They are very different machines with very different capabilities. --Brouhaha 02:52, 12 August 2006 (UTC)
Retrieved from "http://en.wikipedia.org../../../i/b/m/Talk%7EIBM_5100_4a5e.html"