steve...@gmail.com
2022-05-16 01:49:29 UTC
You are correct, SABR is the intermediate output from the FORTRAN II compiler. The OS/8 FORTRAN IV compiler translates into FLAP/RALF. Same basic idea, but very different intermediate languages as FLAP/RALF is oriented to the FPP that FORTRAN IV can take advantage of.
I have used SABR on occasion, but really only when I need to embed some PAL-8 level code into FORTRAN II. For example, I've been (slowly) trying to re-create a library of graphics output code for the Tektronix 4010. Otherwise I do all of my PDP-8 assembly code in regular PAL-8. Dealing with pages the -8 is a bit inconvenient at times but since that's where I actually started machine-level coding in1975, I look at other machine instruction sets as simplifications (smile). The bottom line is that I prefer dealing with pages over wrestling with some of the crazy stuff that SABR sometimes does. My biggest unresolved challenge has been trying get a delay loop in SABR that always causes the same delay regardless of relocation and I haven't been able to make that happen.
I guess my biggest thing was going the other way. I needed to embed some matrix math code (solutions of simultaneous linear equations) into a PAL-8 application and have it run as fast as possible. But rather than writing it in PAL-8 I wrote the original code in FORTRAN II then intercepted the SABR output then rewrote that in regular PAL-8 for inclusion in the application. I recall it saved me a ton of time to get it to work in FORTRAN II first rather than trying to wrestle with both the algorithm and PAL-level code at the very same time.
Cheers,
-- steve
I have used SABR on occasion, but really only when I need to embed some PAL-8 level code into FORTRAN II. For example, I've been (slowly) trying to re-create a library of graphics output code for the Tektronix 4010. Otherwise I do all of my PDP-8 assembly code in regular PAL-8. Dealing with pages the -8 is a bit inconvenient at times but since that's where I actually started machine-level coding in1975, I look at other machine instruction sets as simplifications (smile). The bottom line is that I prefer dealing with pages over wrestling with some of the crazy stuff that SABR sometimes does. My biggest unresolved challenge has been trying get a delay loop in SABR that always causes the same delay regardless of relocation and I haven't been able to make that happen.
I guess my biggest thing was going the other way. I needed to embed some matrix math code (solutions of simultaneous linear equations) into a PAL-8 application and have it run as fast as possible. But rather than writing it in PAL-8 I wrote the original code in FORTRAN II then intercepted the SABR output then rewrote that in regular PAL-8 for inclusion in the application. I recall it saved me a ton of time to get it to work in FORTRAN II first rather than trying to wrestle with both the algorithm and PAL-level code at the very same time.
Cheers,
-- steve