SOB? about python-pdp11-emulator

[outofmbufs]

Well, I'm not entirely sure why I wrote that code that way, it's certainly less-clear than it could be, but I think the two implementations are identical. Here's a test showing that -- unless you think I've overlooked something.

For the test code, I replaced "cpu.r[srcreg]" with just a function argument (see the code below) and have the functions return a tuple: True/False (branch taken or not) and the modified register value. They produce identical results for all inputs 0 .. 65535:

Put this in a file foo.py and "python3 foo.py" to run it.

def sobSIMH(srcreg):

r = (srcreg - 1) & 0xffff
if r != 0:
    return r, True              # branch
return r, False                 # no branch

def sobNW(srcreg):
r = srcreg

if r == 1:
    r = 0
else:
    if r > 0:
        r -= 1
    else:
        r = 0o177777    # 0 means max, that's how SOB is defined

    return r, True

return r, False

if name == "main":
import unittest

class TestMethods(unittest.TestCase):
    def test_compare(self):
        for srcreg in range(65536):
            self.assertEqual(sobNW(srcreg), sobSIMH(srcreg))

unittest.main()

[outofmbufs]

ugh - code got hosed in cut/paste. Get it here for now:
https://www.dropbox.com/scl/fi/mfxyo2gqsv1e4fe5m2an8/sobx.py?rlkey=o41hjnaf389kdp1la5nng6rnh&dl=0

[folkertvanheusden]

Hmmm you're absolutely right indeed. Sorry for the confusion, could've verified that myself! Have you tried to compile the python code by the way? I tried shedskin and nuitka3 but both failed unfortunately.

…

On Sun, Mar 31, 2024 at 3:12 PM Neil Webber ***@***.***> wrote: Well, I'm not entirely sure why I wrote that code that way, it's certainly less-clear than it could be, but I think the two implementations are identical. Here's a test showing that -- unless you think I've overlooked something. For the test code, I replaced "cpu.r[srcreg]" with just a function argument (see the code below) and have the functions return a tuple: True/False (branch taken or not) and the modified register value. They produce identical results for all inputs 0 .. 65535: Put this in a file foo.py and "python3 foo.py" to run it. def sobSIMH(srcreg): r = (srcreg - 1) & 0xffff if r != 0: return r, True # branch return r, False # no branch def sobNW(srcreg): r = srcreg if r == 1: r = 0 else: if r > 0: r -= 1 else: r = 0o177777 # 0 means max, that's how SOB is defined return r, True return r, False if *name* == "*main*": import unittest class TestMethods(unittest.TestCase): def test_compare(self): for srcreg in range(65536): self.assertEqual(sobNW(srcreg), sobSIMH(srcreg)) unittest.main() — Reply to this email directly, view it on GitHub <#9 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AUN5IW4MP52CG4WLGSQUSILY3ADR3AVCNFSM6AAAAABFQJTRS2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRYG4YTSNJVG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[outofmbufs]

Took the pull even though old code worked. New code is simpler.

Never tried compiling. I did spend a little time trying to make the simplest instructions (register-to-register moves in particular) go fast, but at some point if performance is that important "use SIMH" becomes a better answer. I was just happy that it booted v7 (despite all the recently-found bugs! lol) and ran at a reasonable speed.