4K HP 67 Games rom attached for testing

[Dieter]

(11-20-2016 02:31 AM)Gene Wrote:  Until we figure out if there is a problem in the AOS, this is the version that will go in. It gives answers instead of zeroes. :-)

AOS into .rom

Yes, there is a problem that I fixed in the last version I got here – but obviously the wrong one went into the zip file I posted.

The problem is in the Equals-routine (E) which finally calls the clear-all routine (XEQ a) at a point where the final result is in X. Due to the CLST in the clear-routine this value is lost. So your program works correctly, it just clears its result right before returning it. #-)

The attachment includes two versions: AOS_3 uses CLRG and thus clears all the 41's data registers. Here the final result is saved in Z and finally recalled by two RDNs (same as the original progam). AOS_3a uses a CLRGX loop that clears only R00...R22 and saves the result in LastX so that it can be finally recalled from there after the clearing is done. Both versions should work, both have a specific advantage: AOS_3 is faster, but clears all data registers, while AOS_3a clears only the required ones but takes 2-3 seconds longer for each "AOS" and "=" call due to the clear loop. If a complete CLRG is fine for you, use AOS_3.

Edit: I wonder if this CLRG is required at all. After the stack pointers are reset everything should work without setting all registers to zero. Gene, could you try this by loading AOS_3 and removing the initial CLRG? I have filled R00...R22 with random numbers and the program seems to work fine without clearing the registers.

Dieter