Talk:NES 2.0 Mapper 319

typo?

FCEUX's source in boards/hp898f.cpp swaps the location of the two registers relative to what's documented here. —Lidnariq (talk) 17:32, 15 May 2018 (MDT)

Additionally, the description here (modifying overriding PRG A16 with CPU A14, instead of PRG A14 with CPU A14) really sounds like this description is a misdump. —Lidnariq (talk) 19:13, 15 May 2018 (MDT)

I have corrected the relative register location. Not sure what you mean by "misdumped description " in the second part. If the 16 KiB PRG-ROM bank bit 2 (4s) is modified, it's going to affect PRG A16, not A14. NewRisingSun (talk) 23:52, 15 May 2018 (MDT)

Or do you mean the ROM (and not the description) is misdumped, having the banks in a supposedly implausible order? I have not dumped the cart myself, and I am not going to assert bad dump status without having seen the board topology. NewRisingSun (talk) 00:10, 16 May 2018 (MDT)

That is exactly what I mean. There are a wealth of other boards which support both NROM-256 and NROM-128 by relaying or overriding CPU A14; this cart is unique in interpreting that signal as connecting to PRG A16 instead of A14. —Lidnariq (talk) 14:09, 16 May 2018 (MDT)

Here are the values that the menu is writing and the resulting 16 KiB bank numbers at 8000 and C000 according to the A16 description that runs all games on the multicart:

               Written 16K Banks# @
               Value   8000    C000
Menu		38	07	07
Super Mario     D0	02	06 *
Star Force	C8	01	05 *
Tennis		98	03	03
F-1 Race	80	00	00
Lode Runner	A0	04	04
Duck Hunt	88	01	01
Galaxian	38	07	07

(The cart has Inventor's renamed hacks of the games; I used the original names.) I'm struggling to think how to rearrange the banks in such a way that Super Mario could work with an A14 override (yielding in banks 02,03 instead of 02,06) without breaking Tennis. NewRisingSun (talk) 14:29, 16 May 2018 (MDT)