One improves the IRQ emulation for a number of games, so you'll not be seeing shaky things like this or this anymore. Probably of most interest to people would be a couple small bug fixes for Japanese Konami games.
With a code freeze on the horizon I thought I'd share a small selection of things from the past month. NES roundup-because it's not all multicarts! (Ok, it's mostly multicarts.) It may not be the final word on this bootleg, maybe krzysiobal over at NesDev will give us a nice schematic someday soon, but in the meantime-with any luck and the blessing of the MAME devs- enjoy a nice SMB2 bootleg, 12-bit IRQ counter, broken status bar, and crashing at the end of 4-4 like the bootleggers.intended? It further further occurred to me that bootleggers don't typically take pride in polishing up their handiwork and may not even had somebody around who was good enough at this tough little game to play test to the end of 4-4, even with warps! It further occurred to me that bootleggers like to borrow the work of other bootleggers. It occurred to me though, all the other bootlegs of this game have 12-bit IRQ counters.
I almost submitted a patch with a nice formula: PPU cycles * (# of lines after VBLANK + status bar height) per 3 CPU cycles etc etc. In an attempt to get this game running in MAME I *also* mucked around with values around 5750 and had it running smoothly. Your humble narrator comes into the story at this point. At least that got the game running smoothly so all was good. Why? Because, not knowing how the cart actually behaves, clever folks deduced that the number of CPU cycles between VBLANK and the end of the status bar was about 5750. Now, this title has run in various emulators for years (in MAME sadly it's always been just a grey screen) and none of them AFAIK exhibit the real hardware bugs the status bar stays nicely fixed, and players talented enough to make it to World 4-4 are pleasantly greeted by World 5-1 after finding out their princess is in another castle. You'll observe something strange right away: the status bar at the top scrolls with the game! Additionally if you make it to the end of the video you'll see the game apparently freeze on a black screen. One such bootleg can be seen running on hardware here: SMB2 (LE10). It was, however, heavily bootlegged on cartridge and on multicarts. As you may recall the original Japanese Super Mario Bros 2 was only released in Japan on the Famicom Disk System. Here's a mildly amusing case of a NES bootleg.
Fortunately given the abysmal performance of this CPU, emulating at full speed via interpreter is no problem.
With this and the Acorn 32016 second processor now running the Panos operating system successfully, it seems like most of the bugs in the NS32000 CPU core are sorted out the major remaining task is to implement the 32082 MMU. Now we just need the 6100 version of UTek installation media or a hard disk image and we should be able to claim another unique Unix workstation emulation for MAME.
The Am9516 DMA controller used by the floppy controller isn't implemented yet (also needed elsewhere, like Sun 3), and it doesn't like something about the i82586, but otherwise looking pretty good.
It's starting up and passing a number of diagnostics before failing to boot and reverting to a strange monitor on the second serial port. There's a slightly younger 4132 model which is very similar, mainly exchanging the MFM hard disk controller for SCSI, so hopefully that will be fairly straightforward to emulate as soon as firmware shows up. It was supposed to be part of a family of 6100 systems, but it never gained much traction. Reached an initial milestone for the extremely rare NS32016-based Tektronix 6130 Intelligent Graphic Workstation.