MSX2+ vs Zemmix, any real difference?

The 1chipMSX / Zemmix Neo is an emulation in FPGA and the emulation is not perfect, so as a programmer this is sometimes I run into sometimes. E.g. in with VGMPlay I encounted an issue with line interrupts on line 208, and with my RPG project I encountered an overscan issue. KdL has been very helpful fixing those issues, but still… It’s good to keep it in mind.

In other words, don’t use the 1chipMSX or Zemmix Neo to do development on or verify real machine behaviour, it is better to use an accurate emulator like openMSX and a real machine. And when you encounter an issue on the Zemmix Neo, report it to KdL.

I just would like to suggest that you consider a MSX-SM instead of a Zemmix Neo if going through the FPGA route, it has all the perks of Zemmix Neo, some of the disadvantages were fixed, FPGA is almost twice the size of OCM/Zemmix Neo so it has more stuff being done to it (WiFi interface, extra sound chips, built in midi) and has way better 12V/-12V supply that fixes some cartridges that do not work (and its base firmware is KDL OCM, so you start with a pretty good base that is the same as Zemmix and jus has new stuff being added to it :-)), better sound and HDMI output. It sounds like MSX-SM should be more expensive than Zemmix/OCM, but it is not.

P.s.:it is not an advertising on any way, I don't earn a single penny on MSX-SM sales, I'm just a happy owner of a prototype and develop some stuff for it (unapi/WiFi) absolutely for free.

Yeah, I had made a workaround for it before, but I was able to remove it because KdL fixed the underlying issue.

That really depends on your definition of standard. This particular case was about interrupting on a line that is in the vertical blanking period, regardless of scroll offset. The official documentations neither says that it does or that it does not generate interrupts there (and the reality is somewhere in the middle). In other words, it’s ambiguous, and so are many of the other details around splits and other techniques. The documentation is simply not detailed enough to describe it.

If you limit yourself to only using things that are described, well that is your choice, but it e.g. essentially rules out pretty much all use of line interrupts because the way changes in registers during display are presented (e.g. does it take effect immediately or only on the next line, and can you safely use them during VDP commands?) is not at all described in the manual, it’s impossible to make splits without making assumptions on the VDP behaviour.

In the end it is the responsibility of MSX compatible system developers to provide a proper V9938 emulation. An official MSX3 and later by ASCII could’ve used the V9938 logic source code to provide a completely compatible simulation, however since this is not available to the community, the emulations (be it FPGA or PC) need to be based on reverse-engineering, documentation alone is simply not sufficient to base emulations on like any emulator developer will tell you, too many unspecified details are commonly relied on in practice by software.

Well, I meant more if MSX systems had continued to release after the turboR, with the upgrades being a true ASCII development building and improving on the chips they had at the time.

The OCM implementation was by Tsujikawa from ESE Artist’s Factory, I don’t think they had much support from ASCII in the development, very much doubt they were given access to the original V9938 logic designs etc (that would be wonderful tbh).

I am trying to use the RTCSAVE and get UNSUPPORTED KERNEL. How can it be used? No instructions.

SD generated by the sdcreate tool, and booting from SDBIOS. With the Zemmix BIOS I got a "not compatible BIOS" message so the BIOS should be fine.