Waiting to test that ROM!
very nice.
Obviously I'm very curious.
If you want you can email me the rom, I'll put it here online.
Sorry I haven't followed all you posted here about this thread anyway just comment
if Be-Bop Bout! could help you for something... (maybe already discussed before... )
The rom image is uploaded and can be found here:
home.online.nl/smastijn/sft.rom
For controls read the earlier post.
@ARTRAG:
Code already runs from ram. The only reason the emulator is faster is probably that the out is slowed down on a real machine.
And no I do not take into account the end values, so maybe there is some gain there. But this will only affect mode 1 & mode 2.
For all those who go and disassemble the rom please note that the routines are not fully optimized yet.
Especially the "byte" level routines still read from the same line list, and currently only skip the dot mask commands. So lots to gain here, also in line list size.
Anyway have fun!
Wow! Looking really good!!!
The smaller sprites are big enough!
yeah it does look amazing,
indeed the smaller sprites are big enough, in fact, i think the big sprites are way too big ^__^
the small sprites look more or less the same as the original sf2 size.
and at 20 fps this is already looking very very nice.
Good thing is that the player handler won't be slowing the game much.
You only need to read out keys, move the players accordingly, check for attack collision and perform an action.
The only problem will be adding music each interrupt.
I hope 15 fps is doable.
Thanx for spending your weekend on this, it is defenitely NOT thrown away time.
Once again, great job !!!!
Amazing!
@norakomi: GO FOR IT! It is a master piece that deserves an MSX (turboR?) version!
Indeed the bigger sprites are way too much for MSX screen resolution...
The rom image is uploaded and can be found here:
home.online.nl/smastijn/sft.rom
For controls read the earlier post.
@ARTRAG:
Code already runs from ram. The only reason the emulator is faster is probably that the out is slowed down on a real machine.
And no I do not take into account the end values, so maybe there is some gain there. But this will only affect mode 1 & mode 2.
For all those who go and disassemble the rom please note that the routines are not fully optimized yet.
Especially the "byte" level routines still read from the same line list, and currently only skip the dot mask commands. So lots to gain here, also in line list size.
Anyway have fun!
in all cases where you have multiple vdp commands using the final values can speed up the commands on r800 of about 30%
@GhostwriterP
I know it's just a test, but I noticed some interesting results while playing with the ROM. I hope this helps to improve it:
1) I also tested it using the turbo-blitter option of openMSX (set cmdtiming broken). I was hoping that all methods would run at 60fps. What is my surprise to discover that only the BlitMode-0 runs at 60fps. Other BlitModes seem to be slowed down by some other kind of temporization
2) On BlitMode-0, enabling the sprites causes a huge performance impact as expected. But many other BlitModes show no performance drop at all. This probably is related to the problem (1).
Note: Tests done emulating an FS-A1GT.
@sd
That probably because only mode-0 is fully depending on the vdp, all the others are slowed down by the cpu, either due overhead (filling the regs) or simply due the outing.