Whenever you stop moving the joystick, the penguin runs to the right. The joystick has been calibrated under Windows and works fine with other games. I've tried changing the timed= variable in the config file, it makes no difference. Tried this under Win98 and now under WinXP.
I suppose what's needed would be some way (eg. in the keymapper?) to change the mapping of positional values between the host and guest.
Are you using latest DOSBox?
Intel i7 5960X
Gigabye GA-X99-Gaming 5
8 GB DDR4 (2100)
8 GB GeForce GTX 1070 G1 Gaming (Gigabyte)
I confirm it happens in latest svn with both timed true and false.
I can't try for the moment, but maybe try with a very low cycles setting, like 500 or so.
or if you are brave, try changing the internal resistance of the emulated joystick
Water flows down the stream
How to ask questions the smart way!
cycles=520 and timed=true works although i am not sure whether it is not too fast for this game. Minuous you will have to check yourself.
Those settings you mentioned do appear to work under V0.72. (Speed seems about right but it is difficult to tell whether it is running at 4.77MHz like it should be, it would be better IMHO to measure the speed in MHz rather than arbitrary "cycles".)
But the joystick doesn't seem to work at all under V0.73, not even the firebuttons (tested with Pango and Alleycat).
delete mapper.txt
Yep, that's fixed it, works great now. Thanks!
rather than arbitrary "cycles"
cycles are not arbitrary.
He has a point though; an actual MHz display would be pretty useful.
What for? Just provokes a lot of postings "it says xyz MHZ but it feels wrong".
Well, it would make it easier to determine what a game needs for speed, as you could simply look at the system requirements, instead of trying to guess how many cycles to use... 😉
DOSBox counts instructions, not clock cycles, so variations in the clocks consumed by the particular instructions being executed make the relationship fuzzy between DOSBox cycles and MHz. However, approximations can be made, such as the utility that ih8regs made for the purpose.
So, is there some reason that DOSBox *doesn't* calculate the proper timings of each instruction in cycles, like every other emulator I can think of?
I assume it's because this would make the emulation overhead slightly higher, but is this really such an issue in 2010? To lose a small bit of speed in return for greatly improved accuracy and timing would seem to be worthwhile.
greatly improved accuracy ?
There aren't many games that care for it.
Water flows down the stream
How to ask questions the smart way!
So, is there some reason that DOSBox *doesn't* calculate the proper timings of each instruction in cycles, like every other emulator I can think of?
Which PC emulator are you talking about that has cpi accurate emulation?
Always appreciated: post a patch at sourceforge.
The earliest PC games are the ones that could most benefit from counting clocks, I think. When there was only 4.77 MHz, and not much worry about how the code would scale to different clock rates, it was possible to get away with counter loops and other such poor practices. Fortunate, then, that the lowest emulated CPU speeds have the most host CPU to spare for the extra processing that counting clocks would incur. Something like cputype=8088_clocked would be nice to have, but don't hold your breath waiting for someone to code it.
>The earliest PC games are the ones that could most benefit from counting clocks, I think.
Yes, they are the ones I play the most 😉
>Which PC emulator are you talking about that has cpi accurate emulation?
I meant emulators in general, rather than specifically IBM PC emulators.
>Always appreciated: post a patch at sourceforge.
I might do this in the near future then...
