Re: Current status of the Commodore Amiga driver
Kale
02/12/25 09:11 PM
CIA8520 memes (a.k.a. the I don't really comprehend this chip section)
- batman1 (Ocean Batman budget) writes a 0 to all TOD registers for even CIA (the one counting from hsync), wait for 1 vblank sequence then pretends that TOD is higher than > 0x115. With current logic this can't possibly never get right (i.e. I tried several CPU speed combinations and never got a bit 8 high), TOD stay in latched state between these two events.
- cardiaxx / cardiaxxa keeps looping at a timer B check that never trigger ICR bit 1 (game makes sure to mask all irqs by then). Specifically it writes 0x48 / 0x49 (TA with run mode) where m_count_b2 gets first set by m_ta_out then cleared by a mask if crb bit 0 is off. Note that this definitely isn't protection: it does the exact same thing with WHDLoad version ...
33
25,870
Read More
|
|
Re: Is there any other LED clock on MAME?
=CO=Windler
02/09/25 02:57 AM
Early electronic clock chimes are a barely researched fascinating topic. Unlike synths, they are still rated trash and no serious part of clock history. E.g. the "Kienzle Variogong" chime synthesizes only does "bim bam" from a crazy amount of logic ICs and mixing its 2 bell notes from each 4 squarewaves (with tuning trimmers) and analogue decay envelope (i.e. having 8 trimmers and 8 capacitors only to play 2 bells). A later chime was "Junghans Quattro-phon", which already was single-chip digital hardware, but it has 4 polyphonic clock sounds (2 chip variants exists) made from layered squarewave tones and the quartz clockwork still had a big brown design case with stylish knobs of those the designer certainly was proud of, while most later quartz movements became boring black cheap plastic things intended to be hidden and hopefully never to be looked at again.
If anybody is interested in emulating the first digital clock chimes, check out Kienzle Variogong (patent US4270200, DE000002753733A1, will likely need netlist) and Junghans Quattro-phon (patents US4271495, DE000002850286A1, single-chip likely needs decapping). Clocks with the latter were also sold under a couple of different brands, but the the big brown quartz movement with big knobs (Junghans W771, W772, Hermle 1217) is easy to identify.
3
413
Read More
|
|
Re: MAME 0.274
=CO=Windler
02/06/25 05:23 AM
Is there any genuine speed benefit in making MAME enforce now SSE4.2? Else this does solely spin the obsolescence hamster wheel further. (Though shalt not obey Microsoft.)
Technologies tend to spoil after a certain sweet spot of development. Typically it starts with the split between pro and household grade versions of a device class, making the latter artificially shortlived and functions dumbed down.
* sewing machines (many presets but flimsy plastic cogs jam by thick cloth) * washing machines (crack-prone plastic tank with unremoveable bearings, tiny power components self-desolder by overheat) * music keyboards (presets-only instead of editable sounds, button beeps spoil stage use) * TV (very fragile TFT screen, overheat, user+firmware in same flash memory, cloud dependency and spyware, OLED shortlived with "burnout compensation" always-online spy mode) * online devices (designed to spy by unremoveable battery, no lens cover, no mic switch, no mechanical power switch, remote maintenance, SSD causes unexpected dataloss (by write loops & poor cooling, unsuited to keep old devices unpowered), no screws to deter repair, pulsed microwave radiation).
Cloud obsolescence even affects vehicles (Van Moof bicycles got unusable by company bankruptcy) and household appliances those would never need app and online connection at all.
10 years later, bad properties of the household variants tend to seep back to the pro versions, or the latter get designed to be made unusable for non-professional users (enterprise SSD with only 2 weeks unpowered data retention time, non-wifi routers need connected PC to boot before corded phone works, LAN switches hold no setup data when unpowered etc.) Nonprogress should not be furthered.
11
1,268
Read More
|
|
Re: MAME on MacMini M4
estefan3112
02/02/25 08:33 AM
Hi all, just a brief "Thank You" for the implementation of the ARM64 Dynamic Recompiler into MAME 274! Performance really speaks for itself on my M2 Mac mini (v273 compared to v274): gradius4u: from 150% to 537% (OpenGl) gauntleg: from 133% to 600% (Metal) kinst: from 111% to 499% (OpenGl) Now many of the newer games could suddenly make sense on Macs :-). Cheers!
3
248
Read More
|
|
Re: HD6805V1 dumps
Guru
01/25/25 12:23 AM
It would be better to just send it to Sean Riddle as capturing the data using a logic analyzer isn't simple especially if you don't have one.
2
517
Read More
|
|
Re: x86-64-v2 performance
Vas Crabb
01/20/25 11:29 AM
I have mixed feelings about the MAME packages provided by Linux distributions. I understand that they’re a convenience for users, but they all change things and end up providing an experience we can’t really endorse. FWIW Fedora is one of the better distros in this regard – MAME packaged for Fedora is pretty close to stock MAME.
I had a particularly bad experience with the person making the package for Void Linux. Every time they asked for advice, they insisted on doing the opposite of what I recommended. I believe the Arch Linux package is based on the Void Linux package. I would recommend against using either.
Trying to work with the person who wanted to make Debian packages was incredibly frustrating, too. They seemed to have no aptitude for actually handling porting issues, and would immediately start demanding assistance any time they encountered anything that required any thought. They’d need hand-holding for everything. It was like they wanted credit for packaging MAME while pushing the actual work back onto me.
As for how much value the Linux packages provide, I’m not sure. Mac users are often mocked for being non-technical and needing a one-button mouse interface to do anything, but Mac users have had no trouble compiling MAME for themselves since the days when SDLMAME was still a patch set. I would expect Linux users could work it out.
5
563
Read More
|
|
Re: Trying to replace a sound sample in Golden Axe
Vag
01/17/25 08:23 PM
I couldn't sleep at night, thinking that the timer wasn't 100% accurate... So I changed the logic, it is updated independently, when the system timer is updated. The routine for the music just checks it now, it doesn't set it. I also have the routine run in any case, because there were times it wouldn't run (like during the screen being revealed, or when the logo is shown). Finding the right times for each song was really time consuming, I spent a whole night listening to the songs repeatedly (I survived). Now there's 100% accuracy in music during attract mode :-) Also, I managed to exclude the first time the logo is shown. This means it can play a sound when it's shown, without messing with the songs. The music starts afterwards, when the title screen is shown. The other times the logo is shown (by itself), music is playing and I have no other sound. I did that because I'm not sure if I will find a way to have both a sound effect and a song, without one stopping the other. So, even if I won't find it, I can have a "GreekRoms" sound (if I can also find a way to encode NEC ADPCM sounds). By the way, one of the songs loops, but after the song intro (all the others that loop, go to the beginning). It was hard to find the proper time for the timer for this one, but now everything works perfectly :-)
The next thing I'll try to do, is to find the code that shows the dripping blood and the decapitation (if it's possible) and enable it, depending on the settings :-)
83
10,244
Read More
|
|
Forums9
Topics9,331
Posts122,199
Members5,077
|
Most Online1,283 Dec 21st, 2022
|
|
These forums are sponsored by Superior Solitaire, an ad-free card game collection for macOS and iOS. Download it today!
|
|
|
|