138 replies to this topic

[Mclaneinc]

He's got to have a look at a couple of bits / features / bugs on the new 1200 emulation in the lastest beta first

[lemiel]

Wow! Thanks. It works.
Had tested different software, not everything works, but it should be also tested with my second drive (Tygrys Turbo - Turbo, US and Synchromesh), which I own not so long, because probably they could work only with TOMS Drives - sends special code execute on Z80 in drive. Have to check and copy all software to PC.
The same about this initializer. I did not analysed it - it was enough usable in early nineties and works with drive in turbo. I have three more working with Turbo drives. Two based on / similar to on German SpeedStart from BiboSoft and one different.

carmel_andrews, on Tue Mar 20, 2012 7:26 AM, said:

what other 810 or 1050 drive mod's you thinking of trying to emulate phaeron, what about isplate 1050 or the SA/SA2...or of course speedy/super speedy....that's enough to be getting on with

I think that main need is to emulate speed options, and this is now, we have all modes now - UltraSpeed (US), two speed Synchromesh (Indus, XF551) and Turbo.

Other not general like format, etc. commands are for this moment not specially usable.
(edit) And I do not know which other commands are supported for example for Speedy1050.(/edit)

ISplate like most of extensions is using UltraSpeed, so is emulated. Maybe not with original speed 118kbps (HS index 1), but what for - is there any special software to use with this one and only speed?
I am not sure about speed of SA/SA2 but it is probably US 52kbps to.

Happy is an extra option with same speed as US Doubler, I am not sure why, but assume that probably because of $48 command needed to configure high speed write.
And with SpartaDosX I can see that switching drive emulation mode from selected at cold start US Doubler to Happy "online" is switching high speed off, but from Happy to US Doubler high speed is all the time. Because drive was not configured by indus.sys driver initially?
I do not know at which speed were working early Happy 810 versions - probably likeit it said in AtariSIO with 38kbps, but which protocol?

I can see two more emulation modes eventually usable for special purposes (fun coding, etc.):
1. with freely set generic UltraSpeed speed,.
2. with both generic Ultra Speed (do not know which speed, maybe free set to) and Turbo. Similar to TOMS Multi and Tygrys enabled drives. - Have to check on real drive with Tygrys one of modified SuperDOS 3.7T, which starts with Turbo, same way as initializer - not fallback to normal on non turbo drive, but after loading dup.sys all operations are in normal - maybe it tries to use ultraspeed?
Or only second option.
Of course it can bee also fun (but for how long) to have possibility to set different emulation modes and speeds for every 1-15 drive. But...

Edited by lemiel, Tue Mar 20, 2012 12:13 PM.

[phaeron]

There were a couple of reasons i put in the drive emulation options:

• High-speed SIO routine testing. Almost no one has all of the different types of drives to test on, and the result is that people frequently write HSIO routines that malfunction on some drives.
  
• Software that makes improper assumptions about drives. I realized a while back that it was impossible to support all software with a single generic drive emulation profile because of this problem -- programs would assume that because a specific high-speed index value was returned or that a particular command succeeded that all of the rest of the functionality for a particular drive was supported. This is compounded by HSIO routines that don't have proper fallbacks, as well as actual hardware that doesn't do proper validation (XF551 Write PERCOM Block command).

The Fastest Possible setting is basically the extreme stress test -- if your HSIO routine passes this, then it should be able to handle any speed. There are a couple of HSIO routines that actually can work at this speed, such as the one in MyPicoDos, although you may need hardware modifications for it to work on real hardware.

Note that switching the emulation level on the fly isn't a good idea. SpartaDOS X, for instance, tests the drives on first use and remembers which type it has found, so if you switch a drive from US Doubler to XF551 then it will get confused and probably not work, or at least not use high speed mode properly.

What Altirra doesn't actually do is emulate the drive CPU itself. This means it can't handle any extended command that involves writing code to the drive or otherwise executing from specific addresses. There are a couple of reasons for this, one being that I simply don't want to emulate another CPU -- never mind the fact that it can be 6507, 65C02, 8048, or Z80 -- and another being needing uncommon ROM images to do so. I can add emulation for extended commands that don't require full emulation of the drive, but usually I find that the software doing this needs the full enchilada anyway. This is also why you don't need to initialize the drive like you might on real hardware, such as loading actual working Synchromesh code in Indus mode -- the emulated drive doesn't need it.

[Mclaneinc]

I know you are a stickler for exact emulation and that's the way it should be but I've always failed to see the 'need' to fully emulate either a Happy (or the clone, the Lazer) or an Archiver. The only people who actually need to play with these drives are the people with real hardware, it makes no sense to emulate them fully in my book.

In this day and age of images and proper emulation and of course the attempt to archive real disks more exactly via Vapi stuff the need to use a software archiver or happy seems pointless. Of course if it was about exactness then maybe but at the end of the day who would actually 'use' the emulated Happy for more than 5 mins as novelty factor, after that all the time coding it would be wasted.

Well that's my thoughts..

[w1k]

can i record XF551 sounds for altirra?

[phaeron]

Thanks, but I already have XF551 sounds implemented -- it's enabled if you switch the drive emulation mode to XF551 mode.

[w1k]

ok, do you plan new version?

[flashjazzcat]

Emulator keeps crashing under Windows 7 when logging back in after locking the PC. 2.10-test17.

[w1k]

i know about latest version 2.10-test24

[flashjazzcat]

Nice, but I don't. Latest one mentioned here is test18.

[w1k]

http://www.virtualdu...2.10-test24.zip

[flashjazzcat]

Thanks.

[phaeron]

flashjazzcat, on Fri May 25, 2012 12:13 PM, said:

Emulator keeps crashing under Windows 7 when logging back in after locking the PC. 2.10-test17.

I was able to find and fix one cause of this:
http://www.virtualdu...2.10-test25.zip
http://www.virtualdu...-test25-src.zip

[Mclaneinc]

Other bits

Changes:
- Now includes the 64-bit build. There really isn't much point to this since the 32-bit build is faster, but someone asked for it and I'd had it building for a while anyway.
- Fix for the 16-bit cheat issue.
- Fixes for some sprite issues, including the one that masked the glitch in the diarrho Quantizator picture. Sprite overlap is now supported. Unfortunately these fixes do slow things down a little bit.
- BASIC is no longer shown on the caption bar for machines that don't support built-in BASIC.
- Fixed a mode switch related display crash on XP machines.
- Several updates to the debugger.
- Fixed long filename mangling issue in H: device (it was needlessly mangling some filenames between 6-8 chars).