WinUAE 3.5.0 beta series (was 3.4.1)

[DamienD]

Cheers Toni

[mark_k]

Originally Posted by NorthWay

Is that Comspec controller the one that Dragon's Lair supported?

My guess is, probably. Try setting up a config with Comspec controller and installing the game to it.

[mark_k]

Not sure whether this is beta-related, but I set up a config with emulated A1000 NTSC, Comspec SA1000, 2MB fast RAM and 40MB blank/empty HDF. Loaded Kickstart 1.3 then booted comspec-2000.dms to Workbench. Then pressed Ctrl-A-A to reset.

WinUAE kept showing dialog boxes with this text:
hpos = 257!?
hpos = 258!?
hpos = 259!? (and so on)
In the log output after each hpos = ... line was
Tried to capture mouse but window didn't have focus! F=0 A=0

On dismissing the dialog, WinUAE warps the mouse to the centre of the emulation window (I guess the hpos =... dialog box being open causes mouse capture to fail) then the next dialog box appears. No way to exit in that state other than killing it. Can't press F12 or click close box of emulation window since dialog box is open.


Edit to add: Also, does anyone else see the emulated Amiga hanging (i.e. mouse pointer no longer moves) most of the time on SCSI accesses, e.g. after running HD_Utilities, or (after defining a partition with HD_Utilities) selecting Initialize from the Workbench Disk menu? There are several lines like this in the log:
WD33C93 WD_DATA READ without data request!?

[DamienD]

Heya Toni,

Haven't been that active in testing the 3.4.1 betas so far, tonight I thought I better get more involved

...obviously I'm not about to test all 1690 games / configurations that I have in my WinUAE collection so thought I'd test my usual list of *difficult* games over the years that sometimes don't work when a new version is released:

A New Adventure - FaYoh 2 [AGA].uae
ACSYS - Autonomous Cybernetic System [AGA]   [unreleased] [demo].uae
After Burner (Activision).uae
After Burner (Sega).uae
Aladdin [AGA].uae
Alien Breed 3D [AGA].uae
Altered Beast.uae
Agony.uae
Army Moves.uae
Ballistix.uae
Disposable Hero.uae
G-LOC - R360 Air Battle.uae
Full Contact.uae
Hare Raising Havoc.uae
Line of Fire.uae
Roketz [AGA].uae
Moon Child [AGA]   [unreleased] [demo].uae
Obsession.uae
Second Samurai [AGA].uae
Trolls [AGA].uae

I'm happy to say that all is good

...except for "Moon Child [AGA]". Unfortunately the "OldMoonChild" part runs very slowly / jerky when compared with v3.3.0 or v3.4.0

I've uploaded the disks to The Zone! (use an "AGA - 2MB Chip & 4MB Fast RAM" config) but do let me know what I need to provide you with in order to investigate further

[HanSolo]

What was changed in hrtmon 2.37 in compare to 2.36 ?

[Toni Wilen]

Originally Posted by HanSolo

What was changed in hrtmon 2.37 in compare to 2.36 ?

http://whdload.de/whdload/Tools/

Only "real" updates probably are custom chipset color palette restore from "action replay hardware" (=uae or aca500plus) -> always correct colors after exiting on OCS/ECS Amigas. (AGA allows palette reads, OCS/ECS is write-only) and whdload support now works with 68000 CPU, previously it crashed.

Originally Posted by mark_k

WD33C93 WD_DATA READ without data request!?

This is normal (Others probably aren't, later..). Comspec for some reason does extra data port read. Probably used to flush "fake dma" hardware buffer.

EDIT: Workaround = Switch off cycle-exact.

[DamienD]

Originally Posted by DamienD

...except for "Moon Child [AGA]". Unfortunately the "OldMoonChild" part runs very slowly / jerky when compared with v3.3.0 or v3.4.0

I've uploaded the disks to The Zone! (use an "AGA - 2MB Chip & 4MB Fast RAM" config) but do let me know what I need to provide you with in order to investigate further

Heya Toni,

So have tested this further. winuae_3410b1.7z is good but with any beta onwards the issue occurs.

What else do you need from me?

[Toni Wilen]

Originally Posted by DamienD

Heya Toni,

So have tested this further. winuae_3410b1.7z is good but with any beta onwards the issue occurs.

What else do you need from me?

Config and log, as usual (You sure it isn't: "Some 68020 CE mode cycle-counting is back.")

[DamienD]

Originally Posted by Toni Wilen

Config and log, as usual (You sure it isn't: "Some 68020 CE mode cycle-counting is back.")

Good point, so I'll try unticking cycle-exact... be right back

[DamienD]

Move along, nothing to see

Yup, unticking cycle-exact fixed this.

...so the question is; a lot (but not all) of my A1200 configs have this ticked. Should I now be unticking this to be safe?

Don't really feel like testing 300 x AGA configs to see if there's an issue or not...

[Toni Wilen]

In most cases "memory cycle exact" is more than enough and it needs less CPU power. (Note that 68020 CE being "faster" was only temporary change)

EDIT: and this is probably very game specific anyway, it probably won't affect most games that much (or none at all).

[Toni Wilen]

Originally Posted by mark_k

Not sure whether this is beta-related, but I set up a config with emulated A1000 NTSC, Comspec SA1000, 2MB fast RAM and 40MB blank/empty HDF.

Fixed. New PIO WD33C93 support called INTREQ() from WD emulation thread. It is always bad to change custom chipset emulation state from separate thread but is really bad in cycle-exact modes.

[DamienD]

Originally Posted by Toni Wilen

In most cases "memory cycle exact" is more than enough and it needs less CPU power. (Note that 68020 CE being "faster" was only temporary change)

EDIT: and this is probably very game specific anyway, it probably won't affect most games that much (or none at all).

Noted. Cheers Toni

[freddy]

Toni can you put entries to allow translation

Thank you

[AnnaWu]

Originally Posted by Toni Wilen

Beta 8:
- Added two A2088T BIOS 4.0 (390657-01) ROM versions with different internal date (09/11/89 and 14/11/89).
- Added A2386SX 1.00.03 ROM.

Are these ROMs in The Zone!?

[huepper]

Yes they are, look at new expansion_boot_roms.zip.

[Toni Wilen]

http://www.winuae.net/files/b/winuae_3410b9.7z
http://www.winuae.net/files/b/winuae64_3410b9.7z

Beta 9: (RC1, 1-3 weeks to go)

- Comspec emulation was very unstable in cycle-exact modes.
- Amiga-side key(s) pressed when entering pause state was not correctly released when unpaused.
- Add UAE boot ROM (if not 128k full autoconfig device) fake ConfigDev as earlier (in diag init) to make sure possible rom configured MMU setup knows about the fake autoconfig device (for example m68k AROS with MMU enabled)
- Updated AROS ROM.
- Selecting built-in AROS ROM always opened console window (broke when console window serial input was added).
- Added A590/A2091 v6.1 ROM.
- H CPU history debugger command now also shows all DMA accesses if v DMA debugger is enabled.

[staropram]

- Added A590 v4.6 boot ROM to ROM scanner.

Are this ROM in the ZONE ?

[AnnaWu]

Originally Posted by staropram

- Added A590 v4.6 boot ROM to ROM scanner.

Are this ROM in the ZONE ?

For A590/A2091 v6.1 take a look into MAME/MESS BIOS set.
390722-01/390721-01

v4.6 is in progress ...

[bladecgn]

On "Chipset Extra" drop-down "A1200" is listed twice.