WinUAE 2.4.0 beta series

[sink]

thk you Toni.. this beta 14 make my day!!

[Dr.Venom]

Originally Posted by Toni Wilen

http://www.winuae.net/files/b/winuae_2400b14.zip

Beta 14:

A small issue seems to have slipped in between Beta 13 and 14. In Beta 14 the scrolling is not smooth in non-laced screenmodes when using DirectDraw. Tested with Jim Power. In Beta 13 same config has smooth scrolling. Logs + config attached.

[Toni Wilen]

Originally Posted by Dr.Venom

A small issue seems to have slipped in between Beta 13 and 14. In Beta 14 the scrolling is not smooth in non-laced screenmodes when using DirectDraw. Tested with Jim Power. In Beta 13 same config has smooth scrolling. Logs + config attached.

Before I check any logs and config: all buffer settings have same problem?

[Dr.Venom]

Originally Posted by Toni Wilen

Before I check any logs and config: all buffer settings have same problem?

Yes this happens in all three buffer modes.

Update: I found something. The issue in Beta 14 can be solved by explicitly selecting "50hz" in the display tab for the screenmode, instead of having it at "default refresh rate". Config and log attached.

In Beta 13 both "default refresh rate" and "50hz" work correctly.

[Mequa]

I couldn't get CD32 games to boot with JIT enabled in the latest beta, using quickstart options and Amiga Forever ROMs.

The 2 most compatible CD32 quickstarts (without JIT) work fine. Least compatible CD32 quickstart only works if JIT is unchecked.

not work ppjoy LPT

[Toni Wilen]

Originally Posted by rekk

not work ppjoy LPT

winuaebootlog.txt needed.

EDIT: not needed, I think I found the problem already.

[Toni Wilen]

http://www.winuae.net/files/b/winuae_2400b15.zip

Beta 15:

- Append 'i' to PC resolutions that are interlaced. (TV Out, VGA + 15KHz hacks etc..)
- Low latency Amiga interlace mode odd/even fields are now in sync with output's odd/even fields, if output resolution is also interlaced.
- Set cpu speed to "real" if it was fastest possible originally when switching cycle-exact on the fly using uae-configuration.
- MIDI device IDs are correct again (b1)
- Display panel displaydata PAL/NTSC/etc select box now mirrors currently active emulated setting.
- "Default refresh rate" skipped every other frame in DirectDraw mode (b14)
- gfx_autoresolution_min_vertical and gfx_autoresolution_min_horizontal support on the fly adjustment using uae-configuration.
- Write winuaebootlog.txt to Windows temp directory if original directory is read-only.
- Directory harddrives configured as unbootable are now really unbootable under KS 1.3.
- Mequa's optimized Akiko C2P added.
- HID raw input mask overflow when HID bitsize was 32. (PPJoy)

Thanks Toni, PPJoy is OK

[Mequa]

(Moved from Akiko C2P thread as this is directly relevant to WinUAE 2.4.0 beta)

In WinUAE 2.4.0 Beta 15, CD32 games boot fine without JIT, or with JIT enabled immediately after the CD is recognised. Just not with JIT enabled from the beginning (with CD32 or Generic chipset).

But in WinUAE 2.3.3, CD32 games do boot with JIT enabled from the beginning, so this looks like a tiny regression in the beta.

[mingle]

Okay, continuing on from my other post about grey scanlines, I've just tested good old SysInfo on both WinUAE 2.3.3.0 and 2.4.0.0b15 and aside from the difference in speed (assuming it's right!), the 'mips' rating is showing a rather strange "004.75" as opposed to the couple of hundred you'd expect.

See attached screen-shots.

Cheers,

Mike.

[Mequa]

Could be an overflow error perhaps due to higher values than SysInfo expects?

LOL @ Phone Me NOW!!!


P.S. Would you mind trying it with the new Vsync enabled (with JIT+Fastest Possible) to see if that gives a major difference in CPU performance?

[mingle]

@Mequa,

I tried the settings you suggested and it actually went slightly slower.

The mips was showing 975.40 - which explains the "004.75" I got previously - obviously the text field for the mips reading is only 6 characters :-)

The other strange thing is the differences between the mips, mflops and the comparisons against the various amiga models between WinUAE2.3.3.0 and 2.4.0.0.b15 - for 2.4.0.0.b15, the mips are way up, but the mflops are down, as is the speed comparisons vs other amiga models.

[mingle]

Okay, since I'm giving v2.4.0.0.b15 a workout, I've also noticed that full-screen native modes now appear to be alinged on the left side of my monitor (running at 1920x1080 in WinXP).

In the attached screen-dump I was running SysInfo with the settings shown in the two WinUAE program windows I super-imposed on the picture.

In WinUAE 2.3.3.0 these settings result in the native amiga screen appearing in the centre of the screen, as I'd expect.

I did try tinkering with the "Horiz. position" slider, but this seems to 'stretch' the native screen, rather than moving it (a bug?)

The "Vert. position" slider seems to work correctly, however and will move the native amiga screen up/down as expected, without stretching it.

It would be good to have an check-box option to automatically centralise (both vertically and horizontally) the native mode screens.

Cheers,

Mike.

[mingle]

I also just ran a quick side-by-side comparison of WinUAE 2.3.3.0 and 2.4.0.0.b15 using SysSpeed v2.6.

I always find that program to be very difficult to use, but I did manage to call up a comparison between the two.

On identical hardware, using the exact same workbench install and config file, v2.4.0.0.b15 comes in at around 80-85% the speed of v2.3.3.0

[Mequa]

Originally Posted by mingle

On identical hardware, using the exact same workbench install and config file, v2.4.0.0.b15 comes in at around 80-85% the speed of v2.3.3.0

Is that without Vsync?

[mingle]

Yep, without v-sync (I'm assuming you mean the v-sync option under the display settings?)

When I enable Low latency VSsync for Native only, the speed drops by around 50%.

When I enable Low latency VSsync for RTG only, the speed drops by around 5%.

When I enable Low latency VSsync for Native and RTG, the system won't boot - just get a blank screen - or if I change the settings while the emulation is running it freezes.

I'm running ClassicWB, which does have things like FBlit and other stuff running in the background. Are any of these likely to cause problems?

I might actually create a squeaky-clean WB3.1 install and try that out...

Cheers,

Mike.

[Toni Wilen]

Do not bother with SysInfo, it is bad test for emulation (and for real 68040+ CPUs too!)

Use something that actual does something (like using lha to unpack or pack files) and time it using a stopswatch to confirm.

And if there is actual difference (>~5% or so), try to find which beta introduced the slowdown, thanks.

[Toni Wilen]

Originally Posted by mingle

In the attached screen-dump I was running SysInfo with the settings shown in the two WinUAE program windows I super-imposed on the picture.

But it is centered in the attached image?

I did try tinkering with the "Horiz. position" slider, but this seems to 'stretch' the native screen, rather than moving it (a bug?)

I can't duplicate, works fine here.

It would be good to have an check-box option to automatically centralise (both vertically and horizontally) the native mode screens.

Do not talk about centering before reading all other similar thread It isn't that simple at all. Autoscale settings attempt to do that.

EDIT: Logs also needed. ~10-20% speed loss in fastest/JIT vsync modes is usually normal, there isn't any easy ways around it.

[mingle]

Originally Posted by Toni Wilen

But it is centered in the attached image?

Hi Toni,

The above image (native_fullscreen.png) is centred vertically, but not horizontally.

The 'turquoise' background represents that actual size of the screen of my system. The Amiga native PAL screen is aligned on the left-side of the screen.

Thanks for the other comments. I'll have a look at the previous versions if I get a chance.

Cheers,

Mike.