WinUAE 4.9.2 Beta Series

[AMIGASYSTEM]

Author: Toni Wilen


Download WinUAE 4.9.2  Beta 9:

http://download.abime.net/winuae/files/b/winuae_4920b9.7z
http://download.abime.net/winuae/files/b/winuae64_4920b9.7z

A new era begins for WinUAE, listing all the current and future improvements and progress is not easy to do in a few words, attached is the official quote from Toni Wilen.

After some testing I can tell you that the Power in RTG is now remarkable, UAE-GFX and all Picasso96 functions are now fully "hardware accelerated" (uses native host-side code), this means that even high resolution videos will be smooth, as will RTG Games.

If before no Expansion Card even the most powerful ones could not match the power of WinUAE, now they will be even farther away, plus the new UAE-GFX Picasso96 features are now optional (all enabled by default). Configuration files only.

Other big news on KickStart, now WinUAE will also recognize and be available in the SyStem ROM Setting the new KickStart 3.1.4, 3.2 and 3.2.1.

News and improvements also for the RTG board aVRAM, Simplified JIT Direct Exception Manager.

CyberStorm MK I update: Accelerator and SCSI have been split, The CyberSCSI module is now an expansion board and has a separate ROM image (works similarly to the Blizzard SCSI kit).

At the moment with the new WinUAE it does not work OS4 and AROS 68k (reported by me), probably very soon there will be a fix


You can read Toni's full review here Wilen:

https://eab.abime.net/showthread.php?t=112041

[G-linx]

I think the word "wow" sums it up nicely.
However, I'm still AROS all the way!
Who do I talk to to get j-uae updated??

[AMIGASYSTEM]

As expected, Toni's fix arrived, AROS One 68k works perfectly on the new WinUAE, I would say much faster and smoother graphics, also SDL works well, see screnshot (NetSurf SDL)

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 2: 14/10/2022

- Display resync blanking was stuck enabled (~1s blank screen when mode switches)
- Internal emulated serial devices (for example laserdisc player) data transfer was unstable.
- Z3 hardware emulated RTG board autoconfig size data was fixed and then was broken again.
- Some NCR710+ SCSI controllers hung.
- Serial port status bits were not correct in all CPU modes.
- Internal CPU instruction emulation table refactoring, use separate table for instructions that return total cycles and instruction that count cycles internally (cycle-exact modes).
- Quickstart panel floppy type dialog was partially clipped if GUI was using smaller default size.
- D3D9 full install was accidentally required (d3dcompiler_47.dll error message).
- Automatic fall back to GDI mode (if both D3D11 and D3D9 is not available) retried forever.

Pre-4.9.2 fixes:

- Box art path detection didn't parse cdimage0 paths correctly if path had "," character. ("," needs special case because path can be optionally followed by ",delay")
- If only one autoconfig device was enabled and it was removed when emulator was running and then system was hard reset: old autoconfig data was not cleared causing disabled device to get misdetected by AmigaOS which usually caused hang.

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 3:24/10/2022

- 68000 IPL check position is now almost fully accurate (Still some special cases to check and test). Very big thanks to fx68k author!
- W debugger command hang fixed.
- Full drive/RDB mode is now normal checbox. Hopefully it is now less confusing than button that didn't really look like a button. If selected HDF already has RDSK identifer, checkbox is ticked and disabled. Checkbox can be only ticked/unticked if selected HDF does not have RDSK identifier.
- GDI mode didn't force 32-bit screen depth mode.
- FloppyBridge uae-configuration change now really updates all internal variables.
- Screenshots didn't work if 256 color checkbox was not ticked (tried to create 24-bit palette indexed image)
- DMACONR blitter status 1 cycle delay was broken (Circle scroller / UFO)
- Enabling DMA debugger and then returning to emulator continuously reported conflicts until next frame had finished. DMA debugger init was changed but tables were not fully initialized.
- Keyboard reset now behaves more like real hardware. Press keys, hardware gets reset (if keyboard reset warning is not active), reset continues only when at least 1 key is released. If all keys are kept pressed more than 5 seconds: do hard reset.
- DMA debugger CIA access line includes also E-clock state (0, 2, <data access info>, 6, and any "wasted" cycles needed to sync with E-clock are marked with "-".
- DMA debugger is wider. More space for future features.
- New cputester tests done, for example all SR modification instructions (including STOP) tested with both trace + pending interrupt. No errors found.
- If STOP was executed without trace active and STOP enabled trace, STOP still stops normally and only interrupt can wake it up. (b1)
- Blitter busy bit read incorrectly as blit not busy if blitter was in non-nasty mode and not line mode and CPU stole second to last D channel slot.
- Removed all remaining Windows 7/Vista checks (from days when XP and Vista was still supported).
- "Remove interlace artifacts" is broken. Will be fixed later.

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 4: 01/11/2022

- Remaining IPL fixes (LINK, UNLK, MOVE to SR/CCR, AND/EOR/OR SR/CCR, CPU wait state IPL fix).
- Emulated Paula side-effect that affects IPL timing. IPLx line state changes from low to high (IPL is active low) are about 0.5 CCK later than high to low transitions. This delays CPU IPL change detection by 2 CPU clocks if any IPL line had low to high transition. CPU accepts IPL change only if has been stable for 2 CPU clocks. (for example 3 to 5 has delay, 4 to 5 does not)
- Enable only WASAPI audio by default.
- Debugger break point command accepts number of hits before breaking to debugger (f Nxx <address> where xx is number of hits required)
- uaegfx overlay limit checks compared Amiga side VRAM end against physical host side bitmap end which accidentally passed if 64-bit Windows. Overlay window was usually empty if 32-bit Windows. (It might have worked in some configurations when using Z2 uaegfx)
- GUI Help button is now always available and opens if exists: (in priority order) winuae.chm, winuae.pdf. If neither is found, help web page is opened (without asking first like previously)
- It is still possible to run windowed WinUAE if desktop is 16-bit but it can cause graphics problems. Not sure if this needs to be supported anymore.

68000 based A500 like unexpanded (Fast RAM expansion is supported) configurations should be now 100% cycle-accurate. (If there are no bugs, many edge cases need complex and very carefully timed cycle accurate tests that are not easy to do)

Note that this does not mean all features are fully emulated, for example OCS and ECS Denise mid-screen resolution changes are not pixel accurate. Accurate as in if resolution change is not carefully coded, there can be few extra or missing pixels when resolution changes. AGA does not have this side-effect.

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 5: 06/11/2022

- If copper used last cycle in scanline and last cycle was even cycle (PAL 226) and it wanted next possible copper cycle and it was free: copper will allocate cycle 1 (Copper won't use it, neither can CPU or blitter. Cycle is allocated because it is 2 cycles from 226 but it is unusable because it isn't even cycle) and cycle 2 (copper uses this normally) of next scanline. This was implemented before 4.9 but was lost when custom emulation cycle allocation rewrite was done.
- Reading from non-existing register always returned FFFF instead of data that previous cycle's possible DMA access transferred. (Equipose / Complex hang)
- Don't keep "interrupt active" flag set if IPL is >0 but CPU interrupt mask is higher or same as current interrupt level. Only check it if mask gets lowered. Prevents unnecessary status flag checks between each emulated instruction.
- Fill color palette entries with pseudo-random contents at power up. (Except COLOR0 to not get annoying color flash at startup. Which does randomly happen on real HW.)
- EHB mode OCS/ECS vs AGA difference emulated. OCS/ECS number of active plane does not affect EHB (only HAM, DPF or KILLEHB disables EHB). For example if plane count changes mid scanline from 6 to 5, EHB stays active and remaining data in plane 6's shifters still appear as EHB plane. But on AGA plane count change causes switch to normal color mode immediately, plane 6's remaining data selects color palette entries 32 to 63. (G.Rowdy / Desire)
- In some situations if same scan line enabled and disabled EHB, whole scan line was drawn without EHB. (G.Rowdy / Desire)
- Replaced C-library pseudo-random number generator with simple xorshift algorithm.
- Ultra extreme overscan mode screenshot was vertically clipped.
- Remove interlace artifacts option fixed. (b1)
- Added optional chipset interlace mode filter panel settings. If enabled and at least one entry is different than normal native mode filter panel, filter is selected when current native is interlaced and switched back to normal native filter when mode switches back to non-interlaced.
- Added semaphore locking to A2065 emulation, previously it was possible to free A2065 resources while some other thread (for example slirp) was currently accessing its on-board RAM.
- Fixed serial port MIDI emulation hang.

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 6: 19/11/2022

- Interlace mode filter crash fix (b5).
- Disk read DMA started (without DSKSYNC) and reading "nothing" (no drives connected): disk DMA never finished. (CDTV boot hang without DF0:)
- CIA B timer counting timer A underflows or CNT: generated interrupt was 1 timer count too early.
- 68060 without FPU and executing FPU instruction that uses -(an) or (an)+ addressing mode: address register's original value was restored twice when F-line exception stack frame was being created. Second time it used invalid register value causing D7 to be modified.
- Audio interrupts and AUDxDAT processing delays fixed. They were too long (2 CCK) when correct delay is only 1 CCK. 2 CCK appeared to be correct because previously interrupt timing was wrong.
- Keep CPU halted when keyboard reset keys are kept pressed. (b3 reset change update)
- EHB mode was unreliable in AGA mode (b5)
- 68010+ BKPT illegal instruction exception had stacked PC pointing to next instruction but it should point to BKPT.
- 68010 DIVU overflow condition undefined N flag behavior updated. (DIVS overflow undefined flags are still not fully correct)
- Copper/blitter bug tested, updated and confirmed and is enabled by default again (only if 68000 and cycle-exact). This is the infamous chipset bug that can cause copper pointer to blitter pointer copy if copper was waiting, CPU wrote to COPJMPx during odd cycle and blitter was active. Not 100% accurate (modulos are not added correctly when glitch happens) in line mode because line mode internal "micro-operations" are not fully correct (exactly when pointers change, modulo is added etc, invisible operations that don't affect line mode output).

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 7: 26/11/2022

- When on the fly changing floppy type from FloppyBridge drive to normal or vice versa, disk change is now simulated automatically.
- floppyXprofile (x=drive number) config file entry can be used to change drive type on the fly using uae-configuration. Replaces floppyXtype which uses magic number parameters (it is still supported). Floppy type can be for example "35dd" or "35hd" or "floppybridge". (TODO: floppybridge:<parameters> support)
- CPU instruction prefetches are marked with "I" and data accesses with "D" in DMA debugger (for example "CPU-RWI")
- Do not allow JIT FPU option if no FPU is selected.
- If copper used last cycle of frame: following copper instruction (=very last instruction of copper before it restarts) was shown incorrectly in DMA debugger (visual bug only).
- Special empty copper cycles where copper allocates the cycle but leaves it unused are now marked with "C" in DMA debugger.
- Disk DMA write to nowhere (no drives selected) never finished. This (and reading which was fixed in b6) was working in older version, probably broke in b1.
- Disk DMA write DMA slots now have correct order (Read DMA: slots are filled from right to left, write DMA: filled from left to right).
- Floppy index sync (CIA-B flag) CIA interrupt bit was "sticky" in some situations. (b1)
- CHD CD audio sync fix. (Always use logical block access method)
- Some filter options still used non-interlace filter mode settings when interlace mode filter was active.
- D3D11 mode integer scaling artifacts fixed. (b1)
- D3D11 shader support code leaked ID3D11RenderTargetView, causing problems when shader or parameters were changed on the fly.
- Double clicking windowed mode window title bar always switched to fullscreen mode (should only switch if Shift pressed=Fullscreen or Control pressed=Fullwindow is pressed). Possibly Windows version dependant.
- DIWSTRT and DIWSTOP same vertical start and end was unreliable (James Pond 2 intro glitch)

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 8: 03/12/2022

- RTG to chipset mode (non-interlaced) switch didn't enable native mode filter (b7)
- If CIA-A TOD counting was PSU tick based, ticks that were delayed until next scan line (due to CIA internal TOD counting delays) were lost. (b1)
- Added GVP A1208 Z2 RAM + SCSI board. It is basically GVP Series II Z2 SCSI controller + RAM expansion in different form. Uses usual GVP Series 2 SCSI ROMs. If A1208 is emulated and Z2 RAM is enabled, Z2 RAM gets GVP A1208 Z2 RAM autoconfig data. A1208 GVP internal device ID is 0x97.
- JIT mode interrupt handling restored, JIT needs "different" interrupt handling than non-JIT modes for best performance. (b5)
- uaegfx blitter emulation was accidentally disabled since b4.
- uaegfx b1 update had wrong operator in masked 8-bit DST=SRC blitter function (Turbotext scrolling in 8-bit) (b1)
- uaegfx non-basic blitter functions (for example mask set or not commomly used) didn't handle overlapping source and destination correctly. (Turbotext scrolling in 8-bit mode)
- gfx_left_windowed and gfx_top_windowed config file entries replaced with gfx_x_windowed and gfx_y_windowed. Reason was swapped coordinates but to keep backwards compatibility, new config entries have been added instead of fixing it by swapping them. Old config entries still work.
- A4000 Quickstart mode set PSU tick CIA-A TOD timing but A4000 does not have PSU tick hardware.
- 68020 cycle-exact timing adjustment (Lionheart whdload slave v2.3)
- Increased A1200 68020 approximate speed slightly. (Many AGA demos had slowdowns). CE mode is not changed.
- Rare SupraDrive 2000DMA boot ROM was finally dumped. Support added.
- Debugger memwatch command parsing changed. Some parameters got skipped. Value match is now V<value>.
- Debugger string to integer parser error condition support added. Potentially dangerous commands are now aborted if expected address can't be parsed (like 'g <typo in hex address that does not parse>'), previously failed string to integer conversion returned zero value.

This version should restore performance (CPU and uaegfx) back to 4.9.1 levels.

If there is no major issues left, official should get released in next 2 weeks. Plan that rarely works but maybe this time..

[AMIGASYSTEM]

WinUAE 4.9.2 Beta 9: 07/12/2022

- Direct3D9 mode overlay mode didn't show Amiga screen in most configurations.
- JIT interrupt handling fix part 2.
- Debugger SMC detector cache flush does not anymore clear whole SMC state array (which can be huge if Amiga RAM is located in Z3 space). When state array needs to be reset, reset only RAM and ROM regions. Max SMC array size is now automatically same as highest RAM address.
- Autoresolution didn't do anything unless gfx_windowed_multi or gfx_fullscreen_multi had non-zero values (b1)
- Autoscale worked incorrectly in programmed modes. (b1)
- Disable GUI Wait for blitter checkbox in 68000/010 cycle-exact modes. Previously it was allowed but did nothing.
- CDTV/CD32 Quickstart didn't use new NVRAM Files -path (b1)
- Added Supra AMAB2 ROM to ROM scanner. (Seems to support at least 2000DMA and AutoBoot models). AMAB1 is still missing, previously added 2000DMA ROM most likely isn't AMAB1 because it only supports 2000DMA. 2000DMA ROM selection now also lists all AMAB roms because AMAB2-4 have 2000DMA support (AMAB5/6 dropped 2000DMA support) and ROM scanner selection logic does not support filtering ROMs that have same type