PCem

There's a lot of DOS sound stutter in V15 on lower emulated targets and most of it seems to go away when under Win9x. Supposedly it doesn't happen under FreeDOS

The dynarec was ruled out as it does happen on interpreter, though V15 also had a lot of timer changes...

The resolution slider in the V2 panel is intended to pick a resolution to change the refresh rate for. It doesn't affect game resolution choices. Beast Wars only supports 512x384 and 640x480 for Glide/D3D. It's functioning as designed for a typical Playstation port of the era still growing from earl...

Where a Pentium II is recommended, that's where the AMD K6/2 steps in for being the class equivalent and emulated. Pentium II != PMMX. PCem's Mobile PMMX CPUs are faster than desktop PMMX, however, in Real-Life(tm), most fast PMMX laptops tend to have some other bottlenecking factor (video chipset, ...

The new dynamic recompiler has nothing to do with the voodoo emulation choking on DirectDraw buffer writes. There's other games that slow down for the same reason (Omikron's menu, Mechwarrior 3's menu, Urban Chaos's ripples, Metal Gear Solid's water/ninja effects)

What i'm listing are games that use tweaked VGA video modes intended for square pixel ratios. PCem and DOSbox treat these modes as 4:3 for their aspect correction currently, causing balls to stretch vertically in TIM etc...

640x350 EGA (on VGA) was 4:3 for me then.

What were some games that intentionally displayed wide on an old 4:3 CRT? (and I don't mean cutscenes) There's only three I can think of: - Jazz Jackrabbit (letterboxed 320x199 60hz VGA ingame) - Mystic Towers (the "Wide" mode is a letterboxed 320x200 VGA) - The Incredible Machine (letterboxed 640x4...

That "dot pattern" is the ViRGE's overdrawn dither for alpha blending, so you're erroneously using the ViRGE for this game (you shouldn't!)

karr wrote: ↑

Sat 20 Jul, 2019 11:42 pm

pcem is now portable

it isn't anymore as soon as you mount anything. There is no relative path support.

In my period setup (1994) I used an OTI-011 driver with MSCDEX in DOS 6.22, but this seems to be regressed in v15 on some AMI boards, so I've bumped that up to an anachronistic Win95 mscdex+oakcdrom driver. If I could, i'd be using that old crusty Tottori Sanyo Electric driver as that's what my 4X M...

Sam & Max CD audio in that startup menu works for me too, in DOS w/ P75+SB16 at least...

Bumping this with more random screenshots Neverwinter Nights (No hardware mouse cursors on V2, so naturally it's unplayable. Game does run at "playable" speed (similar to having much better video hardware in a real K6-2) and graphically correct otherwise) Post-game completion of Unreal Return to Na ...

switching the CPU to WinChip2 interpreter might be a workaround as there's been a few rare times where the dynarec makes the Voodoo2 drivers get garbage.

The CMDFIFO crash indicates a failure at the Voodoo emulation level. Have you tried it with the voodoo recompiler disabled? Have you also tried it on Voodoo Graphics?

For 3D, there isn't (Virge emulation is in there more for playing some of the earliest pre-3dfx exclusive 3d game releases), but if you like high color depth in your high resolutions for the desktop you'll probably want to use one of the ATI Mach64 cards.

Note that Max Payne will have graphical issues with the Voodoo2 (real behavior). Graphic novels WILL be blurry and decals in maps won't combine properly, making their edges black. Also real behavior is that a K6-2+V2 won't be enough for Max.

I've compiled from the windows release's sources. I don't like sharing binaries and I recommend compiling it yourself (apart from that gcc version fumble... it's easy!!!)

Also i'm expecting the RPI4 to do 486DX40. Being VERY optimistic here... hopefully the VCVI's driver doesn't bottleneck as hard.

Apart from the problem that it's Jim Power, V15's dynarec seems to get a divide error on start. The interpreter works, though there's also a freeze issue, which probably could be a real issue. The freezing also happens in V14 and is less frequent on lower clock speed (i.e. 25mhz), but I don't have a...

abk4523 wrote: ↑

Fri 21 Jun, 2019 4:02 pm

I know. It worked pretty well for a while but the laptop eventually started locking up after logging in.

The behavior would be XP throwing a trap exception before the boot splash could even show up.

abk4523 wrote: ↑

Fri 21 Jun, 2019 11:48 am

I installed Windows Xp on a 486dx2 laptop from cd

XP will refuse to install and run on anything less than a Pentium.

abk4523 wrote: ↑

Thu 20 Jun, 2019 8:59 pm

I know some bioses, as early as 486dx2s, can boot from usb.

The earliest BIOS USB booting i've seen were chipsets for Pentium 4 systems.... Are you sure you're not confusing with a floppy-based bootloader? USB wasn't a thing at all in the 486 era implied. CD booting wasn't yet a thing either.

The K6 is not the source of your missing CD-ROM problem and is unrelated to this thread. Taking an installed Win9x from one motherboard to another will bring up IDE issues which would require either removing/reinstalling all device drivers, or (sometimes easier) a full reinstall. Relatedly, I did no...

Real behavior. Win95 will actually throw protection errors on CPUs that fast! There's a patch for it, but generally it's easier to install Win98 (or better) instead.

My PSX emulator preference is Mednafen. It's also a very good Saturn emulator too. Both are for faithful accuracy reasons with no hi-res cruft to sidetrack like about everything else before it (pSX and SSF being exceptions, but those are closed source)

The stuttering has nothing to do with roms. Check your sound buffer length and execution percentage (it'll stutter when dropping below 100%). Emulating a PMMX233 would call for a 4GHz desktop i5 for example

Is it the gold version or Nitro Riders/Pack? The Glide renderer in that only seems to behave well with V1's with older drivers (and has led to a bug report incident before with a V2 which turned out to be real behavior)

Due to limitations, it's a right click menu on linux.

Right-click the window > Disc > Change Drive A:

Since the M II is quite literally a 6x86MX rebrand with better voltage/heat (according to all those CPU collection sites out there anyway), it was rather trivial to do this. It's mostly dupes of the higher 6x86MXs with that last CPUID bit changed to reflect "M II / IBM 6x86MX", done by trial and err...

I don't have any timing info but I wanted to fill that empty AMD S7 list slightly more. Also curiosity and I was bored one day and I figured what the heck i'll make another preliminary, probably very embarrassing and uncommitable patch... What this does is add the SSA/5 and 5K86 K5 CPUs to the AMD S...

Make sure the rom filenames are lower-cased.

Unfortunately that's the latest choice I have through the Raspbian distro. I could try older GCCs at least probably as there's gcc 4.4-4.9 and 5 packages.


EDIT: Installing gcc-4.9 and editing src/makefile to point to that allowed it to successfully compile and run