That seems highly unlikely. Bryan's hitting 100% because, if everything's set up correctly, Windows 9x will start sleeping when appropriate, dropping the instruction count significantly.
The WoW emulation layer won't affect speed much, as the CPU emulation doesn't use Windows API calls anywhere.
sorry for a noob question: is EGA emulation broken?
Shouldn't be. What issues are you seeing?
Sarah, can you please make PCem not to use the host's time and date?
It will allow me to run Windows betas that require me to change the virtual BIOS time and date..
Learn to DATE, "betas=betta" kid
wrote:Shouldn't be. What issues are you seeing?
no display, gray window for a long while.
it eventually works, but I have to wait on every boot.
screenshot took at some seconds after clicking "Hard reset":
Hi there,
I think I found a bug in PCEm, running Windows 3.1x on PCEm V8.1.
When I run a DOS application within Windows 3.1x and change it from full-screen to windowed mode (CTRL+Enter), I get video corruption in the DOS application's window.
If I edit the PIF file for the DOS application so that it launches in windowed mode immediately, I get a random result ranging from occasional success through to GP faults/integrity violations or complete lock up (of the emulated system that is).
I have tested this on Generic AMI 386/486/WinBIOS configurations, with both Trident 8900D and Tseng ET4000 roms (so basically the core configurations available if you just stick with what's downloadable from Sarah's website). For each configuration I also played around with all sorts of BIOS settings, plus tested DOS versions 5.0a and 6.22 + Windows versions 3.1 and 3.11 with a variety of video drivers/resolutions/color depths. Even with a bare-bones boot (himem.sys only, no emm386/mouse/smartdrv/etc.) the bug still persists. Screenshots are attached so you can see what I'm getting.
I'm not looking for a fix, just thought I'd let Sarah and contributors know seeing as nobody has mentioned this bug (so far a I can see that is).
But while I'm here may I say how awesome this little emulator is. Sarah and contributors: you have a real talent, thanks for sharing this with folks like me who wouldn't even know where to start. I do code a bit myself, but it's all so high-level compared to this!
M
quick question: why is the 430VX not getting recognized even its rom is there in the \430vx folder (55XWUQ0E.BIN) for me? (And I compiled the build on my own with some patches).
No idea. Maybe one of your patches has broken it in some way? (any intention of sharing these patches btw?)
wrote:No idea. Maybe one of your patches has broken it in some way? (any intention of sharing these patches btw?)
those patches are mainly for the CPU emulation, not the configuration menu, also, I had to select in pcem.cfg the machine as 24 (machine = 24) to get to the 430vx, but the configuration menu appears blank.
Screenshot in the attachment.
SA1988 and Sarah: I ran into the same bug, and it was trivial to fix. Current PCem SVN has the PC Jr., which means the total number of machines is one higher than it was previously, but the variable that keeps track of how many machines there are wasn't incremented to match (or at least, not everywhere). I should have pointed it out when I noticed it, but I forgot--sorry about that! Sorry to be so vague about it, but it really is a quick and easy fix, so I don't imagine you'll have any difficulties taking care of it. If necessary, just ask and I'll hunt it down in more detail. In the meantime, my patch collection is coming along quite nicely, and one of the things I'm fixing is some sort of subtle protected mode timing bug that manifests when EMS drivers are loaded. Perhaps that will lead to a fix for the Win 3.x DOS box glitch that was just reported here? Probably a long shot, but I wouldn't know where else to begin hunting that down.
Thanks for that. Fix committed.
Any more details about this protected mode bug?
wrote:Any more details about this protected mode bug?
I don't want to speak too soon, as I'm still pretty early in the process of figuring out exactly what's going on, but the issue I'm seeing is basically this: when in DOS with EMM386.EXE enabled, the emulated system's effective execution rate is reduced to a small fraction of what it should be (often ~25%, though this number isn't quite constant). I've tested this with several versions of MS-DOS and tons of different versions of EMM386.EXE, and the behavior is always the same. The effects of this aren't always apparent, but it badly breaks software that depends on real-time delay loops for speed control. Even so, you should be able to verify it by looking at the MIPS value in the status window with and without EMM386 enabled. I may have been too hasty in declaring it to be a pmode core bug, but that's my current tentative guess. I'm not an expert on any of this, but imperfect or incomplete emulation of V86-mode would seem to be the obvious culprit. I think I can probably fix it, but if you get it done before I do, all the better (but please let me know, so that I don't duplicate your work 😀)
Another, but minor, issue is the CD track listing (CDAudio) when I play a CD Audio disc on PCem (mounted), it gets played well, but the track listing gets screwed up on win3.00a MME 1.0, win3.1x, win95, win98 and winme: for example, when Track 2 of a CD Music disc gets played, the track listing goes towards the successive tracks (like Track 3, track 4 and so on) even though what is playing is Track 2.
allowing nt to run would be a nice feature to!!
About critical bugs, I should say that most Windows 3.1 S3 Trio64 drivers fail to run properly with the emulated Number 9 330FX video card on PCem.
Noticed transparent textures in Direct3D Ramp get inverted in black & white. Bahamas 64, WinChip 240, DirectX 5, Win95C.
(also to be clear to everyone here, WinChip itself is a slow CPU so 240mhz isn't that impressively fast.)
I get following errors in log when running Civilization in dos,ami486 and trying to move a mouse (game hangs):
Serial data 43 34 08
Serial data 43 3E 01
Serial data 43 3A 06
Serial data 43 3D 04
Serial data 43 39 0E
x86 illegal 0011 80000011 017D:000007EA FF
x86 illegal 0011 80000011 017D:000007EA FF
x86 illegal 0011 80000011 017D:000007EA FF
x86 illegal 0011 80000011 017D:000007EA FF
What mouse driver are you using?
now, instead of going to successive CD audio tracks, for mixed mode CDs, when a track is played it doesn't start from 00:00 but rather to a random number (like 05:54, then goes, to 06:54, 07:54 every second), I wonder if it's a MSF issue as dosbox doesn't have such issue (I have looked at ide.c and cdrom-ioctl.c but found nothing or maybe I haven't looked well)
