Hi all,
I hope everyone knows about the kelvar unREAL engine written by Tomasz.
I was testing it out but I didn't get to see how it works
DosBox : Showed errors saying Ilegal Read from FFFFFF1, FFFFFF2 CS:IP 0:0x6B
Virtual PC : Once showed divide overflow, after that halted.
Virtual Box : I could only see it saying kelvar unREAL engine
extended memory <some number>, after that the text faded (quite impressive!) but I could only see a black screen.
I am using MS-DOS 6.22, with HIMEM (if that causes a problem)
Thanks
Sid123

This is an older version of the unREAL engine that is still used by the DOS version of fasm, when DPMI is not available. The improved version that fasm uses first checks whether the processor/emulator is capable of executing the 32-bit instruction encodings in real mode and it shows an error message appropriately. So I would suggest that you first try to run the unREAL fasm in the emulator of your choice (make sure that there is no DPMI server in your DOS, because fasm is going to use it instead of unREAL if available), and if it shows 'processor is not able to enter 32-bit real mode' message, then "kelvar" will not work in that environment either.

However if you are able to run the unREAL fasm but "kelvar" still does not work, then it means that there may be some other problem that I am not aware of.

Hi,
FASM.EXE works fine. But I get an extra string at the end of the program which is

Is this normal?
When I run FASMD.EXE it shows No DPMI Services available.
Thanks
Sid123

you have to download the CWSDMI
http://homer.rice.edu/~sandmann/cwsdpmi/
and run the CWSDPMI.EXE
after then run fasmd again

Hi,
Thanks for the link, but as Tomasz said the unREAL engine used by FASM.EXE and KELVAR are the same.
If unREAL FASM can work, I can't figure out why the unREAL kelvar engine won't work.
I am getting a DOS Boot Disk ready to test it on a real 386 and would try to dig up what's causing the problem.
Will also try it on Bochs and see what I get.
Has anybody tested it on their machines/emulators?
Thanks
Sid123

> If unREAL FASM can work, I can't figure out why the
> unREAL kelvar engine won't work.

> Will also try it on Bochs and see what I get.
> Has anybody tested it on their machines/emulators?

IIRC Kelvar worked for me, just no sound.

Also FASM (DOS, no IDE) used to work (better no DPMI host), but I can't guarantee for the latest version, as I use almost exclusively the IDE.

Of course I run DOS natively on real hardware, but I think both should work inside BOCHS too (untested).

This Tomasz Grysztar has no knowledge on hardware of electronics!. He made a good assembler, sadly he still many years after doesn't know the computer hardwares!. Obviously Tomasz Grysztar is a good copier best human machine! microsoft qualified. Microsoft approves such people!.

sid123! won't work on any 80% or less crappy dos emulators! in shitty world wide web have full of dumb emulators! because they and he followed microsoft and similar shits!. That's why that kelvar engine wont work!.
Work on pure 50%-100% 386 pc hardware emulated emulators! or you just use the standard ibm pc compatible 386 computers! should be work fine!.

I highly disagree.
Making something like FASM that can assemble itself plus can run on something
as useless as DOS is what a copier definitely can't do. I guess is good to have something like FASM since NASM has high requirements, FASM Compiles to 99KB, NASM compiles to 893KB, and for GNU AS, I guess everyone knows about how painful the syntax is, as for other assemblers, they are either proprietary or have some limitations.

This is self contradictory, to understand more about a specific CPU the best way to write an assembler that runs/assembles on an for that CPU.
@All
I got the kelvar engine working on the 386, Quite slow but works
@DOS386
IMHO If possible please lock this thread, I don't want this turn into a flame war, I guess everyone seeing this thread knows what I am talking about.

In my above post I forgot to include the link to my text on the history of the unREAL engine. It may give you some idea about the things that may go wrong. Not only the emulators, but even some physical CPUs do not support this trick. Still it worked on all the Intel and AMD processors I tried back then (only the more "exotic" x86 brands had problems).

As for the kelvar itself, it was based on my very old sources from the time when I was trying some game development (long before fasm). I later ported it to work on the unREAL engine and released it as an example. It contains a lot of suboptimal and perhaps even buggy code, though. And on 386 machines it was very clumsy. The one I have that runs it quite well is the first generation Pentium with 60 MHz clock.

This was already discussed two years ago in VirtualBox (VT-X under Nehalem Westmere 32nm)
. Yes, it works under VirtualBox. Of course, I only tested with advanced VT-X, so it's possible that your cpu still doesn't support it (sadly).