Hello all,


i'm a ex-MSX software developer, after finding out about the 1 chip msx of course i ordered one (and let a good friend who is an ex-MSX developer order one too). these are just 2 cases of a real revival of cooled down love for the MSX!

(just wanted to say that for all the skeptics that claim there was not a revival going on)


one more thing, anyone know about a 1 chip emulator being developed? i would love to write vhdl code for the 1 chip hardware but i was wondering if there is actually some emulation possible where you can verify your code. i imagine that to be able to do this a bit convenient we at least need an vhdl emulator/interpreter and a VGA out to bitmap/window emulation. then a keyboard to ps2 in emulation would be needed too.

anyone know about such an emulator or anyone also is interested in one?

So, are you anyone we might know? (masked multician, remove your mask! ^_^)

i wrote quite a few demo's for the msx in the days, usually showing new techniques that i thought up. then i would sell them for 10 gilders on the MSX meeting days around the country.

after i found out about the 1 chip msx i looked into the MSX again and i actually gave one of the last demo's i wrote to the people of openmsx so they could fix their emulator and make the demo run on it that emulator is quite kool & sophisticated to support the features used in that demo. it used a not so well known feature of the clock chip to measure the exact speed of the Z80 and calculate the amount of t-cycles that fit in 1 display line. this info was then used to draw a full screen fluently moving 3D landscape (colored blocks) with the text "MAZE" hovering over it. the maze text had a shadow that was projected over the landscape. the demo runs on a msx-2 and was never really released.

www.htmlgamez.com/3D-Demo.zip (for all of you who got curious now)

ok, but no generally known nick/group, like those on the old Picturedisks from Genic and Sunrise etc.? (or other demodisks from that era)

i forgot what name i used back then but whatever name it was, i don't think many people know me from that time (might be wrong) some of the names back then do know me however, andre from ANMA was always the first person to look me up on every big meeting to see what i made and to show what he been doing. i also published some article in one of the msx magazines (MCM) and published a game in another one (MSX Info). then i was also part of the MSX Team and was responsible for the documentation library, the software we were supposed to make at the MSX Team with our group of developers was never completed though.

Welcome to the forum!

To answer your question. The full implementation is much too complex to run through a simulator with any real purpose. However, you can probably simulate subsystems if you can isolate them properly and feed them proper inputs. VHDL compilers and simulation software can be found in one of the other threads. If you run Windows you might as well get the Quartus II web editon from the Altera site as you will need this to rebuild the configuration anyway.

Good to have more old developers back on msx. If you have a look around, you'll see that there are still quite a few them.

are you sure that a 2.something gigahertz cpu can't emulate some VHDL code? i can imagine emulating 12.000 LEs is something else, but i mean just interpret or HLE the VHDL/source itself?

Hai Multi, and welcome to the forum. I checked your pseudo 3d field, it reminded me of this:

multi, thanks again for contacting us (openMSX devs) back then. And nice to see you here on the forum

Quote:

are you sure that a 2.something gigahertz cpu can't emulate some VHDL code? i can imagine emulating 12.000 LEs is something else, but i mean just interpret or HLE the VHDL/source itself?

That is actually not much different. VHDL is not like normal programming languages. Apart from process blocks, everything is evaluated in parallel. So for every change in input you'd have to run through the entire thing. With the VDP clock at over 20MHz, you have an input clock that changes 40 million times per second. So you can imagine what a staggering amount of computational time is needed to simulate just one realtime second.

The other problem is that a complete machine requires rather complex input. It requires memory with code to run which it fetches through a memory interface. While I haven't tried anything this complex yet, I doubt simulators were built to accommodate this.

Quote:

The other problem is that a complete machine requires rather complex input. It requires memory with code to run which it fetches through a memory interface. While I haven't tried anything this complex yet, I doubt simulators were built to accommodate this.

yeah, thats what i figured too and exacly the reason why i thought an 1-chip emulator would be kool, lol. well maybe the 1-chip itself would be the perfect platform to test code, i understood that the USB can be used to load code in the device. maybe it's possible to write some simple vhdl code to capture some 1-chip hardware signals (like input) and send it over usb to feed it to an simulator. but like you said the main issue would be an simulated memory controler (high speed & interactive in+out). anyway a simple oscilloscope program that would show pin 0/1 status and lets you flip a pin and see how things react would already be of great help i guess.

well, i am going to let all this info sink in and see by the time the 1-chip device arrives how it could help making vhdl development nicer for it (or not ofcourse if i don't get any brilliant ideas ).