cpmulator VS solvespace

A while back I wrote a game in assembly, for CP/M. Since I have a single-board Z80-based computer on which I can run it.

I later ported the game to the ZX Spectrum, because that was a fun challenge, and I only needed a few basic I/O operations - "write to screen", "read a line of input", etc, etc.

It occurred to me that I could reimplement the very few CP/M BIOS functions and combine those implementatiosn with a Z80 emulator to run it "natively". So I did that, then I wondered what it would take to run Zork and other games.

Slowly I've been reimplementing the necessary CP/M BDOS functions so that I can run more and more applications. I'm not going to go crazy, anything with sectors/disks is out of scope, but adding the file-based I/O functions takes me pretty far.

At the moment I've got an annoying bug where the Aztec C-compiler doesn't quite work under my emulator and I'm trying to track it down. The C-compiler produces an assembly file which is 100% identical to that produced on my real hardware, but for some reason the assembler output from compiling that file is broken - I suspect I've got something wrong with my file-based I/O, but I've not yet resolved the problem.

TLDR; writing a CP/M emulator in golang, and getting more and more software running on it - https://github.com/skx/cpmulator

Can second this!

However, I would recommend https://solvespace.com! It hits a sweet spot between features vs complexity/learning effort.

We use rebase on solvespace, along with sensible squashing so most commits along master are pretty self contained. You can see the clean history here:

https://github.com/solvespace/solvespace/commits/master/

I changed a behavior to the "more standard" one because it felt obviously right. This was a 3 line change. But the was enough backlash right there in the pull request. So I spent a couple hours remembering how to add a configuration option to keep the old way for those guys:

https://github.com/solvespace/solvespace/pull/1425

> If you like Linkage, you might also like Solvespace.

No, I mean Brent Curry's Linkage[1] bicycle design software, not David Rector's Linkage Mechanism Designer and Simulator[2].

You should read Wikipedia article.[0]

N.B. About SolveSpace, as I'm its experienced user[youtube,patreon], I may say next: yes, it could be used for bike mockup, as any other CAD, but it still has a lot of limitations and even does not export correct STEP files yet[3], and in FreeCAD such STEP could fixed only partially.[video]

So, for serious 3D CAD work I highly recommend use FreeCAD (and LibreCAD for 2D CAD work) instead of SolveSpace, and use SolveSpace only as a helper tool like a calc or as a notepad for noting ideas.

About Linkage Mechanism Designer and Simulator, it is only useful for planar (2D) kinematics analyze, and if You are looking an alternative for it take a look on Pyslvs[4], that is in part based on SolveSpace's solver.

[0] https://en.wikipedia.org/wiki/rattleCAD#History

[1] https://bikechecker.com/

[2] https://blog.rectorsquid.com/linkage-mechanism-designer-and-...

[3] https://github.com/solvespace/solvespace/issues/206

[4] https://github.com/KmolYuan/Pyslvs-UI

[video] https://www.youtube.com/watch?v=F3LJMeqUDrU

[youtube] https://www.youtube.com/@appsoft

[patreon] https://patreon.com/app4soft

C++ this file covers all the math for working with NURBS curves and surfaces:

https://github.com/solvespace/solvespace/blob/master/src/srf...

There is a lot more in other files - triangulation, booleans, creation - but the core math functions are there in very readable form.

>> Maybe somebody has statistical survey of how much of the existing deployed CPU core count is typically used?

My guess is very few cores are used on average. I did some testing with Solvespace to see which build options contributed most to performance:

https://github.com/solvespace/solvespace/issues/972

Obviously using OpenMP for multi-core was the big win. But what's not shown is that in typical usage (not the test I ran) if you're dragging some geometry around it will use all cores (in my case 4 cores / 8 threads) at about 50 percent utilization. That percentage probably drops as more cores are thrown at it due to Amdahl's Law. In other words, throwing double the cores at it will give a good boost to a lot of code that is already taking less than half the time (wall clock time, not CPU time).

We added OpenMP to a number of functions for significant performance gains. And in fact, any remining single-thread operation that gets the parallel treatment is likely to have a significant impact on overall performance since that is where most of the time is spent now. At this point we're more focused on features and bugs.

Algorithmic improvements are possible and I'd like to do those in the future, but they are much harder to do than sprinkling some #pragmas around critical loops. That will improve the scalability though, where multithreading really did not.