I’ve been chewing on some retro meat this week around what I think is a real gap in hardware and emulated hardware preservation for gaming of a certain era.
This has come up repeatedly over the past few months while I’ve been researching games for The Play Retro Show.
You may hear me talk often on the show about game preservation. But when I say game preservation, I’m talking about something bigger than just keeping the software alive. I’m passionate about preserving the experience of those games, not just the bits on a disk, but how they felt to play, and making that experience as accessible as possible to as many people as possible.
When it comes to software, preservation is actually in pretty good shape overall. Most classic games are dumped, archived, and widely shared. People are doing fantastic work on that front.
But hardware is a lot tougher. There are incredible hardware emulation projects and dedicated communities working hard to keep vintage hardware alive and patch old games to run on modern systems.
Here’s the problem I keep running into:
A lot of PC games from the late 90s and early 2000s require so much hoop-jumping to get working properly. We have great tools like PCem and 86Box, which do an impressive job emulating machines from that era up to about Pentium II processors around 233 MHz. After that point, things start to get dicey.
Like many folks emulating with these projects, I’ve often wondered why the limit exists. Modern processors are absurdly powerful. Shouldn’t they be able to handle faster emulation? Can I just upgrade my PC and brute-force past it?
So I’ve been digging into documentation and discussions to get at least a basic understanding of why. Feel free to correct me if I’m wrong here:
-
The limitation is mostly practical.
-
Modern CPUs have lots of cores, which are split into logical processors.
-
Projects like 86Box only use a single logical processor.
-
Because of that, you are limited by the performance of a single core or thread rather than the total horsepower of your CPU.
-
Over the last decade, single-core speeds have not dramatically increased, even as CPUs have gained more cores.
Unless you are buying a very high-end CPU, you still hit that same ceiling.
So what happens to all the games that fall into this weird in-between era? They are too new for easy retro emulation, too old for modern Windows, and do not have passionate communities building compatibility patches.
Do I try to build an old Windows PC to cover most of that era? Multiple machines? Should I spend my time and resources going down that road? Or should I bite the bullet and buy a $1,500 i9 and hope that helps?
Are there projects I am not aware of that use a different approach, like leveraging the GPU for emulation or splitting work across multiple cores?
These are not questions I expect anyone to definitively answer, but I would love to hear your thoughts. Feedback is appreciated, and discussion is very much encouraged.
