Hardware manufacturers should have spent the last 20 years focusing on improving power consumption and battery life on consumer electronics and less time giving programmers 4.6 billion yards of rope to hang themselves with
What do modern PCs give you that a z80 CPU couldn't
Is it the bigger display? Do you think we could have done that without making our computers hundreds of times more powerful?
My TI-84+ SE runs a full blown Unix operating system, can boot to a useful graphical interface in less than 10ms, and I change the battery once per YEAR
@sir I have a 4k screen with beautifully rendered fonts and graphics, a lot of which is rendered (at least partially) by the CPU. My CPU can decode beautiful high resolution video at high frame rates. My CPU handles complex physics simulations in games. My CPU lets me edit and encode high resolution video reasonably well.
You can complain about unnecessary software bloat, and it's a real problem. However, even the best software engineers couldn't give me the high fidelity I'm used to on a z80.
@mort we could do all of this on a z80. Well, maybe not 4K, but I did *explicitly address that*.
@sir If you mean your "is it the bigger displays?" comment, then I suppose the answer is just that no, we need many orders of magnitude better CPUs for that.
A raw 4K image is 24.9MB. I want 60 of those per seconds, which means processing 1.5GB of image data per second. _Just iterating through the data with 1 clock cycle per byte_ would require 1.5GHz CPUs. Add more complex operations, and you quickly start to see the need for many multi-GHz cores.
@mort or, you know, dedicated GPUs.
@sir You could go that route, yeah. Don't treat the computer as a general computation machine, instead add dedicated hardware for everything that's too much for your z80 to handle.
Or you could just make them faster. The current state, where we have dedicated hardware for common tasks but CPUs that are fast enough for all tasks other than 3D graphics, seems like a nice balance. Lets software utilize dedicated hardware if possible, and fall back to the fast-enough CPU if necessary.
@mort I never suggested adding new hardware for everything the z80 can't handle. There is a strongly compelling use-case for a highly parallel processor, i.e. a GPU. Much less so for everything else.
@sir I mean, you would lose the ability to decode or encode video on the CPU should the GPU be missing or incompatible. Your z80+GPU would also just be unable to cover highly sequential needs like physics simulation, where one physics tick necessarily depends on the previous.
Fast CPUs are just really useful all around even when we ignore Electron or GTK or whatever you consider bad.
@mort I have personall written real-time physics simulations for a 15 MHz z80 CPU.
@mort personally*
@sir I'm sure you have, but they presumably weren't super complex. There's no upper limit to how fast CPUs we need for physics simulations; faster CPUs always mean we can model more complex systems more accurately.
Accurate, complex physics is nice for games, but it's super fucking important for things like climate science and metrology.
@mort specialized problem -> specialized computer
You just can't yield on anything, can you
@sir software is the superior R&D environment. This is why general purpose computers took off, programmers would do things nobody anticipated would be done, creating markets. Software sold hardware, software shaped hardware. It is a superior evolution path for IT. The problem is that it was proprietary software, which led to unhealthy rivalry and degeneracy. That is the only reason that you are now in a position to argue against general purpose computers and their inefficiency by example.
