Recent content by commodorejohn

Man I have just been banging my head against this kind of thing with my own esoteric Tandy project. 8086 instruction encoding and 8088 prefetch queue voodoo is its own particular brand of Deep Magic. (Wrt. arithmetic between accumulator registers, I get the impression that there's only certain...

I'd forgotten that was even a thing. I really should try it on the RL and see what happens.

Had a chance to work out some initial thoughts on algorithms: Likely the biggest bottleneck: if (Xsource ^ Xdest) = 1, the pixels aren't aligned within the bytes, and will have to be shifted and masked/combined byte-by-byte. When that's not the case, we can mostly blit in 16-bit chunks instead...

♫ Same as it ever was... ♪

The Model 100 might have it beat on the "rock-solid build quality" front, but the 1000 certainly has the upper hand in terms of system architecture.

Thankee, that is quality information =) Hadn't expected that they would've bothered to add that much bus-juggling logic to what was still a machine primarily aimed at the home market; really drives home what an absolute corner-cutting botch-job the PCjr was =/

Yeah, I figured as much on that point - I only have the 1000RL to test against in actual hardware, anyway, but I'm just trying to get my head around it all. (Speaking of the RL/SL/T* series, I'm idly curious if the video memory is 16-bit? I know they revamped the video chip and swapped out the...

Many moons ago, when I was first dabbling with 8086 assembly, one of the miscellaneous projects I undertook was a little demo moving a masked sprite across a bitmapped playfield in Tandy 1000 320x200x16 mode. Being very inexperienced with the 8086 at the time, I didn't expect spectacular results...

It's less you specifically and more that it's just gotten furshlugginer exhausting how inescapable the "I asked ELIZA about XYZ and here's what she babbled, can anyone tell me if this is anything approximating boring fleshbrain notions of 'correct' plzkthx?" threads choking up tech forums have...

LLMs are a glorified Markov-chain "travesty generator." They have no understanding of anything - all they do is string words together according to probability - and should never, ever be relied on for factual information, especially where important questions in matters like finance or law are...

Purely spitballing here, but if it's a stack-based architecture coming out of HP, have you compared notes against the instruction set for the original HP 3000 architecture?

Hah, that's wild :D

Welcome aboard!

I mean, I want that to be true.

I've been waiting for ReactOS to be ready for primetime since about 2008...unfortunately, I'm still waiting :/