4 ms·

I had an Amiga 1000 and loved it. That said, a lot of the things we thought were so epic about it were just smoke and mirrors. The iconic bouncing ball demo,

by 13of40 9y ago

I had an Amiga 1000 and loved it. That said, a lot of the things we thought were so epic about it were just smoke and mirrors. The iconic bouncing ball demo, for example, was made to spin by switching the color palette and made to move by changing the address of the video buffer in memory. The same elegant system couldn't have extended to something as simple as a second ball on a different trajectory or rotating at a different speed. A modern computer, on the other hand, could render 1000 of those balls and grab coffee between frames. It's hard to imagine what you could throw together on some breadboards in your garage that would be streets ahead of the mid-range PC you're lamenting.

deaddodo 9y ago
But that's what you have to do on limited hardware. Raycasting is awesome, but it's a hacked, inverse raytracing limited to a single ray plane. What made the bouncing ball awesome was the fact that it could be done thanks to the planar video setup and powerful blitter compared to the chunked video system IBM-compatible's were using up through SVGA.
- bitwize 9y ago
  Most of the PC's graphical memory layouts in the 80s and early 90s were planar; the major exceptions were CGA and 320x200 MCGA 256-color. That had a lot to do with why PC gaming was so bad; without a blitter, graphics had to be drawn on-screen by shifting the bit patterns "by hand" and drawing them on the appropriate bitplanes. Many games got around this by only drawing sprites or scenery elements on 8-pixel boundaries. John Carmack figured out a way to use the EGA's scrolling registers and off-screen video memory to create smoothly scrolling backgrounds; for sprites he stored four copies of each sprite, each shifted to a different bit position so they could easily be drawn on screen. Chunky graphics is what saved the PC, gaming-wise, and made Doom and Quake possible. Drawing an image became a matter of writing the color-table index for each pixel to that pixel's memory location. Screens of considerable complexity could be drawn with the CPU alone.
danielbarla 9y ago
On the other hand, I distinctly remember having Imagine or Lightwave doing actual ray tracing in the background (clearly taking up 100% CPU), while I was playing some simple game in a window in the foreground. Having set the foreground window to high priority somehow seems to have worked back then, but for whatever reason doesn't seem to work on today's typical PCs. Random things seem to make ridiculously powerful machines pause for whatever reason, even without having processes take up 100% of the CPU.
- AstralStorm 9y ago
  This is the difference between a real time operating system and one that is not. See BeOS for a newer take. Or MorphOS. (X86 has a few things that can break realtime performance, such as broken firmware handling of SMI.)
  - wolfgke 9y ago
    What you told is probably an important reason. But another one is that modern x86 cores clock down if there is load on multiple cores (to prevent heating issues) or even clock up for a short time (e.g. Intel Turbo Boost) if only a single core gets a load peak without much load on other cores. In practice also often insufficient cooling is a reason why the computer gets "slower/sluggish" if there is a high load in particular on multiple cores. The cooling of practically every laptop available is in my opinion undersized (to save space) and if dust gets into the fan (it will gradually) this will get even worse. Luckily most gamers ("PC Master Race") know how important a sufficiently dimensioned cooling is for their self-built PCs.

by 13of40 9y ago

deaddodo 9y ago

But that's what you have to do on limited hardware. Raycasting is awesome, but it's a hacked, inverse raytracing limited to a single ray plane. What made the bouncing ball awesome was the fact that it could be done thanks to the planar video setup and powerful blitter compared to the chunked video system IBM-compatible's were using up through SVGA.

bitwize 9y ago
Most of the PC's graphical memory layouts in the 80s and early 90s were planar; the major exceptions were CGA and 320x200 MCGA 256-color. That had a lot to do with why PC gaming was so bad; without a blitter, graphics had to be drawn on-screen by shifting the bit patterns "by hand" and drawing them on the appropriate bitplanes. Many games got around this by only drawing sprites or scenery elements on 8-pixel boundaries. John Carmack figured out a way to use the EGA's scrolling registers and off-screen video memory to create smoothly scrolling backgrounds; for sprites he stored four copies of each sprite, each shifted to a different bit position so they could easily be drawn on screen. Chunky graphics is what saved the PC, gaming-wise, and made Doom and Quake possible. Drawing an image became a matter of writing the color-table index for each pixel to that pixel's memory location. Screens of considerable complexity could be drawn with the CPU alone.

danielbarla 9y ago

On the other hand, I distinctly remember having Imagine or Lightwave doing actual ray tracing in the background (clearly taking up 100% CPU), while I was playing some simple game in a window in the foreground. Having set the foreground window to high priority somehow seems to have worked back then, but for whatever reason doesn't seem to work on today's typical PCs. Random things seem to make ridiculously powerful machines pause for whatever reason, even without having processes take up 100% of the CPU.

AstralStorm 9y ago
This is the difference between a real time operating system and one that is not. See BeOS for a newer take. Or MorphOS. (X86 has a few things that can break realtime performance, such as broken firmware handling of SMI.)
- wolfgke 9y ago
  What you told is probably an important reason. But another one is that modern x86 cores clock down if there is load on multiple cores (to prevent heating issues) or even clock up for a short time (e.g. Intel Turbo Boost) if only a single core gets a load peak without much load on other cores. In practice also often insufficient cooling is a reason why the computer gets "slower/sluggish" if there is a high load in particular on multiple cores. The cooling of practically every laptop available is in my opinion undersized (to save space) and if dust gets into the fan (it will gradually) this will get even worse. Luckily most gamers ("PC Master Race") know how important a sufficiently dimensioned cooling is for their self-built PCs.