Same for the 8514/A - which, although it was released a couple of years earlier, formed the basis for most "SuperVGA"-style clones.
In the first half of the 90s, pretty much every ISA graphics card had a 1024x768 @ 43i mode, and fancier ones could often do 1280x1024 at the same refresh rate (often accompanied by an ear-splitting whine coming from the monitor!)
This situation lasted until programmable PLLs that could do 60 or 72p began to take over in the mid 90s, with 90p following swiftly behind. But by then the idea that anything above 800x600 was a recipe for eye-strain had been baked into the lore, and corporate IT departments in particular refused to risk going beyond it until well into the LCD era.
For that matter: how does "self-driving only" even work for a personal car, from a practical perspective - e.g. how do you get it to park in a particular spot? Can you take it through a drive-thru, or a toll booth? How about a car wash? Will it respond appropriately to a police car trying to pull it over or guide it around an accident? Etc, etc.
All of these are problems which can be ignored or worked around for a taxi, or for a car with an optional self-driving mode - but if there's really no way to control the car manually, the car needs to be able to handle everything itself. And I'm not sure it can.
> That might be a hardware limit of the current architecture.
It is, at the very least, a practical limit. There are unlikely to be enough customers willing to pay $50k+ for a hypothetical 1TB RAM version to make it worth manufacturing.
The pipeline from new ARM cores to shipping silicon is surprisingly long. Cortex-M85 was announced in 2022, and there's still almost nothing on the market that implements it.
I don't think it's merely long pipeline. Simpler explanation is that there is just not much demand for these middle-class cores; A55 cores are from same generation and started appearing in socs pretty quickly. I suspect that the benefits of A3x (compared to A5x) just werent there for vendors to justify adding them to their portfolios
> This is like desktop computers from 30-40 years ago I guess.
30 years ago was 1990; a typical desktop computer of that era would have a single CPU at 20ish MHz, and between 512 kB and 2 MB of RAM. The Ox64 easily outclasses anything that was available at the time.
Your overall point is valid, though I think you'll find that this particular calculation isn't correct. (The current year ends in a 6 so 30 years ago also ends in a 6.)
However, I just did a quick search for the Doom release date, which was 1993 and the original system requirements which is Memory (RAM): 4 MB and Processor (CPU): Intel 486DX (66 MHz).[1] So in terms of RAM this is 16x as much and in terms of MHz it's 6x faster on one of the cores. So it's definitely desktop-class for 1993. (33 years ago). As you state, it's much faster than desktop class in 1990.
66MHz probably wasn't an 80486DX; as far as I can recall, those only went up to 50MHz. It would have to be a 486DX2, with the RAM running at only 33MHz. https://netlib.org/performance/html/dhrystone.data.col0.html says that a 486DX2/66 was about 31 to 51 Dhrystone MIPS.
Whoops! Either way - 480 MHz and 64 MB of RAM would put this more on par with a 2001 iMac (400-600 MHz, 64 MB RAM on the base model). And that's not even counting the other CPU cores or the (probably much better) IPC of a modern CPU design.
> So it's definitely desktop-class for 1993. (33 years ago).
I have no idea how you reached this conclusion, given that the numbers you have provided are in total contradiction.
A ~400+ MHz processor today is definitely NOT desktop-class for 1993 when the average was a 486DX. I mean, I was there. You are a decade off, if you want to include all the architectural improvements that even a cheap CPU today has.
reply