🤯
Comment on Data centers will look ridiculous with tiny future servers.
Kolanaki@pawb.social 3 weeks ago
I sometimes wonder how powerful a computer could be made if we kept the current transistor size we have now, but still built the machine to take up an entire room. At what point would the number of transistors become more of a problem than a solution? 🤔
UniversalFlamingo@lemmy.world 3 weeks ago
Deepus@lemm.ee 3 weeks ago
Isnt the main limiting factor signal integrity? Like we could do a CPU the size of a room now but it’s pointless as the stuff at one end wouldnt be able to even talk to the stuff in the middle as the signal just get fucked up on the way?
LH0ezVT@sh.itjust.works 3 weeks ago
Signal integrity will probably be fine, you can always go with optical signalling for the long routes. What would be more of an issue is absurd complexity, latency from one end to the other, that kind of stuff. At some point, just breaking it down into a lot of semi-autonomous nodes in a cluster makes more sense. We kind of already started this with multi-core CPUs (and GPUs are essentially a lot of pretty dumb cores). The currently biggest CPUs all have a lot of cores, for a reason.
Jolteon@lemmy.zip 3 weeks ago
IRC, light speed delay (or technically, electricity speed delay) it’s also a factor, but I can’t remember how much of a factor.
BullishUtensil@lemmy.world 3 weeks ago
It’s significant already. If I get the math right (warning, I’m on my phone in bed at 3am and it’s been 10 years) I think that a 1 inch chip running at 3GHz clock rate could, if you aren’t careful with the design of the clock network, end up with half a clock cycle physically fitting on the chip. That is, the trace that was supposed to move the signal from one end of the chip to the other, would instead see the clock signal as a standing wave, not moving at all. (Of course people has (tried?) to make use of that effect. I think it was called “resonant clock distribution” or some such)