5 ms·

Isn't this just a trade off? Is there never a scenario where you would trade transistor density for switching speed and lower power consumption?

by pg_bot 5y ago

dumbfoundded 5y ago
Yes. IIRC, amd chips have been beating intel chips for a while now on transistor sizes but intel even with larger transistors still have a greater density on a chip (maybe it's changed in the latest gen). Another benefit of lower density is cooling.
dogber1 5y ago
That's indeed done all the time in electronics: for example, RF CMOS usually trailing on a node three or four generations behind the bleeding edge. However, all-optical/photonic computing is just intrinsically so much worse than electronics. On top of the issues that I touched on, there are also other fundamental problems, e.g. distribution of power: photons like to get absorbed by nearby electrons. How do you then supply all the active devices (switches/lasers/etc.) with power while maintaining some semblance of signal integrity and dense integration?
- singularity2001 5y ago
  Could nonlinear wave interactions be applied in near vacuum, isolated from the lasers, amplifiers and counters? Think 100000*100000 imprecise loss-full tensor/matrix multiplications.
- jacquesm 5y ago
  There is a special case: pumped laser amplification of signal in underwater fiber optic cables. That's all optical for the signal path as far as I know. https://www.laserfocusworld.com/fiber-optics/article/16551090/optical-amplifiers-speed-data-flow-undersea https://www.laserfocusworld.com/fiber-optics/article/1655109...
singularity2001 5y ago
Exactly. While this speed vs space trade off makes less sense in mobiles, it might make perfect sense in industrial settings. Imagine 3D computers the size of a room (Craigh 2) but a 1000 times faster than any TPU only cluster.
- ajuc 5y ago
  If you can trade single-core speed for parallelism cluster of traditional electronics make more sense, but for some algorithms you can't. Imagine a CPU with the complexity of Arduino but running at 100 GHz.

Isn't this just a trade off? Is there never a scenario where you would trade transistor density for switching speed and lower power consumption?

by pg_bot 5y ago

Isn't this just a trade off? Is there never a scenario where you would trade transistor density for switching speed and lower power consumption?

dumbfoundded 5y ago

Yes. IIRC, amd chips have been beating intel chips for a while now on transistor sizes but intel even with larger transistors still have a greater density on a chip (maybe it's changed in the latest gen). Another benefit of lower density is cooling.

dogber1 5y ago

That's indeed done all the time in electronics: for example, RF CMOS usually trailing on a node three or four generations behind the bleeding edge. However, all-optical/photonic computing is just intrinsically so much worse than electronics. On top of the issues that I touched on, there are also other fundamental problems, e.g. distribution of power: photons like to get absorbed by nearby electrons. How do you then supply all the active devices (switches/lasers/etc.) with power while maintaining some semblance of signal integrity and dense integration?

singularity2001 5y ago
Could nonlinear wave interactions be applied in near vacuum, isolated from the lasers, amplifiers and counters? Think 100000*100000 imprecise loss-full tensor/matrix multiplications.
jacquesm 5y ago
There is a special case: pumped laser amplification of signal in underwater fiber optic cables. That's all optical for the signal path as far as I know. https://www.laserfocusworld.com/fiber-optics/article/16551090/optical-amplifiers-speed-data-flow-undersea https://www.laserfocusworld.com/fiber-optics/article/1655109...

singularity2001 5y ago

Exactly. While this speed vs space trade off makes less sense in mobiles, it might make perfect sense in industrial settings. Imagine 3D computers the size of a room (Craigh 2) but a 1000 times faster than any TPU only cluster.

ajuc 5y ago
If you can trade single-core speed for parallelism cluster of traditional electronics make more sense, but for some algorithms you can't. Imagine a CPU with the complexity of Arduino but running at 100 GHz.