Apple's M1 Ultra delivers a stunning 114 billion transistors that make up 20 processing cores and 64 graphics cores.


Joined Mar 14, 2008
And to think how impressed I was when MOS Technology squeezed 4,528 transistors into the MOS6502!
Yes, I remember when a 1k bit RAM was a big development.

Amazing that the number of transistor on one chip has now gone to well over a billion.
Can it be long before they hit a trillion?

Of course with Apple, stunning performance usually comes with a stunning price. :rolleyes:


Joined Mar 19, 2019
With the several hundred billion USD they are talking about in new chip foundries in the next year it won't be long to hit the trillion mark. Even though some of that is being spent on larger trace sized legacy chip production to meet market demand.

panic mode

Joined Oct 10, 2011
at 4-5 thousands per chip this is not going to make much of a difference. but $300-$400 chips that sell in much larger volume will.


Joined Jan 17, 2007
When I was at Purdue, back in 1983, I held their Cyber 205 supercomputer in awe ... it was capable of handling up to 100 million arithmetic operations per second! (100 megaflops). What on earth were they doing with so much power?!!!

Nowadays, my slightly outdated desktop computer runs at 3.2 GHz and has 4 cores. I don't know how many operations per cycle each core can execute. But according to the "flop" equation, it has a basic capability of at least 3.2G x 4 x 1 = 12.8 Gigaflops! ... and that is without taking into account the graphic card installed in it, which admittedly has much more power than the computer's CPU itself!


Joined Sep 8, 2017
it has now reached nuron count in human's brain,
maybe ready for stochastic behaviours of a mind in a CPU. :E
If there is any "stochastic behaviour" at all in binary computer systems it is because of noise, not consciousness. And that is more a function of proximity rather than the particular number of transistors involved. (The clock signal otherwise prevents any sort of randomness in computation, which is precisely why results can be so readily duplicated across machines.)

That said analog computers might not be so easy to dismiss. The random fluctuations in their invariably distinct computations could perhaps be ascribed to a "mind" of sorts. It is all speculation of course...