Comment Laws of Physics (Score 1) 1009
This is more a case of performance I expect, the higher I/O speeds of the newer memory and graphics ports and PCI-e are such that the costs to maintain signal integrity through a socket adaptor are horrid. If your really performance minded, you'd shotgun the socket and solder existing PC's to the motherboard today, over clockers would likely benefit from the lower noise due to removing the inductance of the socket leads on the power and ground pins. This would yield higher reliability at higher clock frequencies. The concerns over socket signal integrity have been an issue testing these high end CPU's since the I/O bandwidths approached 100Mhz.
I agree with Jump! Jump! Jump! To quite a dramatic conclusion.
Also... in the IC testing community, regardless of the physical structure (flip chip/BGA/PGA), we still refer to the electrical contacts of a device under test aka (DUT) as "pins". So these devices are likely to have "pins" forever, even when we start interfacing to these chips with photons...
I agree with Jump! Jump! Jump! To quite a dramatic conclusion.
Also... in the IC testing community, regardless of the physical structure (flip chip/BGA/PGA), we still refer to the electrical contacts of a device under test aka (DUT) as "pins". So these devices are likely to have "pins" forever, even when we start interfacing to these chips with photons...