Sunday, August 03, 2014
Rock's Law now rocks: Why Moore's law effectively failed at Intel's 14nm ?
___________________________
99% BAD HARDWARE WEEK: We should have by now Intel's 4+B processors. But even possible, those transistors should be dark due to physically limited power consumption. 22 nm node was a tipping point.
After that, chip manufacturing costs were rising faster than the number of transistors on chip.
Though Barnum circus style marketed around for a while, Intel's 14 nm process is actually only process improvement, not a real shrunk node. Process benefits on chip's high performance are neglectible. However, in low power consumption version effects are somewhat better. 30 years after 386 processor introduction in 1,5 micron and 100 times smaller node at 14nm, seeems that shrinking magic definitely failed. ALAS, NO MORE FREE BIG CHIP LUNCH FOR CONSUMERS.
99% BAD HARDWARE WEEK: We should have by now Intel's 4+B processors. But even possible, those transistors should be dark due to physically limited power consumption. 22 nm node was a tipping point.
After that, chip manufacturing costs were rising faster than the number of transistors on chip.
Though Barnum circus style marketed around for a while, Intel's 14 nm process is actually only process improvement, not a real shrunk node. Process benefits on chip's high performance are neglectible. However, in low power consumption version effects are somewhat better. 30 years after 386 processor introduction in 1,5 micron and 100 times smaller node at 14nm, seeems that shrinking magic definitely failed. ALAS, NO MORE FREE BIG CHIP LUNCH FOR CONSUMERS.