"ARM's A15 (28nm Process) supposidly delivers ATOM like performance at moderate power consumption (1.9W)"
so an A15 at 28nm actually hasn't actually been measured for power consumption...
Wow, one opinion post stirs the ARM camp so heavily. OK Samsung Galaxy S4 said to have high Antutu score, Intel Lenovo K900 (on 32nm process) may be comparable. Don't be surprised if Intel are trialing some newer tweaks to the architecture prior to moving Atom to 22nm process with completely new architecture.
My suggestion is that ATOM is already comparable with A15, though Atom is a process node behind and an aged atchitecture.
Ok, seems a little strange that ARM fans are so irate about my post. Ynot go to ARM board and proclaim how fast and superior 96 cores out perform 16 core intel 32nm Xeon (3 year old) cores in server application. Why not compare to intels latest Xeon phi :829 GF/s @ 195W, possibly a 300 ARM core server may achieve such results at 1000W.
So much previously said about intels power inefficient and ARM efficient designs. Suggestion is that there really is no efficiency gained by architecture. Manufacturing process, software and internal design features determine the efficiency of a processor.
So little mentioned now a day about the supposid power efficiency advantage ARM has/had over intel?
"Wow, one opinion post stirs the ARM camp so heavily. "
I see... so a fact has turned into an opinion when proven false...
"Don't be surprised if Intel are trialing some newer tweaks to the architecture prior to moving Atom to 22nm process with completely new architecture."
Why would I be surprised... instead it's about time !!!
"My suggestion is that ATOM is already comparable with A15, though Atom is a process node behind and an aged atchitecture."
If you understood that ARM, with its load/store and wimpy register set, is being compared to an ATOM
And Tegra 4 benchmarks were 3x that of the Clovertrail...
And when CloverTrail + is available... what die size will it be...
Bet you it's bigger than Tegra 4 at 80 sq mm...
do u think much diff going from 32nm to 28nm? A15 is power hog.
Any info anywhere on next atom going from 32nm planar to 22nm Tri-gate? If current atom on 32nm on planar already competitive with 28nm A15, where do you think next Atom on 22nm Tri-gate will stand? Fool.
The Axxia 5500 product family will include a range of communication processors with different core counts and throughput capabilities and are the first in the industry to deliver 16 ARM Cortex™-A15 cores using ARM’s CoreLink™ CCN-504 low-latency interconnect in 28nm process technology. The combination of LSI’s networking expertise, specialized acceleration engines and Virtual Pipeline™ technology with ARM’s power-efficient processors and interconnect IP delivers communication processors that are uniquely suited for building intelligent, heterogeneous networks.
Sentiment: Strong Sell
Anandtech did in depth power analysis of Atom, Krait and A15...shows A15 is a power hog:
"The ARM vs x86 Wars Have Begun: In-Depth Power Analysis of Atom, Krait & Cortex A15"