View Single Post
Old December 16th, 2012, 11:20 PM   #10 (permalink)
EarlyMon
The PearlyMon
 
EarlyMon's Avatar
 
Join Date: Jun 2010
Location: New Mexico, USA
Posts: 46,323
 
Device(s): M8, LTEvo, 3vo, and Shift - Evo retired
Carrier: Sprint

Thanks: 42,706
Thanked 57,146 Times in 22,962 Posts
Default

Because you only split things out so far for parallelism, and that includes loaded tasks against actual ram.

If you want an active 8-core where everything gets used, you need a lot more ram than today, or apps written to exploit more parallelism. Probably both.

And unlike cores, sections of ram can't go to sleep, so that's going to be a power draw just for real estate.

BIGlittle has already been out and proven itself to a certain point, this is its next evolution. It's designed as an either-or deal to go fast or go efficient, on the fly.

The Tegra 3 quad core does this with a single little core (it's the quad with 5 cpus). Think of this like that, only the BIG gets a big boost going from the A9 to A15 architecture and the little gets a boost from one to quad to ease the transition from a single little core to multiple BIG cores.

Is this the best way? Is this a good way? Will it come through on promises? Will others just go on the big V-8 kinda engine? How much power in real life? How hot?

Your guess is as good as mine, I think we'll have to just see them.

But on paper, does it make sense? Yes, I think so.
__________________
|

Minutus cantorum, minutus balorum, minutus carborata descendum pantorum.

Links: Site Rules / Guidelines -and- Zero Tolerance Policy (All Members Read)


For right-on help, the Thanks button is on the right of the post.
For anything out in left field, the /!\ report button is to the left.

Remember, it's our forums and we're all in this together - so let's keep it cool!

Shoot the breeze at the best new gun forum!
EarlyMon is offline  
Reply With Quote