11 replies to this topic

[Arceles]

http://www.notebookc...he.84601.0.html

MT6599, LTE... just for those that always complain for more cores in a smart phone

[Astra.Xtreme]

I wonder how many seconds the battery lasts.

[+Azusa]

Astra.Xtreme, on 27 November 2012 - 00:21, said:

I wonder how many seconds the battery lasts.

8 seconds 1 for each core.

[tiagosilva29]

Nice!

[(Spork)]

Astra.Xtreme, on 27 November 2012 - 00:21, said:

I wonder how many seconds the battery lasts.

well that didnt take long lol

[+Xinok]

I still only have a dual-core processor in my PC.

[Draconian Guppy]

You have to wonder, the real world benefits.. heck only windows 8 can fully manage AMD's octocore, how are they gonna go about this on smartphones?
This is why I like windows Phone... No need for crazy ass specs !

[yowanvista]

MediaTek's SoC are usually behind the competition and the introduction of that 8-core thing sounds a bit too overkill. Why on earth would a smartphone possibly need this?

Seems to be another hyped gimmick SoC that will bring nothing new to the table unless they use the Cortex-A15 design.

Recent versions of Android will play nice with single core processors let alone dual cores. They're going crazy with this 'who has more cores' race.

[+Phouchg]

Fooling general public with numbers game. While it's not a bad idea as such (because we are ever so close to the point where clocks can't be raised much), efficient parallel processing has failed to materialise, because it's friggin' hard to even do at times, not to mention do correctly. And many of today's mainstream programmers cannot be bothered to do hard things - they've lived on the legacy of the old, hiding behind tons of abstractions, so all new systems totally suck.

[S7R1K3R]

Phouchg, on 27 November 2012 - 07:30, said:

Fooling general public with numbers game. While it's not a bad idea as such (because we are ever so close to the point where clocks can't be raised much), efficient parallel processing has failed to materialise, because it's friggin' hard to even do at times, not to mention do correctly. And many of today's mainstream programmers cannot be bothered to do hard things - they've lived on the legacy of the old, hiding behind tons of abstractions, so all new systems totally suck.

What comes first, the hardware or the software to run on said hardware?

[+SharpGreen]

S7R1K3R, on 28 November 2012 - 20:52, said:

What comes first, the hardware or the software to run on said hardware?

The hardware always comes first.

[+Phouchg]

S7R1K3R, on 28 November 2012 - 20:52, said:

What comes first, the hardware or the software to run on said hardware?

SharpGreen, on 28 November 2012 - 20:56, said:

The hardware always comes first.

It depends. In short - software can be designed to be scalable.

Long story - in most cases it even is - rarely a program (or an app... dog, I hate this term) is single-threaded these days. Just that managing multiple threads so that they don't race each other, spinlock, starve for data or deadlock - that's a damn pain in the ass for most programmers these days. And with a good reason, because it actually is very hard and complicated thing, requiring to draw flowcharts and doing hand optimizations for pretty much anything. OS schedulers and CPUs are very slowly getting better at being relied on for that, however often neither has a clue of the importance of the task at hand, not as it's being perceived by the user - they just have a priority list of hundreds of threads that each cry for more cycles. That way we are left with throwing bigger and bigger iron at the problem. I genuinely hope that there won't be a technological breakthrough for reducing process sizes so that it this unwillingness to do hard work will become problematic, even cease.