SoylentNews is people
SoylentNews
The recently released ASUS smartphone zenfone 2 has hit a new price/perf benchmark point with an MSRP of $199 but mid-high range specs:
-4G LTE
-Quadcore x86 processor
-5.5 inch IPS 1080P screen with gorilla glass
-2GB of RAM
-16GB storage
-3000 mAh battery
The low price is in part because Intel has been desperately trying to get a foothold in the mobile market and likely playing contra-revenue games. Unlike past low-cost options like the oneplus phone, this phone has wide release being sold at online retailers like Amazon.
Is this setting a new standard in low-cost, high-performance phones, or is this a temporary ploy until Intel starts charging for their SoCs? Will this lead to a price war between Mediatek, Qualcomm, and Intel? All of which have already released phones this year for the North American marketplace supporting the 4G spectrum. How low-priced can these smartphones with laptop-like specs go?
Reviewed here: http://anandtech.com/show/9251/the-asus-zenfone-2-review
[Editor's Comment: Original Submission]
(Score: 2) by Gravis on Thursday May 28 2015, @09:27PM
i got modded down for saying x86 eats too much power [soylentnews.org] but this is vindication. the battery capacity on this thing a monstrously high is 3000 mAh and yet it has poor battery life [anandtech.com] but the excuse of "because it has higher performance!" doesn't add up because poor performance scores [anandtech.com] too!
(Score: 1, Funny) by Anonymous Coward on Thursday May 28 2015, @09:29PM
But it cranks out a good amount of BTUs just like any good x86 chip.
(Score: 2) by MichaelDavidCrawford on Thursday May 28 2015, @09:38PM
I haven't actually tried but I would be unsurprised were my Xeon box to run DOS 2.0 just fine.
Yes 16-bit is used for BIOS but now we have UEFI.
According to "Pentium Processor Optimization", the pentium ran a lot faster if you treated it as a RISC chip - that is, if you programmed it in much the same way as you would a powerpc.
Suppose you patched llvm or gcc so that their x86 code generators did that. I expect they already do, but try removing all the support for 16-bit, string instructions, memory-to-memory operations and so on. Now you have a compiler for an "x86 RISC".
Then Intel or AMD could produce a CPU that would run that compiler's binaries. It would be smaller, cheaper, use less power, generate less heat. It would still boot but with EFI.
It's been obvious to me for well over a decade that this would be a good move; why doesn't anyone actually do it?
Yes I Have No Bananas. [gofundme.com]
(Score: 0) by Anonymous Coward on Thursday May 28 2015, @09:41PM
People run lots of legacy software.
(Score: 2) by MichaelDavidCrawford on Thursday May 28 2015, @10:35PM
I didn't think to make that clear - there are lots of people who don't run legacy software. Consider a typical webhost, it's all built from recent source.
Yes I Have No Bananas. [gofundme.com]
(Score: 1, Insightful) by Anonymous Coward on Thursday May 28 2015, @09:51PM
The answer is that those extra features are a lot less wasteful than you think. Sure the instruction decoder has to be a little bigger, but compared to the rest of the chip it's such a tiny savings that it's not worth the effort. Internally, Intel processors convert x86 to processor-specific microcode and the conversion can take a slow path for rarely used instructions.
(Score: 2) by MichaelDavidCrawford on Thursday May 28 2015, @10:37PM
if you support fewer instructions it's more tractable to implement them directly in circuitry, rather than in microcode. Were that the case the chip could be a lot faster.
Yes I Have No Bananas. [gofundme.com]
(Score: 0) by Anonymous Coward on Thursday May 28 2015, @10:23PM
There is a slight difference between a datacenter and a mobile phone.
(Score: 2) by FatPhil on Saturday May 30 2015, @07:52AM
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves