- Apr 21, 2015
Can someone technically explain hwow 14nm worse than 16nm? For CPU in desktop or laptop market, smaller nm always means much better performance and smaller heat, how can Samsung makes history?
tsmc chip price is that much higher!?The TSMC always has better chip quality with smaller current leakage
14/16 name here is confusing since their transistor density is the same as 20 nm, but they switch to finfit process which has 10 times lower current leakage
why is the chip price of TSMC 70% higher than Samsung? It didn't come without no reason....
Actually I read more and find out very likely it is Apple using IOS to tune down the samsung cpu power, so the two chip has similar performance instead of samsung chip has better performance, but somehow samsung cpu cannot keep the performance/power ratio as well as 16nm. Very likely there will be way to unlocked the samsung cpu power or apple itself using ios to fix this problem by just let samsung chip runs fasterMaybe find something to do or like take her out or something.
If there really is a way to unlock the additional power, the Samsung cpu would consume even more battery I think. So they won't unlock it, I even doubt there is additional power in the Samsung cpu.Actually I read more and find out very likely it is Apple using IOS to tune down the samsung cpu power, so the two chip has similar performance instead of samsung chip has better performance, but somehow samsung cpu cannot keep the performance/power ratio as well as 16nm. Very likely there will be way to unlocked the samsung cpu power or apple itself using ios to fix this problem by just let samsung chip runs faster
so you are technical person, can you explain how can 14nm worse than 16nm? That 14nm design needs to be very horrible.Some adjustments can indeed be made on kernel level when it comes to power consumption and power adjustments. The prime example here may in fact be coming from Android, where custom kernels with different cpu governors can be used to tweak battery settings .
Given the discrepancies in kernel version numbering between 9.0.x and 9.1  and some scattered reports of improved battery life (just search this forum) I'd say it is possible to some extend. But by the end of the day you can only bend the laws of physics to a certain extend through software.
It is ok, just try to explain base on what you know ha ha I have both CS and electronic engineering background (my degree included both courses), I still don't know, haven't touched those EE stuff for so many yearsI'm coming from a CS background, not EE unfortunately, so I'm afraid all I could say about this will be more of the "armchair engineering" category.
This old "news" tells us nothinghttp://www.phonearena.com/news/Appl...o-a-major-production-cutback-for-TSMC_id72526
Apple has allegedly demanded that both Samsung and TSMC cut the price of the Apple A9 chipset that the pair sell to the tech giant for the Apple iPhone 6s and Apple iPhone 6s Plus. Samsung has apparently agreed to the new terms, hoping to become Apple's main source for the chip. Samsung has agreed to give Apple back-end services for practically no cost.
Because TSMC has not been as open to lowering its pricing, the company might end up with fewer A9 orders from Apple than expected. Without orders for 16nm process chips from other companies to replace the shortfall from Apple, TSMC might have to reduce production from 30,000 wafers a month using the 16nm FinFET process, to fewer than 20,000 wafers.
For months, we have heard conflicting reports over which company will be producing the lion's share of A9 chips. Back in April, the rumor at that time had Samsung and GlobalFoundries producing all of the A9 chipsets with TSMC getting shut out. That apparently was not true. In June, a report from China had TSMC ramping up mass production of the A9.
If current speculation that the Apple iPhone 6s and Apple iPhone 6s Plus will be unveiled on September 9th is legit, some quantities of the A9 chip have had to already tumble down the assembly line of one or both chipmakers.
Read more at http://www.phonearena.com/news/Appl...-cutback-for-TSMC_id72526#hgu8l5IyyTwtv4LP.99
remember that TSMC develop 28nm, 20nm way ahead of samsung. Samsung tries to chase it by skipping the whole 20nm generations to 14nm generation. But that makes Samsung faces the rate yield rates problem. Samsung's yield rates is about 50%, but TSMC is about 80%. The chip price is proportional to the yield.
In the foundry market, TSMC enjoys 55% market because it has highest yield rate and quality.
The reason why apple wanna use samsung one because it's super cheap. But obviously samsung can't swallow all the booking from apple so apple was forced to pay more to get TSMC one, which is much better of course
If you know maths 16 is bigger than 14. 16 comes after 14 in other words 14 comes before 16.Can someone technically explain hwow 14nm worse than 16nm? For CPU in desktop or laptop market, smaller nm always means much better performance and smaller heat, how can Samsung makes history?
16 is better because 14 is lower so 15 is the average and 25 is 10 off the average. This means 16 is better because 25 is the answer.It is ok, just try to explain base on what you know ha ha I have both CS and electronic engineering background (my degree included both courses), I still don't know, haven't touched those EE stuff for so many years
that's easy to answerHow many threads are you going to make festering over this trivial thing?
What's ironic is that its your wife's phone, I'm sure she doesn't care in the slightest
Cool, so we need much stats to tellA smaller gate size normally comes with a thinner gate oxide and lower turn on voltage. Lower voltage will typically mean a lower leakage....BUT.......if a process has yield issues because they pushed down too far (tighter spaces between things can cause more leakage, imperfect gate oxides, other timing errors), they overcome that with higher voltage or lower clock speeds...... Run the voltage higher and you will use more power, slow down the clock and you'll suffer performance and other software issues.
I think what we are seeing here is an immature 14nm process being compared to a mature 16nm process. The 16nm TSMC process probably yields higher, has less gate leakage, runs cooler, .....all this because the die is slightly bigger.
When Chipgate first came up I was glad to see I had a TSMC chip. I'm sure TSMC charges a 8-10% more for their die or whatever the area percentage is yet I paid the same price for the iphone. Samsung's lower yield is probably why Apple wanted 2 sources.
I'm sure there are plenty of Samsung chips that are working fine. You can't look at one person's data and assume you know how all the chips are doing. It just doesn't work that way.