Samsung: our next-gen smartphones will have 64-bit chips, too

[OneSpeed]

Hey guest thats all great on a desktop computer.

But who the hell is working on 4GB+ files and large data sets on a 4 inch Iphone?

64bit on cell phone will be the same as it was when it came out on computers at the start there will only be small benefit and as time goes on it will increase but its not going to make your phone 100% faster no matter what the marketing slides show you.


Stop drinking apple juice dude.

Hey I work with 4 GB files on my phone all the time all 4, 8 or 16 of them and can't wait to start working with 5 GB files... well all 3, 6 or 12 of them :-D

What applet you running that requires 64bit? Didn't know it took an iPhone to need 64bit processing to play music.

 

[St1ckM4n]

Apple shoved a 64-bit CPU in there because they have no other 'big leaps' cards left to play. They have cornered themselves in their own market.

- Screen size: the move to 4 inches was completely against company ethos and personal belief of Jobs - unless they all lied. The screen size won't change any time soon on their devices, because it is linked to:

- Resolution: The almighty Retina screen, there is none better! Or, so we are led to believe. While Samsung/HTC and everyone else didn't treat their customers like dummies, they continued to improve aspects of the device which are tangible and still have room for improvement. Unfortunately, Apple cannot increase their resolution easily. It is heavily linked to screen size and even more critically their app dev program. AFAIK Apple has no custom resolution/screen scaling. If res changes, the entire dev community is screwed.

Basically, shoving a 64bit chip in there was the cheapest, quickest and easiest way to give the iPhone a 'bump'.

 

[Darth Shiv]

There is no rush for 64bit because no phone will use more than 2GB with current OSes. Maybe as we get 4K smartphones we will need more than 2GB of RAM.

Let me ask what you would prefer to arrive on the scene first. The problem, or the solution to the problem? As some of the TechSpot staff would probably say it, "There is no practical need for this technology". You used word "rush" instead of "need" but the mindset is similar.

Do you want there to be a rush for it? Do you want to pay exorbitant prices for technology that is in high demand.Technology arriving on the scene in advance of a rush for it(or a high demand) will always allow the technology to mature and the cost to come down.

I'll probably be demonized again by some superhero trolling these forums speaking for those without a voice but my take on all of these technologies is let them come, let them come.

"Nobody Will Ever Need More Than 640k RAM!" -- Bill Gates, 1981"

I am not 100% sure about this but isn't Windows 8 Pro a 64-bit OS...doesn't it work on cellphone?

The fact alone that a technology exists sometimes is enough to increase the demand for it as well. In other words, "build it and they will come".

Yes sadly our Government is using the same "Nobody Will Ever Need More Than 640k RAM!" argument for our internet. Australia ranked 48th in world internet speeds. The Govt wants to supply a "guaranteed" 25mbit down and 4-6mbit up in 3-6 years for almost $30 billion.

The previous plan which is essentially now going to be scrapped was to rollout FTTP to all premises that have access to ADSL2+ currently (possibly more than that).

 

[p51d007]

Same debate when the switch was made from x32 to x64 in the PC world. 99% of users have no idea, so it won't make a difference to them anyway.

 

[Geforcepat]

I figured somebody was gonna come out and say me too sometime this week. and looky here

 

[TheBigFatClown]

Of course "we people" know you can calculate bigger numbers, more precision, etc. What the h*ck would you be doing in a smartphone? Managing databases? ARM 64-bit for servers is fine, but for a smartphone? Come on. I'm not even sure those chips have FPU that can take advantage of the much needed higher floating point precision [more accurate decimals].

As the article states: supporting 64-bit environment doesn't guarantee performance, but it does guarantee more memory usage -just think about memory consumed by pointers and the incalculable amount of pointers apps may use and all at the same time; they would simply use double the amount of memory in all those declared pointers [not double the total memory used by the app].

Did you make this same argument when the world shifted from 16-bit to 32-bit? It doesn't sound very plausible to me.

 

[EEatGDL]

Did you make this same argument when the world shifted from 16-bit to 32-bit? It doesn't sound very plausible to me.

When the world did that shift I was too young to make this argument. I don't know what half of what I wrote you're referring to, I think you're talking about the second half, because on the first half I didn't state anything.
Back then when we had 16-bit processors, Intel was already "cheating" using 20-bit for addressing and the integer number limit was easily reachable (signed or unsigned) by daily countability and calculations among with other things. Let me illustrate my argument: let's say a program has 1000 pointers (variables that store the address of other variable or a function), by the simple architecture fact, the same program will allocate 2000, 4000 and 8000 bytes for 16, 32, and 64 bits respectively for just pointers.
Anyway, two technical points about Android: it doesn't have swapping due to the limited write cycles of the flash memory it uses, so its way to make room for more programs that require RAM when it's full is simply killing "old and no longer used apps to free resources"; the second thing is that you only have the illusion of multitasking, you can't actually see several apps at a time and simply shift focus by finger-pointing on the apps you're watching on screen [from all the apps you would see at once on screen]. So we can only focus in one app at a time and it simply kills non-essential processes; what app would use that much RAM alone?
If you compare a program compiled in x86-32 and x64 you won't see double the performance; only the slight gains provided for arithmetic operations with long (64-bit) variables that the program may execute or BigInteger (as big as you need the integers to be technically without limit, imagine adding two 512-bit variables) variables that the program may use through an API in several object-oriented programming languages.

 

[TheBigFatClown]

When the world did that shift I was too young to make this argument. I don't know what half of what I wrote you're referring to, I think you're talking about the second half, because on the first half I didn't state anything.
Back then when we had 16-bit processors, Intel was already "cheating" using 20-bit for addressing and the integer number limit was easily reachable (signed or unsigned) by daily countability and calculations among with other things. Let me illustrate my argument: let's say a program has 1000 pointers (variables that store the address of other variable or a function), by the simple architecture fact, the same program will allocate 2000, 4000 and 8000 bytes for 16, 32, and 64 bits respectively for just pointers.
Anyway, two technical points about Android: it doesn't have swapping due to the limited write cycles of the flash memory it uses, so its way to make room for more programs that require RAM when it's full is simply killing "old and no longer used apps to free resources"; the second thing is that you only have the illusion of multitasking, you can't actually see several apps at a time and simply shift focus by finger-pointing on the apps you're watching on screen [from all the apps you would see at once on screen]. So we can only focus in one app at a time and it simply kills non-essential processes; what app would use that much RAM alone?
If you compare a program compiled in x86-32 and x64 you won't see double the performance; only the slight gains provided for arithmetic operations with long (64-bit) variables that the program may execute or BigInteger (as big as you need the integers to be technically without limit, imagine adding two 512-bit variables) variables that the program may use through an API in several object-oriented programming languages.

Sounds like a good case against Google putting out a 64-bit phone since they use Android. And anybody else who uses Android. But it doesn't sound like this is inherently a problem with 64-bit devices. Only issues with the operating system that might drive them.

You have made the general point clear about the cost of memory each pointer requires as the memory address space increases exponentially. It's not a linear progression. When you double the size of a pointer you get way more than double the memory address space. So, from my point view, it seems you get more bang for you buck.

 

[Guest]

When are the 64 bit apps coming??? As an apple user, I can hardly wait. BTW, what app on a phone would use 64 bits? Anyone? Will we have those apps before android comes out with 64 bits of computing power? Do I really need it now?

 

[EEatGDL]

So, from my point view, it seems you get more bang for you buck.

I'm not saying 64-bit itself is a problem; but there's no use for that other than marketing, they're totally forgetting the main purpose of those embedded systems, they want to see them like full computers when they're actually not that -that's why for servers is great, but for a smartphone? And maybe you got the idea wrong, let's stick to the same example: imagine we have 10,000 bytes (10 KiB) of memory, and we have nothing in memory but 1,000 pointers (nothing useful but for illustrating) so while the 16-bit processor would set the pointer size to 2 bytes, those pointers would consume 2/10 KiB while in the 64-bit processor 8/10 KiB would be used for that sole purpose; so you actually get less for your buck.

It is very hard to calculate the amount of pointers needed by running applications and system, but you would find more memory consumption in the same scenario between 32-bit and 64-bit.

 

[OneSpeed]

I think I'll wait to see what, if any, apps are 64 bits before I jump onto that bandwagon. I mean, c'mon.

 

[TheBigFatClown]

I'm not saying 64-bit itself is a problem; but there's no use for that other than marketing, they're totally forgetting the main purpose of those embedded systems, they want to see them like full computers when they're actually not that -that's why for servers is great, but for a smartphone? And maybe you got the idea wrong, let's stick to the same example: imagine we have 10,000 bytes (10 KiB) of memory, and we have nothing in memory but 1,000 pointers (nothing useful but for illustrating) so while the 16-bit processor would set the pointer size to 2 bytes, those pointers would consume 2/10 KiB while in the 64-bit processor 8/10 KiB would be used for that sole purpose; so you actually get less for your buck.

It is very hard to calculate the amount of pointers needed by running applications and system, but you would find more memory consumption in the same scenario between 32-bit and 64-bit.

I must be missing something. If you use double the memory pointer references moving from 32-bit to 64-bit, so what? The memory address space has way more than doubled. So your still way ahead of the game in terms of how much "more" free memory address space you have using a 64-bit memory address space.

 

[EEatGDL]

I must be missing something. If you use double the memory pointer references moving from 32-bit to 64-bit, so what? The memory address space has way more than doubled. So your still way ahead of the game in terms of how much "more" free memory address space you have using a 64-bit memory address space.

Yes of course; but smartphones haven't got near 3GB yet; they are still launching 2GB flagships.

 

[OneSpeed]

I'm not saying 64-bit itself is a problem; but there's no use for that other than marketing, they're totally forgetting the main purpose of those embedded systems, they want to see them like full computers when they're actually not that -that's why for servers is great, but for a smartphone? And maybe you got the idea wrong, let's stick to the same example: imagine we have 10,000 bytes (10 KiB) of memory, and we have nothing in memory but 1,000 pointers (nothing useful but for illustrating) so while the 16-bit processor would set the pointer size to 2 bytes, those pointers would consume 2/10 KiB while in the 64-bit processor 8/10 KiB would be used for that sole purpose; so you actually get less for your buck.

It is very hard to calculate the amount of pointers needed by running applications and system, but you would find more memory consumption in the same scenario between 32-bit and 64-bit.

I must be missing something. If you use double the memory pointer references moving from 32-bit to 64-bit, so what? The memory address space has way more than doubled. So your still way ahead of the game in terms of how much "more" free memory address space you have using a 64-bit memory address space.

OK, 64 bit CPU on iPhone 5 which has 1 GB RAM (Androids have 2 GB RAM on 32 bit), and no micro sd (Samsung Galaxy S4 has micro sd). Now how much better off are you with freed up memory address space? Flawed logic with 64 bit on a 1 GB RAM phone. If you want more RAM memory, you are better off with an Android, and not the iPhone 5.

 

[TheBigFatClown]

OK, 64 bit CPU on iPhone 5 which has 1 GB RAM (Androids have 2 GB RAM on 32 bit), and no micro sd (Samsung Galaxy S4 has micro sd). Now how much better off are you with freed up memory address space? Flawed logic with 64 bit on a 1 GB RAM phone. If you want more RAM memory, you are better off with an Android, and not the iPhone 5.

Aha, I think I see where everybody is coming from. Your saying that 64-bit is of little value when you have little memory on the device. Yes, I would have no problem agreeing to that. It wouldn't even make sense to use 64-bit memory address space until the device you are using it on has, lets say, double the memory space of what can be achieved with 32-bit memory address space. Maximum 32-bit memory address space is 4GB. So, any cellphone/tablet that touts a minimum memory configuration of at least 8 GB would reap the benefits of 64-bit memory address space. Any cellphone/tablet with less than 8GB of memory would be silly. Can we all agree on that?

 

[OneSpeed]

OK, 64 bit CPU on iPhone 5 which has 1 GB RAM (Androids have 2 GB RAM on 32 bit), and no micro sd (Samsung Galaxy S4 has micro sd). Now how much better off are you with freed up memory address space? Flawed logic with 64 bit on a 1 GB RAM phone. If you want more RAM memory, you are better off with an Android, and not the iPhone 5.

Aha, I think I see where everybody is coming from. Your saying that 64-bit is of little value when you have little memory on the device. Yes, I would have no problem agreeing to that. It wouldn't even make sense to use 64-bit memory address space until the device you are using it on has, lets say, double the memory space of what can be achieved with 32-bit memory address space. Maximum 32-bit memory address space is 4GB. So, any cellphone/tablet that touts a minimum memory configuration of at least 8 GB would reap the benefits of 64-bit memory address space. Any cellphone/tablet with less than 8GB of memory would be silly. Can we all agree on that?

Peace. I would agree with that, but would say at minimum " more than 4 GB of RAM.

 

[Railman]

These new chips could lead to a future 64bit Raspberry Pi!

 

[EEatGDL]

Aha, I think I see where everybody is coming from. Your saying that 64-bit is of little value when you have little memory on the device. Yes, I would have no problem agreeing to that. It wouldn't even make sense to use 64-bit memory address space until the device you are using it on has, lets say, double the memory space of what can be achieved with 32-bit memory address space. Maximum 32-bit memory address space is 4GB. So, any cellphone/tablet that touts a minimum memory configuration of at least 8 GB would reap the benefits of 64-bit memory address space. Any cellphone/tablet with less than 8GB of memory would be silly. Can we all agree on that?

I agree on that, that's why I'm saying "right now" is pointless. They're launching with 2GB still not making use of them and to make good use of it, they'll have to at least triple that and who would develop to consume so much in so little time? It would be like having 4GB on a 64-bit computer, when you can actually make a better use of it byte per byte on a 32-bit computer [even if it detects ~3.25GB]; if you're using a 64-bit PC, go at least for 6GB and make sure you can efficiently use multi-core that can process all that data in the worst case scenario.

 

[Lionvibez]

Aha, I think I see where everybody is coming from. Your saying that 64-bit is of little value when you have little memory on the device. Yes, I would have no problem agreeing to that. It wouldn't even make sense to use 64-bit memory address space until the device you are using it on has, lets say, double the memory space of what can be achieved with 32-bit memory address space. Maximum 32-bit memory address space is 4GB. So, any cellphone/tablet that touts a minimum memory configuration of at least 8 GB would reap the benefits of 64-bit memory address space. Any cellphone/tablet with less than 8GB of memory would be silly. Can we all agree on that?

+1

Agreed!

 

[OneSpeed]

These new chips could lead to a future 64bit Raspberry Pi!

64 bit computing power on your smartwatch

 

[Guest]

http://electronics.howstuffworks.com/tech-myths/5-myths-about-bill-gates3.htm

regarding 640k RAM quote.

 

[Alecks]

I would just like to point out that Android, being Linux-based, can still use the x32 ABI, which was designed specifically with mobile devices in mind. The benefits of 64bit while using less memory.

 

[backo]

And may I ask what phone you have and its current storage capacity including any SD cards?

One by Sarcasm Phones... that was exactly my point... I do not see any real world necessity for 64-bit on a mobile.