ARM is Already Considering 128-bit Mobile CPU

[JOSHSKORN]

What am I missing here? What's the point of a 128-bit SoC? Just to say it's not on an iPhone, yet but on Android? We DO need a 64-bit Smartphone, badly. But, 128-bit is overkill. Even for a desktop, the consumers only need a 64-bit chip. Businesses might benefit, though. Possibly, libraries and internet cafes.

 

[airplanegeek]

they'll put a i7 in a phone someday

 

[ericburnby]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Let's say you had a 64bit program that performed image manipulation. The program itself was 10MB and it allocated 240MB of RAM for data to work with for a total of 250MB of RAM.

Now here's a question for every one on Tom's who continually posts that you need 4GB of RAM to make use of 64bit.

Do you think this program, only occupying 250MB of RAM, will not give you any advantage on a system with less then 4GB of RAM? Think very carefully about your answer. This should weed out the people who actually understand 64bit from those that don't.

I think you need to understand the architecture a bit better. The 10 Meg program and the 240 Meg alloc'd space are stored in different segments. The OS takes care of the program while the program takes care of the data. Ignoring a bunch of details, essentially the OS is only going to be storing offsets, not an entire 64 bit pointer and the offsets are shorter than 64 bits. The virtual address space uses some creative techniques to appear larger and segments the pieces for easy conversion. Thus your 64 bits aren't useful except for much except physical addresses.

Are you going to answer my question or not?

BTW, I've been coding since the early 80's, so don't worry about whether or not I understand the architecture or what's going on. I'm trying to dumb things down for a "certain" group of users here, hence the wording of my example.

So the question stands. Is a 64bit program running on a 64bit processor going to gain any advantages if it's only using a portion of the available memory? It's not a difficult question.

 

[InvalidError]

Meanwhile 99.99% of mobile apps are still 32-bit....

Most mobile apps on Android are written in Java. They will use however as many bits as the Java recompiler and platform CPU support anywhere the JRE developers deem it necessary or beneficial to use more bits.

So if there are parts of the JRE where 128bits may be useful, the JRE will make use of it if available. Read my previous posts in this threads for examples of where this may happen.

This is one of the advantages of using a programming language with intermediate bytecode: the runtime environment can recompile and optimize the code for whatever hardware is available instead of requiring a platform-specific binary from the original developer.

 

[tomfreak]

This must be a joke created by Ifags.

 

[SessouXFX]

There's a an article saying otherwise....

http://www.maximumpc.com/arm_denies_claims_working_128-bit_chips_2013

 

[Pinhedd]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Let's say you had a 64bit program that performed image manipulation. The program itself was 10MB and it allocated 240MB of RAM for data to work with for a total of 250MB of RAM.

Now here's a question for every one on Tom's who continually posts that you need 4GB of RAM to make use of 64bit.

Do you think this program, only occupying 250MB of RAM, will not give you any advantage on a system with less then 4GB of RAM? Think very carefully about your answer. This should weed out the people who actually understand 64bit from those that don't.

I think you need to understand the architecture a bit better. The 10 Meg program and the 240 Meg alloc'd space are stored in different segments. The OS takes care of the program while the program takes care of the data. Ignoring a bunch of details, essentially the OS is only going to be storing offsets, not an entire 64 bit pointer and the offsets are shorter than 64 bits. The virtual address space uses some creative techniques to appear larger and segments the pieces for easy conversion. Thus your 64 bits aren't useful except for much except physical addresses.

Are you going to answer my question or not?

BTW, I've been coding since the early 80's, so don't worry about whether or not I understand the architecture or what's going on. I'm trying to dumb things down for a "certain" group of users here, hence the wording of my example.

So the question stands. Is a 64bit program running on a 64bit processor going to gain any advantages if it's only using a portion of the available memory? It's not a difficult question.

Surely you must be joking.

X86 microprocessors have been using >32 bit physical address spaces since 1995 with Windows support since 2000 (server versions only). AMD64/EM64T extended the virtual address space from 32 bits to 64 bits (of which 48 bits are used for addressing) but this represents only one of many changes introduced by AMD64/EM64T. It's somewhat silly that the most touted benefit of the 64 bit extensions is one that was not in fact introduced by the 64 bit extensions at all while the actual benefits are completely overlooked.

Other benefits include:

native support for NX/XD bit which protects against buffer overflows

native 64 bit operations in the scalar and memory units

access to twice as many GP and XMM CPU registers, each of which is 64 bits wide rather than 32

removal of numerous legacy features

software defined multitasking rather than hardware defined multitasking

 

[falchard]

lol they must not have computer engineers vetting these press releases. No point in 64 bit if you are running less than 2GB of ram. When you are under you actually have a performance hit for using 64-bit since all the int sizes are twice as big, and casting to a lower sized int means you will not be utilizing the CPU effectively.
I would imagine 128-bit would be a much larger hit.

 

[fil1p]

While I appreciate all advancements in technology, I am really looking forward to the increase in battery life in mobile devices more than more power from smartphone CPU's. Advancements are being made, but it would be nice to have a phone that can easily last a couple of days on one charge while retaining the processing power we have/will have.

 

[Cons29]

for mobile, i'd rather have a good battery that can last for at least 2 days of good usage. don't care if they reduce power consumption or develop a new battery type (or do both)

 

[Pinhedd]

lol they must not have computer engineers vetting these press releases. No point in 64 bit if you are running less than 2GB of ram. When you are under you actually have a performance hit for using 64-bit since all the int sizes are twice as big, and casting to a lower sized int means you will not be utilizing the CPU effectively.
I would imagine 128-bit would be a much larger hit.

Computer engineer here, you have absolutely no idea what you are talking about

 

[InvalidError]

No point in 64 bit if you are running less than 2GB of ram.

There are plenty of uses for 64+bits aside from memory addressing.... plenty of them in the comments if you had bothered reading before posting.

 

[bnot]

By Jane McEntegartNOVEMBER 24, 2013 9:00 AM should try harder, apparently not even actually picking up the phone and asked an arm PR rep for clarification etc, , they could have told her
as doron ,did on November 24, 2013 10:20 AM here

that on Nov 22, 2013 10:45:00 AM on the official ARM community blog space Ian Drew CMO and Executive VP said this

"128 bits is 64 bits too many
Posted by Ian Drew in ARM Processors on Nov 22, 2013 10:45:00 AM
Did you hear the joke about the 128 bit processor? Let me offer some factual corrections to an article which appeared a couple of days ago in the Korea Herald and which has fueled speculation…"

".... News reports have suggested that ARM is developing 128-bit processor technology: this is not true. 64-bit processors are capable of supporting the needs of the computing industry now and for many years to come. There are absolutely no plans underway for 128 bit ARM-based chips because they simply aren’t needed. Rumors to the contrary are simply incorrect.

Furthermore, comments attributed to any ARM executive including my colleague Antonio Viana that allegedly discuss any specific partner’s chip plans for the future or 128 bit development are inaccurate: no such comments have been made."....

 

[pocketdrummer]

I cannot fathom a situation within the next 10 years that would necessitate a 128-bit processor in a cell phone.

Put away the tape measure.

(also, can we see some verification this isn't a hoax?)

 

[pacomac]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Sounds like a typical noob response! There is more to 64bit than simply the anount of addressible memory as Apple proved with the A7. And what the heck is a 64bit structure when it's at home? Pointer and some variable type sizes are the only thing that change when using 64bit.

 

[falchard]

No point in 64 bit if you are running less than 2GB of ram.

There are plenty of uses for 64+bits aside from memory addressing.... plenty of them in the comments if you had bothered reading before posting.

My bad read through the comments and someone posted Vector4 and I immediately realized my error.

 

[ssarathe]

I had kept quite all these days but now this is the limits - Tomshardware used to be one of my most favorite sites for news but lately all the news is just LATE... You guys are easily 3 - 5 days late on the news and they are not really NEW anymore...

This news essentially is a week old and what's more embrassing is that after this news was floated out ARM's CMO or some executive guy came out to shed this piece of information as a rumour and gave a press note stating ARM has no plans in the near future for a 128 bit process because there simply isn't any need...

Seriously guys - reporting news a day or a week isn't a huge deal but reporting something that's 10 days old and was even rubbished by the very own company as NEWS is just plain ridiculous...

 

[Khalan]

This has been linked to in the comments already but seems to have been overlooked:
"There are absolutely no plans underway for 128 bit ARM-based chips because they simply aren’t needed. Rumors to the contrary are simply incorrect."

http://community.arm.com/groups/processors/blog/2013/11/22/128-bits-is-64-bits-too-many?sf19780825=1

 

[JOSHSKORN]

If we do end up with 128bit CPUs, what would this do for integrated GPU?

 

[Steveymoo]

64 bit is necessary to future proof mobile chips. 64 bit for desktop is really only necessary for heavy work users, but also allows game developers to expand a bit (which is a good thing.)

128 bit however, is completely unnecessary. I can't think of any application on earth that would require 1 yottabyte of memory. Marketing BS.

 

[InvalidError]

128 bit however, is completely unnecessary. I can't think of any application on earth that would require 1 yottabyte of memory. Marketing BS.

Cryptographic applications require less than 1MB RAM excluding the data being fed through it yet could use a 4096bits CPU to manipulate keys using single integer operations instead of breaking them down into thousands of 32-64bits steps.

There is more to "bitness" than just addressable memory.

 

[ddpruitt]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Let's say you had a 64bit program that performed image manipulation. The program itself was 10MB and it allocated 240MB of RAM for data to work with for a total of 250MB of RAM.

Now here's a question for every one on Tom's who continually posts that you need 4GB of RAM to make use of 64bit.

Do you think this program, only occupying 250MB of RAM, will not give you any advantage on a system with less then 4GB of RAM? Think very carefully about your answer. This should weed out the people who actually understand 64bit from those that don't.

I think you need to understand the architecture a bit better. The 10 Meg program and the 240 Meg alloc'd space are stored in different segments. The OS takes care of the program while the program takes care of the data. Ignoring a bunch of details, essentially the OS is only going to be storing offsets, not an entire 64 bit pointer and the offsets are shorter than 64 bits. The virtual address space uses some creative techniques to appear larger and segments the pieces for easy conversion. Thus your 64 bits aren't useful except for much except physical addresses.

Are you going to answer my question or not?

BTW, I've been coding since the early 80's, so don't worry about whether or not I understand the architecture or what's going on. I'm trying to dumb things down for a "certain" group of users here, hence the wording of my example.

So the question stands. Is a 64bit program running on a 64bit processor going to gain any advantages if it's only using a portion of the available memory? It's not a difficult question.

Short answer is no, but now I'm thinking you agree with me that 128 bit is useless. And I've known lot's of people that have coded for decades that have no clue how the underlying system works. I always cringe when I hear "I've coded...", it tells me you understand programming and not much else.

 

[ddpruitt]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Sounds like a typical noob response! There is more to 64bit than simply the anount of addressible memory as Apple proved with the A7. And what the heck is a 64bit structure when it's at home? Pointer and some variable type sizes are the only thing that change when using 64bit.

64 bit refers to the word size, I suggest you look it up.

 

[ddpruitt]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Let's say you had a 64bit program that performed image manipulation. The program itself was 10MB and it allocated 240MB of RAM for data to work with for a total of 250MB of RAM.

Now here's a question for every one on Tom's who continually posts that you need 4GB of RAM to make use of 64bit.

Do you think this program, only occupying 250MB of RAM, will not give you any advantage on a system with less then 4GB of RAM? Think very carefully about your answer. This should weed out the people who actually understand 64bit from those that don't.

I think you need to understand the architecture a bit better. The 10 Meg program and the 240 Meg alloc'd space are stored in different segments. The OS takes care of the program while the program takes care of the data. Ignoring a bunch of details, essentially the OS is only going to be storing offsets, not an entire 64 bit pointer and the offsets are shorter than 64 bits. The virtual address space uses some creative techniques to appear larger and segments the pieces for easy conversion. Thus your 64 bits aren't useful except for much except physical addresses.

Are you going to answer my question or not?

BTW, I've been coding since the early 80's, so don't worry about whether or not I understand the architecture or what's going on. I'm trying to dumb things down for a "certain" group of users here, hence the wording of my example.

So the question stands. Is a 64bit program running on a 64bit processor going to gain any advantages if it's only using a portion of the available memory? It's not a difficult question.

Surely you must be joking.

X86 microprocessors have been using >32 bit physical address spaces since 1995 with Windows support since 2000 (server versions only). AMD64/EM64T extended the virtual address space from 32 bits to 64 bits (of which 48 bits are used for addressing) but this represents only one of many changes introduced by AMD64/EM64T. It's somewhat silly that the most touted benefit of the 64 bit extensions is one that was not in fact introduced by the 64 bit extensions at all while the actual benefits are completely overlooked.

Other benefits include:

native support for NX/XD bit which protects against buffer overflows

native 64 bit operations in the scalar and memory units

access to twice as many GP and XMM CPU registers, each of which is 64 bits wide rather than 32

removal of numerous legacy features

software defined multitasking rather than hardware defined multitasking

I think you proved his point, your mixing a whole lot of different concepts into your response. Sadly none of them deal with the xx bit debate.

 

[Pinhedd]

128 bit is pointless. I can't even remember the last time I needed to deal with a 64 bit structure. 64 bit is useful for addressing more than 4 gigs of RAM after that it's use is limited.

Let's say you had a 64bit program that performed image manipulation. The program itself was 10MB and it allocated 240MB of RAM for data to work with for a total of 250MB of RAM.

Now here's a question for every one on Tom's who continually posts that you need 4GB of RAM to make use of 64bit.

Do you think this program, only occupying 250MB of RAM, will not give you any advantage on a system with less then 4GB of RAM? Think very carefully about your answer. This should weed out the people who actually understand 64bit from those that don't.

I think you need to understand the architecture a bit better. The 10 Meg program and the 240 Meg alloc'd space are stored in different segments. The OS takes care of the program while the program takes care of the data. Ignoring a bunch of details, essentially the OS is only going to be storing offsets, not an entire 64 bit pointer and the offsets are shorter than 64 bits. The virtual address space uses some creative techniques to appear larger and segments the pieces for easy conversion. Thus your 64 bits aren't useful except for much except physical addresses.

Are you going to answer my question or not?

BTW, I've been coding since the early 80's, so don't worry about whether or not I understand the architecture or what's going on. I'm trying to dumb things down for a "certain" group of users here, hence the wording of my example.

So the question stands. Is a 64bit program running on a 64bit processor going to gain any advantages if it's only using a portion of the available memory? It's not a difficult question.

Surely you must be joking.

X86 microprocessors have been using >32 bit physical address spaces since 1995 with Windows support since 2000 (server versions only). AMD64/EM64T extended the virtual address space from 32 bits to 64 bits (of which 48 bits are used for addressing) but this represents only one of many changes introduced by AMD64/EM64T. It's somewhat silly that the most touted benefit of the 64 bit extensions is one that was not in fact introduced by the 64 bit extensions at all while the actual benefits are completely overlooked.

Other benefits include:

native support for NX/XD bit which protects against buffer overflows

native 64 bit operations in the scalar and memory units

access to twice as many GP and XMM CPU registers, each of which is 64 bits wide rather than 32

removal of numerous legacy features

software defined multitasking rather than hardware defined multitasking

I think you proved his point, your mixing a whole lot of different concepts into your response. Sadly none of them deal with the xx bit debate.

Most of those benefits refer to the AMD64/EM64T extensions in particular, but some of them are generic and would apply to any microarchitecture.

With a full 64 bit ALUs 64 bit instructions can be performed without breaking them down into multiple 32 bit instructions

With full 64 bit physical addressing the physical address space is wide enough that it allows for the NX/XD bit to be present in the page table entry (it occupies the MSB). The virtual address space may still be limited to 32 bits if desired (ala PAE).