64-bit ARM coming

[Cogman]

Completely incorrect. x86 is alive today because neither ARM nor any other platform has been used in any significant way to create the content we all consume with a myriad of devices (smartphones, tablets, etc.)

Wat?

He is completely correct. x86 is alive and well because of legacy. Nothing more. Intel would GLADLY leave the x86 arch for something more sane. In fact, several years ago they tried to do it (Itanium, which didn't take off, legacy being a big issue).

There will always be a need for devices that excel at creating content... and that demands a suitable nexus between processing power and cost, which isn't met by anything other than x86.

People don't create high powered desktop CPUs because they know that it is impossible to break into that market. Think about, how would a high powered arm fair for the desktop? It would need an desktop operating system, software, and to be nearly as fast as current offerings. Not only that, it would have to be able to play nice with existing hardware. For example, it would need DDR2 support, PCI-Express support, SATA support, etc.

In other words, you wouldn't just have to develop the CPU, but the entire architecture around it. That architecture also has to be able to support existing hardware (which means getting licenses etc). It is a HUGE investment.

Mobile platforms are much simpler because you don't have to worry about being compliant with standard techs, you have full control over all of the devices connected, and you don't have to worry so much about being as fast as your competitors.

Content creation is dominated by the x86 platforms because x86 platforms are the most popular content creation platforms. It has nothing to do with the architecture and everything to do with the fact that it has been around forever.

Intel lucked out. IBM chose them for their personal computer to keep the computer cheap (Ironic, huh. Intel has a SLOWER and less powerful arch than IBMs current RISC processor). IBM hit a homerun with their open architecture and as a result, intel hit a home run with their processor. Had the IBM PC not been as widely popular as it was, x86 would never have been the desktop processor of choice. The architecture is not superior, it is familiar.

 

[Ferzerp]

You realize that they are bringing this to the table by 2014 - likely 2015 - at the earliest?

And that Intel, not to mention AMD, will have spent gobs of money AND time increasing the performance and value of their platforms, not just their CPUs, in that amount of time?

Getting traction in the server market is not just about performance per watt. Intel is moving so far up the chain with Nehalem-EX that that by 2015 they will likely be seriously not only endangering Itanium into complete obsolescence but likely Power and Sparc as well (hot swap everything, hardware partitioning, etc.)

Finally, their (Intel) ability to scale their manufacturing processes so quickly means that they will be more power effecient than arm not solely due to their architecture (x86) but do to their ability to be consistently to be (at least) one manufacturing node ahead of their competition - typically with better technology around that manufacturing process to boot. What will change about that by 2015?

Magic tech journalist page view pixie dust is what will change about it. That's what ARM has that Intel lacks.

 

[blckgrffn]

Magic tech journalist page view pixie dust is what will change about it. That's what ARM has that Intel lacks.

Dammit, I knew I was overlooking something...

 

[ncalipari]

Do you know what you call an application server that has low IPC needs but can benefit from many threads?

A VM.

You can stack tens of them on a server now, hundreds by 2015. Running on big, nasty x86 CPUs.

ARM server hardware is going to have a huge uphill battle.

I never said that the actual ARM is going to take over the VM market.

There is no good CPU for everything.

ARM is great for NAS, Firewalls, Web Proxies, High density Blades....


Or the lower/mobile end of the Consumer market.


In 25 years I see a new architecture, spreading across Consumer and Server market, which is much more similar to ARM than x86

 

[blckgrffn]

I never said that the actual ARM is going to take over the VM market.

There is no good CPU for everything.

ARM is great for NAS, Firewalls, Web Proxies, High density Blades....


Or the lower/mobile end of the Consumer market.


In 25 years I see a new architecture, spreading across Consumer and Server market, which is much more similar to ARM than x86

For those purposes what is the point is 64 bit memory addressing? Already ARM can address a ton of ram, this is basically for individual processes... so... this is a check-box feature? High density blades exist to host VMs, FWIW.

25 years from now? In the computing world you can say anything in that time frame... the Singularity event will have happened by then in my view and we'll all be jacked in and able to take snapshots of our brains so that we can play the same video game over and over again for the first time.

 

[Ferzerp]

And lo these prophesies will come to pass when 2*10=10. Of course we will need a new architecture. Are old ones won't work with the new math.

 

[ncalipari]

For those purposes what is the point is 64 bit memory addressing? Already ARM can address a ton of ram, this is basically for individual processes... so... this is a check-box feature? High density blades exist to host VMs, FWIW.

A ton is not enough. There is always a need for more. And a 64 Bit CPU is better than using PAE.

25 years from now? In the computing world you can say anything in that time frame...

Businesses reason in longer timeframes than customers.

 

[ncalipari]

And lo these prophesies will come to pass when 2*10=10. Of course we will need a new architecture. Are old ones won't work with the new math.

it's sad to see that the technical level of anand forums dropped so low. Ignorance coupled with arrogance is a very hard to beat.

Spend more time at your local library

http://www.amazon.com/Access-Math-Ex...9813845&sr=1-1

 

[Ferzerp]

Ignorance coupled with arrogance is a very hard to beat.

I couldn't agree with you more.

I would like your reasoning how values that differ by a factor of 5 and not 10 can be apart by an order of magnitude. Your own reasoning without links. We all know what orders of magnitude are, but your logic appears to suggest that 9 and 10 are an order of magnitude apart.

 

[ncalipari]

9 and 10 are an order of magnitude apart.

9.0 and 10.0 are an order of magnitude aparts.

9.90 and 10.00 are an order of magnitude aparts.

9.990 and 10.000 are an order of magnitude aparts.


They taught me this in the second year of high school, a long time ago, when I learnt scientific notation.

Several years later, after a bachelor in math and a master in Statistics, I still believe this.

 

[Ferzerp]

Wow, now I want to know how deep the rabbit hole goes.

 

[Genx87]

People still believing Intels garbage about getting into tablets and mobile phones with devices using 10 watts? Have fun making a phone call with a 10 watt CPU next to your head.

Anyways Intel has been saying this for years. And for years I have yet to see an Intel based mobile phone in the wild. Tablets are out there with Intel chips. They also represent about 4% of the market. In other words they are irrelevant.

 

[Cogman]

9.0 and 10.0 are an order of magnitude aparts.

9.90 and 10.00 are an order of magnitude aparts.

9.990 and 10.000 are an order of magnitude aparts.


They taught me this in the second year of high school, a long time ago, when I learnt scientific notation.

Several years later, after a bachelor in math and a master in Statistics, I still believe this.

You did not learn your notation well enough...

By this logic, every number is an order of magnitude different from another number.

In science and engineering, when we say order of magnitude we mean 10x bigger.

 

[Cogman]

People still believing Intels garbage about getting into tablets and mobile phones with devices using 10 watts? Have fun making a phone call with a 10 watt CPU next to your head.

Anyways Intel has been saying this for years. And for years I have yet to see an Intel based mobile phone in the wild. Tablets are out there with Intel chips. They also represent about 4% of the market. In other words they are irrelevant.

They existed... And failed miserably (Xscale processor).. Though they weren't x86.

 

[ncalipari]

By this logic, every number is an order of magnitude different from another number.

Sorry?

The order of magnitude of the ratio of the masses can be obtained by comparing the exponents instead of the more error-prone task of counting the leading zeros

I even highlighted the important part, so it's easier to understand

 

[zsdersw]

Wat?

He is completely correct. x86 is alive and well because of legacy. Nothing more. Intel would GLADLY leave the x86 arch for something more sane. In fact, several years ago they tried to do it (Itanium, which didn't take off, legacy being a big issue).

The fact that you can't leave a legacy architecture without a huge investment of both time and money doesn't make EPIC (VLIW) or anything else superior or "more sane".

Content creation is dominated by the x86 platforms because x86 platforms are the most popular content creation platforms. It has nothing to do with the architecture and everything to do with the fact that it has been around forever.

Actually it does have something to do with the architecture. x86 wasn't always the popular or usual choice for many aspects of content creation. PowerPC, SPARC, and other purely RISC architectures (SGI, anyone?) were the traditional choices... with entire hardware/software ecosystems built around them, until the performance of x86, combined with the lower cost afforded by x86's popularity in general computing, made a good case for stepping out of those ecosystems. Lower cost and ubiquity alone wouldn't have made x86 compelling for those markets... but the performance improvements and other architectural changes to x86, made by both AMD and Intel, sealed the deal.

 

[ncalipari]

Actually it does have something to do with the architecture. x86 wasn't always the popular or usual choice for many aspects of content creation. PowerPC, SPARC, and other purely RISC architectures (SGI, anyone?) were the traditional choices... with entire hardware/software ecosystems built around them, until the performance of x86, combined with the lower cost afforded by x86's popularity in general computing, made a good case for stepping out of those ecosystems. Lower cost and ubiquity alone wouldn't have made x86 compelling for those markets... but the performance improvements and other architectural changes to x86, made by both AMD and Intel, sealed the deal.

And now everything is changing again by going with GPGPU

 

[Cogman]

The fact that you can't leave a legacy architecture without a huge investment of both time and money doesn't make EPIC (VLIW) or anything else superior or "more sane".

Actually it does have something to do with the architecture. x86 wasn't always the popular or usual choice for many aspects of content creation. PowerPC, SPARC, and other purely RISC architectures (SGI, anyone?) were the traditional choices... with entire hardware/software ecosystems built around them, until the performance of x86, combined with the lower cost afforded by x86's popularity in general computing, made a good case for stepping out of those ecosystems. Lower cost and ubiquity alone wouldn't have made x86 compelling for those markets... but the performance improvements and other architectural changes to x86, made by both AMD and Intel, sealed the deal.

Check your facts. IBM had a higher performing processor that they could have used for the IBM PC, but they chose the intel's 8088 instead because it was cheaper, not because it performed better. It was good enough.

IBM's PC took off, not because of performance, but because it was cheap and modular. With the old style of PCs, you bought everything, all the hardware was made by one manufacturer. With IBM's PC was designed from the beginning to be completely modular. It was the modularity that lead to IBM PC's being cheap and the modularity that ultimately caused IBM PCs to be the dominate desktop arrangement.

http://en.wikipedia.org/wiki/Intel_8088
http://en.wikipedia.org/wiki/IBM_801

 

[blckgrffn]

A ton is not enough. There is always a need for more. And a 64 Bit CPU is better than using PAE.



Businesses reason in longer timeframes than customers.

Businesses reason about a quarter at a time. Maybe a year if you are lucky.

If a business has real vision they are looking down the road further but if you think they are going to bet on anything technologically being sure further out than say, five years, that's crazy talk.

The path of least resistance seems to come out on top time and time again when you look how we got to where we are today with regards to the current technological landscape. Healthy skepticism is needed for the role ARM will play in any market where it has not already gained traction, ie there is currently resistance to it entering that market. IE, good enough technology is good enough Credit to Cogman for getting that put succinctly

 

[dguy6789]

9.0 and 10.0 are an order of magnitude aparts.

9.90 and 10.00 are an order of magnitude aparts.

9.990 and 10.000 are an order of magnitude aparts.


They taught me this in the second year of high school, a long time ago, when I learnt scientific notation.

Several years later, after a bachelor in math and a master in Statistics, I still believe this.

*pure disbelief*

 

[Ferzerp]

I gave up.

 

[kevinsbane]

The order of magnitude of the ratio of the masses can be obtained by comparing the exponents instead of the more error-prone task of counting the leading zeros

10^x vs 10^y, order of magnitude compares the difference between x and y.

2 = 10^X
log(2) = X * log 10
0.301029995 = X
2 = 10^0.301029995

10 = 10^X
log(10)=xlog(10)
1=X

X=0.3, Y =1, therefore they are definitely not an order of magnitude of difference; at best, if you could give partial orders of magnitude, then it would be a difference of 0.7, not 1.

9.0 and 10.0 are an order of magnitude aparts.

9.90 and 10.00 are an order of magnitude aparts.

9.990 and 10.000 are an order of magnitude aparts.

9 = 10^x
X = 0.954

9.0 is "0.046" orders of magnitude smaller than 10. If you could compare them that way.

 

[exar333]

[/B]

10^x vs 10^y, order of magnitude compares the difference between x and y.

2 = 10^X
log(2) = X * log 10
0.301029995 = X
2 = 10^0.301029995

10 = 10^X
log(10)=xlog(10)
1=X

X=0.3, Y =1, therefore they are definitely not an order of magnitude of difference; at best, if you could give partial orders of magnitude, then it would be a difference of 0.7, not 1.





9 = 10^x
X = 0.954

9.0 is "0.046" orders of magnitude smaller than 10. If you could compare them that way.

:thumbsup:

Thank God someone on here knows math!

The only application I am aware of where a 9.0 is an order of magnitude smaller than 10.0 is when measuring earthquakes.

It's pretty simple...

10 - 100 One order of magnitude larger
10 - 1000 Two orders of magnitude larger
So forth

 

[Ferzerp]

:thumbsup:

Thank God someone on here knows math!

The only application I am aware of where a 9.0 is an order of magnitude smaller than 10.0 is when measuring earthquakes.

It's pretty simple...

10 - 100 One order of magnitude larger
10 - 1000 Two orders of magnitude larger
So forth

Exar, *almost* everyone here knows math. The rest of us were more trying to figure out *why* he seems to think that. The guy who explained it was just being much nicer than we were

 

[Arkadrel]

*pure disbelief*

^this.

ARM vs x86 brings out the weird.

Intel is to big a bussiness to just get pushed out.
They *sit* on 80%+ marketshare.

ARM eventually gets close to x86, Intel works out something new, gets AMD to go along with it, and put out a product that can match ARM's current tech, or best it.

Intel has so much money, so much marketshare, and brilliant engineers.
Hell I can imagine even just removeing some old legacy stuff in x86 would make it more power effecient (enough so to compete with arm for along time).

If x86 falls, its because Intel came up with something better, and got AMD to go along with it.