Recent

% | $
Quotes you view appear here for quick access.

Intel Corporation Message Board

  • semiwiki semiwiki Sep 1, 2011 1:00 PM Flag

    Apple’s $399 Plan to Win Consumer Market in Summer 2012

    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • A 64-bit ARM is not needed for a smartphone or a tablet that is running small code-footprint applications. A 64-bit ARM would probably run at the SAME performance or slower than the 32-bit version.

      A 64-bit ARM is needed when you start solving problems that need more than 32-bits. IMO, you need a 64-bit CPU when you get to the laptop level problems. When a 32-bit CPU starts calliing a 64-bit math library to solve problems, you have hit the wall. I had problems with a 32-bit TurboTax. TurboTax hit the wall with some intermediate computation results.

      When ARM hits 64-bit, they will need some much, much better CPU designs to handle the apps that will run on them. It will take some time (if at all) for them to be successful.
      ------

      ARM 64bit WILL run slower than ARM 32bit. However, I am expecting ARM to have a thumb ISA (instruction compression) for 64 bit that will reduce much of the performance/power problems of going to 64bit.

      The jump to 64 bit is actually very simple. The floating point unit is 64 bits and the SIMD unit is 128 bits, it's only the integer and addressing modes that need 'building'. Clearly bandwidth needs improving, but that's for the SoC designers to sort out.

      At this stage we really don't know what the target market for these new 64bit cores are. They could just be for deeply embedded applications (such as network equipment) or they could be for laptop/desktop/server usage. Even then, they could still be targeting sub 1watt power consumption so wouldn't come close to Intel's high end in terms of performance.

    • The PowerPC looked very much like an Intel CPU to the end customer. It was produced by Moto or IBM? and a company could not really modify it to their own tastes. They could request and lobby for features but not implement what they needed. It was a general purpose CPU.

      ARM is a roll-your-own CPU where you can buy a license and buy access to a body of basic software, programmer knowledge and tools that do not exist to the same degree on the PowerPC. There is a low cost barrier for entry. Apply wrote a check and bought a company. Low cost barrier.

      A 64-bit ARM is not needed for a smartphone or a tablet that is running small code-footprint applications. A 64-bit ARM would probably run at the SAME performance or slower than the 32-bit version.

      A 64-bit ARM is needed when you start solving problems that need more than 32-bits. IMO, you need a 64-bit CPU when you get to the laptop level problems. When a 32-bit CPU starts calliing a 64-bit math library to solve problems, you have hit the wall. I had problems with a 32-bit TurboTax. TurboTax hit the wall with some intermediate computation results.

      When ARM hits 64-bit, they will need some much, much better CPU designs to handle the apps that will run on them. It will take some time (if at all) for them to be successful.

      The beauty of ARM is that ARMH does not really have to build and compete with their designs. Their customers do.

    • ltisteve@verizon.net ltisteve Sep 4, 2011 5:13 PM Flag

      Will Arm get any further against Intel with a 64bit CPU than the PowerPC did compared to Intel?

    • Well, yes, Trust Zone is a plan. I wonder how well it rolls out in the real world. Nobody really cares too much about security until all hell brakes loose.

      Until an cyber attack makes front page news, this won't be a buying factor for most consumers. But, once it does happen, things can change right away.

      What would be the advantages to a 64bit ARM processor aside from addressing more memory than phones and tablets currently offer? It would mostly make sense for desktop and laptops with larger amounts of memory, but other than that, why the excitement?
      ------

      Trustzone has been out for years. It's likely in your smart phone (even if it's not been used by the OS).

      Other than addressing, very little advantage. If anything, there are a lot of disadvantages. Larger core, more power. Will run slower than a 32bit core (like for like basis [like all RISC cores]). Yet another ISA to support (likely more than one). etc etc.

      Yes, the 64 bit part would be for laptops, desktops and servers not for phones/tablets. As for the excitement - there is none. If Apple is serious in dumping Intel (in years to come) it needs 64 bit asap.

    • ltisteve@verizon.net ltisteve Sep 2, 2011 12:06 PM Flag

      Well, yes, Trust Zone is a plan. I wonder how well it rolls out in the real world. Nobody really cares too much about security until all hell brakes loose.

      Until an cyber attack makes front page news, this won't be a buying factor for most consumers. But, once it does happen, things can change right away.

      What would be the advantages to a 64bit ARM processor aside from addressing more memory than phones and tablets currently offer? It would mostly make sense for desktop and laptops with larger amounts of memory, but other than that, why the excitement?

    • This thread has morphed into one about security.

      ARM currently has better on chip 'security' than Intel:

      http://www.arm.com/products/processors/technologies/trustzone.php

      Trustzone is a hardware virtual processor which would run all secure code. You'll see something similar in future Intel chips.

      As for specific instructions for encryption/random number generation, ARM doesn't contain any. In embedded applications you'll often find hardware random number generators and encryption engines. For real end user computing there is no ARM standard.

      Getting back to Apple. At this stage it doesn't matter how good (or bad) ARM performance is as there is no ARM 64 bit core. This should be announced in the next few months. Very little is known about this core and it's capabilities, it could be designed for deeply embedded applications and not end user computing at all.

    • Intel has leadership in an industry where there are only 2 or 3 companies can compete with the Intel fabrication muscle in place today. The capital equipment cost for any new entries is $10bil to $20bil to just become #2 or #3 or .... This is a tremdous competitive barrier.

      Intel has intellectual property patent rights on CPU, cache, chip bus, .... silicon designs. Intel has similar advanced IP patents granted on a number of advanced algorithms that they invented that make it very difficult for others to come close to matching performance.

      You can make an ARM CPU cheaper, but there are very high barriers to adding modern high speed cache and bus support without stepping on Intel IP toes. Security algorithms ..... same thing and I DOUBT VERY MUCH that these are included in any cross licensing agreement.

      When ATOM silicon gets close to the ARM power envelop, you will see the battle turn to PERFORMANCE and USER SECURITY. ARM has no answer. Methinks that Apple may not abandon the security of their customers. They are too enlightened to risk the silver lining of their iCloud.

    • ltisteve@verizon.net ltisteve Sep 1, 2011 9:39 PM Flag

      Oh, another article from my favorite Apple science fiction writer, Whacky Ed! Let's see today Whacky Ed had divulged top secret Apple information that Steve Jobs is privy to (BTW, has anybody told Whacky Ed Steve quit last week?).

      Well, anyway. Apple is putting Intel and AMD in a bidding war (though Apple has enough money in it's left pocket to buy AMD) and the end result is that they need a $75 processor. But, atlas Apple's plan to rule the world is almost complete. By next summer they will have single handed starved out the computer competition with $399 desktop computers. But there are over 300 Apple stores world wide. That means by next summer we'll have one Apple Store for every 20.6 million people on the planet! So, the old days of the former Soviet Union will look so good as people stand in line for weeks at a time to buy an IPad.

      Hey, someone please pass along to Whacky Ed that some guy in California will bet him that by Dell, Acer and Asus will still be in business, still selling computers. Not only that but Apple won't have over a 25% marketshare worldwide in computers. I will bet him $5000. I don't have that kind of money to lose, but that's how confident I am that this guy continues to write out of his A$$.

      BTW, you end up looking ridiculous posting this guy's thoughts. You know this is total fabrication, don't you?

      • 1 Reply to ltisteve
      • The price of the base machine is $399 and you are required to sign up for 2 years of iCloud at $25 per month. Sounds like cell phone pricing.

        $399 + 24*$25 =
        $399 + $600 = $999

        This appears to be a way of pricing a $999 machine cheap, to bootstrap iCloud subscriptions and therefore iCloud success.

        Consumer PC replacements based on a (PowerPC -> Intel -> ARM) migrating application environment will have to have all the applications either ported or emulation supported.

        Fortune500 company IT departments are not anxious to support yet another set of hardware and software applications.

        It is unlikely that ARM will be able to come close Intel performance in the near future.

    • More of the semi-wacky attack on Intel with the idea that Apple is going to take over the world with ARM processors.

      Ain't gonna happen.

      The idea that a low end capability Apple/ARM laptop is going to be a game changer is just silly.

      Just think about the legacy issues.

      When the ultrabook price points are coming down, real users will want the real deal.

      Dunno, if a low end ARM Apple laptop happens (big if), maybe it can compete with Chromebooks.

      Given the waves of Intel fabrication improvements on the way, Apple has to be giving a move to 3-D serious consideration across all lines.

      Who needs a low end laptop that won't run most applications?

      Nobody...

      PS. I'm getting really tired of the semi-wacky, all the way tacky yellow journalistic approach to trying to build readership...

    • AMD inside?????????????????????

 
INTC
33.27-0.63(-1.87%)Jun 2 4:00 PMEDT