ARM for the future?
  • Order of the Butterfly
    Order of the Butterfly
    In_Correct
    Posts: 245 from 2012/10/14
    From: DFW, TX, USA
    Quote:

    Jim wrote:
    @takemehomegrandma

    You are on quite a roll tonight.
    I hope we continue to support discrete Radon video cards too, but AMD APUs offer similar gpus in an integrated package.
    It wouldn't bother me in the least if we supported those.
    And I have never favored supporting only the largest player in the industry, as lack of competition defeats innovation.
    Plus, when Intel made the mistake of betting on the Pentium 4, AMD products were kicking their ass.
    I don't anticipate Intel failing this way in the future, but without competition they WILL try to feed us similar crap.
    As to the comparable power of PPCs vs. X64 processors, they are not nearly as far apart as you seem to think. Even ARM is approaching a competitive stance with X64.
    And, as you have pointed out, its all microcode based anyway.



    +1 Thanks. 8-)
    :-) I Support Quark Microkernel. :-D
  • »29.08.15 - 01:28
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    takemehomegrandma
    Posts: 2720 from 2003/2/24
    Quote:

    Jim wrote:
    @takemehomegrandma

    AMD APUs offer similar gpus in an integrated package.


    You can't be serious?! You don't see the difference between a proper Radeon R9 GFX card (or R7 or most other current Radeon GFX cards) on a proper Intel Broadwell Core i5 platform with decent performance Intel chipset, and some integrated AMD APU crap? There was a point in time where you could say about AMD CPU's that "well hey, at least it's cheap". But that was a looong time ago, IMHO AMD lost its relevance on desktop computing somewhere back when PPC became obsolete.

    Quote:

    It wouldn't bother me in the least if we supported those.


    I bet it wouldn't, after all you are advocating a "future" based on PPC...

    Quote:

    And I have never favored supporting only the largest player in the industry, as lack of competition defeats innovation.


    Yeah, let us make the purpose of next generation MorphOS to even out the x86 market shares out of some ideological "fairness" reason, and MorphOS betting its life on the notorious number two will really make a difference and change the x86 market in this regard.

    Quote:

    Plus, when Intel made the mistake of betting on the Pentium 4, AMD products were kicking their ass.


    Pentium 4?!?? Really...?

    Quote:

    I don't anticipate Intel failing this way in the future, but without competition they WILL try to feed us similar crap.


    Yes, because feeding the world crap is how Intel became the single relevant player on the Desktop market...

    Quote:

    As to the comparable power of PPCs vs. X64 processors, they are not nearly as far apart as you seem to think.


    Yes they are. Not even in the same dimension.

    Quote:

    Even ARM is approaching a competitive stance with X64.


    No, not really, not in a desktop context.

    Quote:

    And, as you have pointed out, its all microcode based anyway.


    Oh my god...
    MorphOS is Amiga done right! :-)
    MorphOS NG will be AROS done right! :-)
  • »29.08.15 - 04:49
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > There was a point in time where you could say about AMD CPU's that "well hey,
    > at least it's cheap".

    There was a point in time when AMD CPUs were faster than Intel CPUs. Jim is correct here.
  • »29.08.15 - 11:59
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Andreas_Wolf wrote:
    >> the 68060 is completely hard-wired while the 68040 and prior use microcode.

    > I don't believe this to be true of any 68k processor, but we can check.

    "MC68060 [...] is all hardwired - there is no microcode in it."
    http://www.faqs.org/faqs/motorola/68k-chips-faq/

    From the Freescale website:
    "The Motorola 68K family consists of a wide range of members from the micro-coded MC68000 to the super-scalar, hard-wired MC68060."
    http://www.freescale.com/files/32bit/doc/reports_presentations/MC680X0OPTAPP.txt

    From Motorola's Jim Reinhart:
    "the 060 execution units are fully logic driven (i.e. no microcode)"
    http://www.verycomputer.com/154_1d1b5950dff465d3_1.htm#p4


    Thanks Andreas,
    That really confirms what I've heard over the years (that there conflicting views on this).
    My information came directly from Motorola engineering in the early '90s.
    So, as we were using the 68000, getting ready to use the '020, and the '030 was on the horizon (but the '060 was long off), I'd be willing to say the 68K family is entirely microcoded.

    BUT...here is the problem, there is no way to confirm anything.
    The structure of these chips is totally "black box".
    We have diagrams showing us the logical parts and we have the instruction set, programming modes, and other useful info but no real info on the actual logic.
    "Never attribute to malice what can more readily explained by incompetence"
  • »29.08.15 - 12:11
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > My information came directly from Motorola engineering in the early '90s.

    That was before the 68060, wasn't it?

    > I'd be willing to say the 68K family is entirely microcoded.

    I still think that's not true for the 68060.

    > here is the problem, there is no way to confirm anything.

    I have no reason to doubt the sources I quoted. Do you have any serious sources for the 68060 using microcode?
  • »29.08.15 - 13:37
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Andreas_Wolf wrote:
    > My information came directly from Motorola engineering in the early '90s.

    That was before the 68060, wasn't it?

    > I'd be willing to say the 68K family is entirely microcoded.

    I still think that's not true for the 68060.

    > here is the problem, there is no way to confirm anything.

    I have no reason to doubt the sources I quoted. Do you have any serious sources for the 68060 using microcode?


    No info I can rely on, but if the previous generations were, why change for the '060?
    And the '060's design really derives from the '060 (really a kind of super scalar version of the '40).

    And I'm trying not to even dignify takemehomegrandma's latest posts.
    Apparently he has not been paying attention to your reports on the latest 64 bit ARM cpus.

    And my about ten year old G5 beats some of the scores I get with one Core2 based system I have.
    Some PPCs certainly beat lower end X64 systems and are light years away from what Amigas started with.
    "Never attribute to malice what can more readily explained by incompetence"
  • »29.08.15 - 15:41
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > if the previous generations were, why change for the '060?

    I don't know, but I read that after the massively microcoded 68000, the following m68k generations used less and less microcode, so that the 68040 had relatively little microcode and the 68060 none at all. Maybe it was for performance reasons?

    > the '060's design really derives from the '060 (really a kind of super scalar
    > version of the '40).

    There are significant opcode differences between the two. And as you said, it's a black box, so ISA design of the 060 may be derived from the 040's, but this isn't necessarily true for the microarchitectural design.

    > Apparently he has not been paying attention to your reports on the latest 64 bit ARM cpus.

    I don't remember I said anything about performance. But there you go:

    http://www.anandtech.com/show/8357/exploring-the-low-end-and-micro-server-platforms/18
    http://www.cnx-software.com/2014/10/26/applied-micro-x-gene-64-bit-arm-vs-intel-xeon-64-bit-x86-performance-and-power-usage/
  • »29.08.15 - 20:07
    Profile
  • Order of the Butterfly
    Order of the Butterfly
    Posts: 165 from 2004/11/18
    Takmehomegrandma: i have never said that morphos does not support new radeon cards in the future ;-) what i say, it's that for me amd64+radeon cards is the best choice. It's not a troll.
    For me AMD processors are great, i use a ATHLON64 X2 since 2008 and it's allways suffisant today. I like also intel processors but the are a little more expansive.

    [ Edited by acepeg 29.08.2015 - 22:06 ]
  • »29.08.15 - 21:01
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Andreas_Wolf wrote:
    > if the previous generations were, why change for the '060?

    I don't know, but I read that after the massively microcoded 68000, the following m68k generations used less and less microcode, so that the 68040 had relatively little microcode and the 68060 none at all. Maybe it was for performance reasons?

    > the '060's design really derives from the '060 (really a kind of super scalar
    > version of the '40).

    There are significant opcode differences between the two. And as you said, it's a black box, so ISA design of the 060 may be derived from the 040's, but this isn't necessarily true for the microarchitectural design.

    > Apparently he has not been paying attention to your reports on the latest 64 bit ARM cpus.

    I don't remember I said anything about performance. But there you go:

    http://www.anandtech.com/show/8357/exploring-the-low-end-and-micro-server-platforms/18
    http://www.cnx-software.com/2014/10/26/applied-micro-x-gene-64-bit-arm-vs-intel-xeon-64-bit-x86-performance-and-power-usage/


    I think you may have a point on the '60 performance.
    It may also point to why Motorola eventually gave up on the series.
    It would make it much more work to create new cpus.
    "Never attribute to malice what can more readily explained by incompetence"
  • »29.08.15 - 22:31
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    "However, when it comes to performance-per-watt, APM X-Gene 1 is clearly ahead of Intel Xeon E5-2650 and there’s no comparison against Xeon Phi systems."

    OK, lower performance than X64, but...

    [ Edited by Jim 29.08.2015 - 17:35 ]
    "Never attribute to malice what can more readily explained by incompetence"
  • »29.08.15 - 22:34
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    Update:

    > Kryo from Qualcomm

    More details:
    https://www.qualcomm.com/news/snapdragon/2015/09/02/snapdragon-820-and-kryo-cpu-heterogeneous-computing-and-role-custom
  • »02.09.15 - 16:02
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    >> the latter idea may be tackled by another contender according to rumours:
    >> http://www.extremetech.com/computing/185888-vias-new-isaiah-x86arm-hybrid-cpu-outperforms-intel-in-benchmarks-but-will-it-ever-come-to-market

    > Neat approach.

    No release from VIA/Centaur so far, but an announcement of a MIPS core and chips with x86 and ARM compatibility from BLX/Loongson:

    http://blog.imgtec.com/mips-processors/loongson-mips64-processors-performance-barrier
    http://venturebeat.com/2015/09/03/chinas-loongson-makes-a-64-bit-mips-processor-that-can-run-x86-and-arm-code/
    http://www.cnx-software.com/2015/09/03/loongson-introduces-mips64-3a2000-3b2000-processors-based-on-gs464e-architecture/
    http://translate.google.com/translate?&sl=zh-CN&tl=en&u=http://www.loongson.cn/news_info.php?id=326
  • »07.09.15 - 14:17
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > http://en.wikipedia.org/w/index.php?title=Apple_A8 [...]
    > From AnandTech:
    > CPU: Enhanced Cyclone (i.e. "Cyclone gen 2") @ 1.4GHz, dual core

    Introduced yesterday:

    https://en.wikipedia.org/wiki/Apple_A9
    https://en.wikipedia.org/wiki/Apple_A9X

    Details are sparse so far. The core has been enhanced over Typhoon (aka "Cyclone gen 2") no doubt. Naming of the new core and performance parameters will trickle through eventually.


    Edit: A9 is dual-core at 1.85 GHz clock rate.
    Edit2: The new core is named Twister.

    http://www.idownloadblog.com/2015/09/21/phone-6s-customer-receives-early-benchmarks/
    http://www.anandtech.com/show/9662/iphone-6s-and-iphone-6s-plus-preliminary-results

    [ Edited by Andreas_Wolf 29.09.2015 - 10:04 ]
  • »10.09.15 - 13:27
    Profile
  • Order of the Butterfly
    Order of the Butterfly
    minator
    Posts: 365 from 2003/3/28
    Quote:

    Andreas_Wolf wrote:

    Edit: A9 is dual-core at 1.85 GHz clock rate.
    Edit2: The new core is named Twister.


    Apple are claiming it's faster than something like 80% of laptops sold in the last year.
    Bit of a bold claim but most laptops sold will be low end, on some benchmarks ARM have been hitting around low end laptop scores since the A15.

    Talking of which, this was an interesting comment:

    "2016 will be the year of the ARM laptop" - Linus Torvalds.

    http://news.softpedia.com/news/linus-torvalds-say-2016-will-be-the-year-of-the-arm-laptop-494044.shtml



    For servers, the APM X-Gene 1 isn't exactly the best poster child as it's built on a process 3 generations behind. The results should be much better on a more up to date process. Haven't heard much recently but X-Gene 2 was due soon with 3 fairly soon after.

    However, now the big boys are getting in:

    http://www.extremetech.com/computing/194701-qualcomm-will-enter-arm-server-market-with-major-partners-broad-solutions
  • »10.10.15 - 13:49
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > However, now the big boys are getting in:
    > http://www.extremetech.com/computing/194701-qualcomm-will-enter-arm-server-market-with-major-partners-broad-solutions

    That's a 2014 article. Wrong link?
  • »10.10.15 - 20:05
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12058 from 2003/5/22
    From: Germany
    > "K12" [...] designed by a core development team led by Chief CPU Architect​ Jim Keller.

    One month ago:

    http://www.anandtech.com/show/9643/jim-keller-leaves-amd
  • »18.10.15 - 11:00
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    minator wrote:

    "2016 will be the year of the ARM laptop" - Linus Torvalds.

    http://news.softpedia.com/news/linus-torvalds-say-2016-will-be-the-year-of-the-arm-laptop-494044.shtml




    Don't bet on it.

    Good thing he just focuses on a kernel based on UNIX (and doesn't really dabble in OS', other software, or hardware).

    At what point did we start worshiping the opinion of a creator of a plagiarized work?

    BTW - I'm not satisfied with Android or ChromeOS based ARM laptops.

    Come to think of it, I regularly consider ditching my Android based cell phone for an iPhone (but the idea of joining that zombie horde troubles me).
    "Never attribute to malice what can more readily explained by incompetence"
  • »08.11.15 - 15:15
    Profile
  • Moderator
    Kronos
    Posts: 2231 from 2003/2/24
    Quote:

    Jim wrote:

    At what point did we start worshiping the opinion of a creator of a plagiarized work?





    One of the few points in time when I wished I was on the red side.

    How I could ridicule Jim ...... how I could troll him with this statement for years to come.....

    ..... all the fun missed ......... *facepalm*
  • »08.11.15 - 16:18
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    ausPPC
    Posts: 543 from 2007/8/6
    From: Pending...
    Quote:

    Jim wrote:

    At what point did we start worshiping the opinion of a creator of a plagiarized work?




    SCO shareholder? Even if that was a reasonable description I wouldn't care because Linux gets the job done.
    PPC assembly ain't so bad... ;)
  • »08.11.15 - 20:47
    Profile Visit Website