SMP/AMP, or other multiprocessing for MorphOS???
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12079 from 2003/5/22
    From: Germany
    > Has there been any new statements about the "Q-Box" being a future project of the
    > MorphOS Dev. Team? [...] I ask if any of the MorphOS Dev. Team members have made
    > an announcement, or even hinted at new work beginning on the "Q-Box". [...] are there
    > new statements or hints from any Dev. Team members that suggest a different impression
    > of what is happening, or what might happen (or be started) within the next year?

    No.

    > I thought that the last word from the Dev. Team [...] was that the "Q-Box" was dead and
    > had been abandoned

    The last word from a MorphOS Team member was that the QBox was never planned and also never mentioned as future prospect by the MorphOS Team (which is not true, of course).

    > I also thought that one or more of the Dev. Team members have stated that NO work has
    > been done on any architecture change, and that "IF" any change to a different architecture
    > would be attempted, it would be a long time (years) before it could be completed. Are my
    > impressions correct [...]?

    Yes.
  • »17.06.14 - 21:13
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    We can call it Q-box, and the concept is quite similar.
    But at this point we can consider multi-core SMP support and 64 bit addressing.
    "Never attribute to malice what can more readily explained by incompetence"
  • »17.06.14 - 23:15
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    Well, this has all been speculated about with mostly the same conclusions since I joined MZ 2 years ago. Without the input of the MOS Team it's quite pointless. After all, it's them who has to do all the work.

    I just hope they will announce their plans before people get tired of waiting. Used macs are, as we all already know, a nice stop gap compromise and they will do just fine for a few more years. And it can work even longer if people know what the next step will be, even if it takes years to materialize IMHO. Old macs works, but it's not sexy :-)

    However, like the AOS 4 laptop, making promises that may not materialize is even worse than saying nothing at all.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »17.06.14 - 23:59
    Profile Visit Website
  • Paladin of the Pegasos
    Paladin of the Pegasos
    redrumloa
    Posts: 1424 from 2003/4/13
    Quote:

    Yasu wrote:
    Well, this has all been speculated about with mostly the same conclusions since I joined MZ 2 years ago. Without the input of the MOS Team it's quite pointless. After all, it's them who has to do all the work.

    I just hope they will announce their plans before people get tired of waiting. Used macs are, as we all already know, a nice stop gap compromise and they will do just fine for a few more years. And it can work even longer if people know what the next step will be, even if it takes years to materialize IMHO. Old macs works, but it's not sexy :-)

    However, like the AOS 4 laptop, making promises that may not materialize is even worse than saying nothing at all.


    Come on man, don't compare MorphOS to other NG solutions that are practically abandonware!

    We end users are allowed to speculate, it is part of the fun as long as the users are grounded in reality. I'd love to see an ISA move and I'd love to see SMP in the near future, but if neither ever happened I'd be happy to just have had the opportunity to use MorphOS on PPC Macs like we have now.
  • »18.06.14 - 05:25
    Profile
  • Acolyte of the Butterfly
    Acolyte of the Butterfly
    KimmoK
    Posts: 102 from 2003/5/19
    Quote:

    an ISA move


    I love to see how ARIX manages to handle it and what comes.
    (so far, to me it seems AROS and Linux have had huge trouble without custom designed, 100% supported HW)

    In the meanwhile, with SMP capable 64bit PPC HW the "next dimension" can be prepared while users can boot pack to legacy (single core, 32bit) system when needed.

    UPDATE: nice graph that shows why SMP is very important:
    (x86&cinebench)

    [ Edited by KimmoK 01.07.2014 - 09:56 ]
    :-x :-P 8-)
  • »18.06.14 - 11:19
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    KimmoK wrote:
    Quote:

    an ISA move


    I love to see how ARIX manages to handle it and what comes.
    (so far, to me it seems AROS and Linux have had huge trouble without custom designed, 100% supported HW)

    In the meanwhile, with SMP capable 64bit PPC HW the "next dimension" can be prepared while users can boot pack to legacy (single core, 32bit) system when needed.


    ARIX uses a Linux kernal, and Linux based Amiga like OS' should have an advantage in implementing SMP.
    The core question is the same though, how do you get legacy Amiga apps to run in an SMP environment.
    "Never attribute to malice what can more readily explained by incompetence"
  • »18.06.14 - 12:19
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Zylesea
    Posts: 2053 from 2003/6/4
    I think the Hypervisor idea is a really neat one. While I am rather more in favor of making a clean cut and probably jump ISA. the hypervisor idea on PPC actually is neat. A multicore ppc where the processor itself warrants the environment for each OS instance. Two OS instances: MorphOS as we knwo it for one core, MorphOS NG on all the other cores . Both running alongside. The actual silicon taking care that everything will be fine.
    The good thing: MorphOS as we know it wouldl stay as it is (and stay compatible). MorphOS NG can evolve.
    Sure, no easy thing. And to be nicely useable both OS instances should be glued togetther somehow. But that approach would offer a very smooth transition. So to say a clear cut (MorphOS NG) without the clear cut (MorphOS stays as it is one _one_ core).
    The big, big, big disadvantage (aside from the required work): it would need to stay on PowerPC where hardware supply of current (QorIQ based) "normal" general purpose computers is - well - problematic (but at least there are the Freescale dev boards). But the idea to use hypervisor to run two instances simultaneously is brilliant to provide a smooth upgrade path for the OS.
    --
    http://via.bckrs.de

    Whenever you're sad just remember the world is 4.543 billion years old and you somehow managed to exist at the same time as David Bowie.
    ...and Matthias , my friend - RIP
  • »18.06.14 - 16:43
    Profile Visit Website
  • Order of the Butterfly
    Order of the Butterfly
    In_Correct
    Posts: 245 from 2012/10/14
    From: DFW, TX, USA
    This SMP or ASMP, Hypervisor, Q Box, etc. Improvements certainly would make MorphOS shine yet again. :-)

    The following is not a good enough reason to add these feature(s) but I still want to share my experience the last two and a half days:

    I have hundreds and hundreds of pages on my Watchlist from Wikipedia. There is a Power Architecture page which includes information on Genesi computers, AmigaOne computers, Hyperion, and the three AmigaOS 4, AROS, and MorphOS.

    An IP Address had removed every mentioning of Hyperion, the OSs, and AmigaOne claiming it to be not encyclopedic. The edit war continued until earlier this morning.

    All of the IP Address reasons to remove the information were unacceptable except for one:

    None of the Amiga-like OS fully supports Power Architecture. MorphOS usually does, except for the systems which contain more than one core.

    Are there many (and current) Power Architecture systems containing more than one processor core? The only ones I am familiar with is AmigaOne series. (Which kept getting removed from article).

    It does not matter because the edit war ended with a cease fire and all Amiga clones, and Operating systems, and related companies are still in the article.
    :-) I Support Quark Microkernel. :-D
  • »29.06.14 - 05:07
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12079 from 2003/5/22
    From: Germany
    > None of the Amiga-like OS fully supports Power Architecture. MorphOS
    > usually does, except for the systems which contain more than one core.

    Power Architecture is an ISA, not a chip or system. Therefore it makes no sense claiming an OS doesn't fully support Power Architecture just because it is a single-core OS. After all, the ISA allows for single-core implementations, of course. A single-core OS can't fully support a multi-core chip or system, but the ISA is per core.

    > Are there many (and current) Power Architecture systems containing more
    > than one processor core?

    Yes, there are hundreds of systems using more than one core that complies with Power Architecture currently on the market from a broad range of vendors. They are mostly specialized systems for embedded, networking, telco, industrial, imaging, storage, military etc. use.
  • »29.06.14 - 16:01
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12079 from 2003/5/22
    From: Germany
    > did he give any hints about what kind of multicore support will be used in AmigaOS4.2?

    Hyperionmp mentions SMP as goal for OS4.2:

    http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=35072&forum=32&start=340#738096
  • »28.07.14 - 22:23
    Profile
  • Moderator
    Kronos
    Posts: 2239 from 2003/2/24
    Quote:

    Andreas_Wolf wrote:
    > did he give any hints about what kind of multicore support will be used in AmigaOS4.2?

    Hyperionmp mentions SMP as goal for OS4.2:

    http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=35072&forum=32&start=340#738096


    Which means SMP can be ruled out with 100% certainty.

    [ Edited by Kronos 29.07.2014 - 03:03 ]
  • »29.07.14 - 05:02
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    @Kronos

    It does? I found Hyperions answer rather cryptic.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »29.07.14 - 16:42
    Profile Visit Website
  • Paladin of the Pegasos
    Paladin of the Pegasos
    redrumloa
    Posts: 1424 from 2003/4/13
    Quote:

    Yasu wrote:
    @Kronos

    It does? I found Hyperions answer rather cryptic.


    Of course they are cryptic, that is the whole idea. Something to do with Plausible deniability.
  • »29.07.14 - 19:20
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Yasu
    Posts: 1724 from 2012/3/22
    From: Stockholm, Sweden
    @redrumloa

    Maybe, but I don't think it qualifies as "100% certainty". Plausable maybe (what do I know?) but not definite.
    AMIGA FORUM - Hela Sveriges Amigatidning!
    AMIGA FORUM - Sweden's Amiga Magazine!

    My MorphOS blog
  • »29.07.14 - 21:25
    Profile Visit Website
  • Acolyte of the Butterfly
    Acolyte of the Butterfly
    KimmoK
    Posts: 102 from 2003/5/19
    from Hyperion MP:
    //Please bear in mind however that some of the goals (SMP e.g.) were deemed "impossible" to achieve barring sandboxing the current API, at least by a quite vocal group.
    Since we realize that many of you are getting somewhat impatient, there will be a "nice surprise" before the end of the year.
    (We realise that "nice" is not the work some would use but there you go).//

    To me it seems they have managed to somehow the stretch the (AOS4) API so that SMP like threading has become possible. (or perhaps they use HW assisted virtualization as part of the solution)

    It will be interesting to see. (and to see if it's "SMP" truly usefull for normal users)

    (there definitely is a lot of ways to get SMP... perhaps some will nag if lecacy code is left to run only on core0, but so is life...)

    [ Edited by KimmoK 30.07.2014 - 14:14 ]
    :-x :-P 8-)
  • »30.07.14 - 15:11
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    BSzili
    Posts: 559 from 2012/6/8
    From: Hungary
    Seeing is believing :)
    This is just like television, only you can see much further.
  • »30.07.14 - 15:22
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12079 from 2003/5/22
    From: Germany
    > To me it seems they have managed to somehow the stretch the (AOS4) API so
    > that SMP like threading has become possible.

    More information:

    http://blog.hyperion-entertainment.biz/?p=1184
  • »26.02.15 - 08:59
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    Cego
    Posts: 712 from 2006/5/28
    From: Germany
    well you just can't expect some solid work here. everything that came from hyperion was buggy and unstable even in their old days when they ported games...
    Pegasos II G4 @1.0GHz, 1GB DDR Ram, Radeon 9200Pro, 240GB SSD+160GB HD, MorphOS 3.18, AmigaOS4.1 FE, Debian 8
  • »26.02.15 - 09:54
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Edited - Sorry that sounded snide.

    It does look like Thomas is facing the same realities we've been discussing.

    [ Edited by Jim 28.02.2015 - 17:59 ]
    "Never attribute to malice what can more readily explained by incompetence"
  • »28.02.15 - 22:09
    Profile
  • Moderator
    Kronos
    Posts: 2239 from 2003/2/24
    Quote:

    "they assume SMP means everything running on all cores as soon as AROS starts.


    The S in SMP stands for "symetric" meaning all cores are treated equally (and therefore all tasks can be scheduled on every core at any given time).

    If it ain't doing that for apps compiled against the Amiga-API, it ain't SMP.



    Still a step in the right direction, and IMO the best possible way to may use of multicore/CPU systems within a reasonable timeframe.

    [ Edited by Kronos 01.03.2017 - 15:18 ]
  • »01.03.17 - 15:16
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Intuition
    Posts: 1110 from 2013/5/24
    From: Nederland
    Quote:

    Kronos wrote:
    Quote:

    "they assume SMP means everything running on all cores as soon as AROS starts.


    The S in SMP stands for "symetric" meaning all cores are treated equally (and therefore all tasks can be scheduled on every core at any given time).

    If it ain't doing that for apps compiled against the Amiga-API, it ain't SMP.



    Still a step in the right direction, and IMO the best possible way to may use of multicore/CPU systems within a reasonable timeframe.


    If i understand what he is trying to achieve then it's SMP for stuff using a new API and single core for legacy code.
    1.67GHz 15" PowerBook G4, 1GB RAM, 128MB Radeon 9700M Pro, 64GB SSD, MorphOS 3.15

    2.7GHz DP G5, 4GB RAM, 512MB Radeon X1950 Pro, 500GB SSHD, MorphOS 3.9
  • »01.03.17 - 15:52
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Intuition wrote:
    Quote:

    Kronos wrote:
    Quote:

    "they assume SMP means everything running on all cores as soon as AROS starts.


    The S in SMP stands for "symetric" meaning all cores are treated equally (and therefore all tasks can be scheduled on every core at any given time).

    If it ain't doing that for apps compiled against the Amiga-API, it ain't SMP.



    Still a step in the right direction, and IMO the best possible way to may use of multicore/CPU systems within a reasonable timeframe.


    If i understand what he is trying to achieve then it's SMP for stuff using a new API and single core for legacy code.



    Which Kronos has accurately pointed out is an kind of ASMP/SMP hybrid.
    And, as I mentioned to Andreas, is something we have been discussing for awhile now ourselves.

    So...AROS implements it first, and THEY get to take credit? ;-)
    Sheesh!
    "Never attribute to malice what can more readily explained by incompetence"
  • »07.03.17 - 13:56
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Intuition
    Posts: 1110 from 2013/5/24
    From: Nederland
    Quote:

    Jim wrote:
    Quote:

    Intuition wrote:
    Quote:

    Kronos wrote:
    Quote:

    "they assume SMP means everything running on all cores as soon as AROS starts.


    The S in SMP stands for "symetric" meaning all cores are treated equally (and therefore all tasks can be scheduled on every core at any given time).

    If it ain't doing that for apps compiled against the Amiga-API, it ain't SMP.



    Still a step in the right direction, and IMO the best possible way to may use of multicore/CPU systems within a reasonable timeframe.


    If i understand what he is trying to achieve then it's SMP for stuff using a new API and single core for legacy code.



    Which Kronos has accurately pointed out is an kind of ASMP/SMP hybrid.
    And, as I mentioned to Andreas, is something we have been discussing for awhile now ourselves.

    So...AROS implements it first, and THEY get to take credit? ;-)
    Sheesh!


    A bit like the US getting credit for the British Jet Engine, Television and Silicon Chip. ;)
    1.67GHz 15" PowerBook G4, 1GB RAM, 128MB Radeon 9700M Pro, 64GB SSD, MorphOS 3.15

    2.7GHz DP G5, 4GB RAM, 512MB Radeon X1950 Pro, 500GB SSHD, MorphOS 3.9
  • »08.03.17 - 10:42
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Intuition wrote:
    Quote:

    Jim wrote:
    Quote:

    Intuition wrote:
    Quote:

    Kronos wrote:
    Quote:

    "they assume SMP means everything running on all cores as soon as AROS starts.


    The S in SMP stands for "symetric" meaning all cores are treated equally (and therefore all tasks can be scheduled on every core at any given time).

    If it ain't doing that for apps compiled against the Amiga-API, it ain't SMP.



    Still a step in the right direction, and IMO the best possible way to may use of multicore/CPU systems within a reasonable timeframe.


    If i understand what he is trying to achieve then it's SMP for stuff using a new API and single core for legacy code.



    Which Kronos has accurately pointed out is an kind of ASMP/SMP hybrid.
    And, as I mentioned to Andreas, is something we have been discussing for awhile now ourselves.

    So...AROS implements it first, and THEY get to take credit? ;-)
    Sheesh!


    A bit like the US getting credit for the British Jet Engine, Television and Silicon Chip. ;)



    Well, lets be fair, the British got really dangerous nuclear power plant technology from the US. ;-)
    And the jet engine got its real start in Germany, not Britain.
    We all strip mined German technology after WWII (an old tradition, that was well underway after WWI, AND may have helped lead the way to the second war).
    "Never attribute to malice what can more readily explained by incompetence"
  • »08.03.17 - 13:59
    Profile