MorphOS on AmigaOne X5000?
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Andreas_Wolf wrote:
    > Can the X5000 (single core) decode and display 1080p x.264 streams in CPU/SW?

    "The second case is FullHD movies. Practically it's impossible to play them without loss of FPS. [...] it's a little bit funny that we can't play FullHD videos."
    http://www.amigapodcast.com/2017/04/amigaone-x5000-first-impression.html


    That may be an issue limited to OS4.
    Who knows?

    We will have to see when they get around to releasing MorphOS for the X5000.

    Edit - Actually, after thinking about it, this will be problematic.
    We simply shouldn't be decoding these streams with only the cpu.

    [ Edited by Jim 24.04.2017 - 04:06 ]
    "Never attribute to malice what can more readily explained by incompetence"
  • »23.04.17 - 23:10
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    takemehomegrandma
    Posts: 2720 from 2003/2/24
    Quote:

    Jim wrote:
    Quote:

    Andreas_Wolf wrote:
    > Can the X5000 (single core) decode and display 1080p x.264 streams in CPU/SW?

    "The second case is FullHD movies. Practically it's impossible to play them without loss of FPS. [...] it's a little bit funny that we can't play FullHD videos."
    http://www.amigapodcast.com/2017/04/amigaone-x5000-first-impression.html


    That may be an issue limited to OS4.
    Who knows?


    Only to a certain degree. The CPU lacks the grunt power, there is no way around that. The X1000 moved in 2005 G4 Powerbook realms concerning performance, and the G4 even won several benchmarks, that's well established. And the X5000 turned out to be a sideway move from the X1000 rather than a forward move.

    So the above comment makes sense and is perfectly understandable.

    Quote:

    Edit - Actually, after thinking about it, this will be problematic.
    We simply shouldn't be decoding these streams with only the cpu.


    That's a different question altogether. The issue here wasn't how movie streams are potentially decoded the best way, but whether the CPU is powerful enough to do it.

    This reminds me of the late Commodore days and thereafter, where the rest of the world (who already moved to powerful HW) began using mp3's and stuff. The Amiga was hopelessly behind. It lacked grunt power. But some enthusiasts persisted: "Untrue! The Amiga can of course also play mp3's! All you need is a HW decoder!" :lol:

    Maybe *this* is the true "Amiga Curse"? The constant lack of powerful HW?

    ;-)
    MorphOS is Amiga done right! :-)
    MorphOS NG will be AROS done right! :-)
  • »24.04.17 - 09:43
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    > That may be an issue limited to OS4. Who knows? We will have
    > to see when they get around to releasing MorphOS for the X5000.

    Or pampers, who we know is very willing to do whatever tests we demand of him on his MorphOS-driven X5000, could report here :-)
  • »24.04.17 - 13:38
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    OK...I PM'd Wiktor.
    Wish I had retained a direct email address, but its been a couple of years since I shipped him anything, and I no longer have access to the account I was using back then.

    The P5020 ought to be right at the borderline for this kind of application.

    I'd hate to find out that the X5000 can't do something a G5 can.

    Although, it does sound like a potential application for ASMP, IF we have to keep discussing cpu driven decoding (which IS damned primitive).

    [ Edited by Jim 24.04.2017 - 16:40 ]
    "Never attribute to malice what can more readily explained by incompetence"
  • »24.04.17 - 20:35
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    @Andreas

    Thanks, email inquiry sent.
    And I saved the email address.
    I lost quite a few of those when I suddenly lost access to my Comcast e-mail account.
    "Never attribute to malice what can more readily explained by incompetence"
  • »24.04.17 - 21:54
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    takemehomegrandma
    Posts: 2720 from 2003/2/24
    Quote:

    Jim wrote:

    I'd hate to find out that the X5000 can't do something a G5 can.


    What do you mean "find out"?!? This has been a known fact for how long now, a year? Two? Since its specs were first announced?

    The X1000, X5000 and year 2005 Powerbook G4's are all in the same performance realm, roughly speaking. And that means about 720p x.264 streams (depending on clip/encoding). Back at the time, many were surprised by the poor performance of the PA6T compared to G4. In several benchmarks, the G4 won, despite the clock difference. But at least the PA6T had Altivec, like the G4. The only reason to why the X5000 cope with the other two, despite not having Altivec (which really helps applications like these) is probably because the higher clock frequency compensates the lack of it. But still, it's only 2GHz. The G5 however goes to 2.7GHz *and* it has Altivec. You see? On G5's, 1080p x.264 streams play at 2.3GHz I think I recall. At 2.5GHz for sure.

    How is it possible that soeone like you, who posts so much about the splendidness of the X5000 all the time everywhere you go, didn't know about the X5000/G5 situation? It's not exactly news, you know? Do you ever listen to what others say?

    Quote:

    Although, it does sound like a potential application for ASMP, IF we have to keep discussing cpu driven decoding (which IS damned primitive).


    Sure, you could send the undecoded binary stream to a GPU and let it decode and show the film, or maybe to an Apple TV via AirPlay and hav it do exactly the same thing. Then you could say "Look, my X5000 can play 1080p streams." :lol: But the discussion here was about CPU performance, and whether the CPU was up for a task like this, or not. It turned out it wasn't. But G5's are! :-)
    MorphOS is Amiga done right! :-)
    MorphOS NG will be AROS done right! :-)
  • »24.04.17 - 22:42
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    >> it does sound like a potential application for ASMP, IF we have to
    >> keep discussing cpu driven decoding (which IS damned primitive).

    > the discussion here was about CPU performance, and whether the CPU
    > was up for a task like this, or not.

    I guess Jim's mention of ASMP refers to involving the 2nd CPU core.
  • »24.04.17 - 23:01
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    Quote:

    Andreas_Wolf wrote:
    >> it does sound like a potential application for ASMP, IF we have to
    >> keep discussing cpu driven decoding (which IS damned primitive).

    > the discussion here was about CPU performance, and whether the CPU
    > was up for a task like this, or not.

    I guess Jim's mention of ASMP refers to involving the 2nd CPU core.


    Thanks Andreas, at least you got that idea.
    As to the P5020 being SO much worse than the G4, I really don't think that is the case, or I wouldn't be planning on buying one.
    It has a clock speed advantage, and from the dmips figures posted should perform fairly well.
    As to AltiVec, that's only a small group of routines, and the cpu still has a standard fpu.
    Its definitely a better performer than the PA6T, and with the P5040's four cores it has the potential (via SMP or ASMP) of outperforming a G4 laptop by about 300%.

    Further, the X5000 supports MODERN gpus. The best that will ever plug into an AGP G5 is a Radeon 3850 or 4650/70 (IF Bigfoot continues to support AGP variants of PCIe cards).

    I'd like to see us one day have support for AMD GCN gpus.

    And we need PCIe if we are going to support USB3 eventually, as PCI doesn't have the needed bandwidth.

    Now, IF we could get support for the later PCIe G5s (like the Quad 2.5 GHz)...
    Eh, I keep one around, just in case, but I'm not placing bets on that.

    But we WILL have support for the X5000, and that DOES have advantages over our current hardware.

    And, btw, I don't like the idea of wasting a huge part of my cpu power on video decoding when it could be done in alternative ways (even with a G5).
    We ought to be able to display these streams on G4s.
    Not being able to do so on processors that run as fast as 1.67 GHz doesn't make sense when, if these tasks were offloaded, the cpu would still have plenty of power to spare.
    "Never attribute to malice what can more readily explained by incompetence"
  • »24.04.17 - 23:33
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    >> X5000 and year 2005 Powerbook G4's are all in the same performance realm

    > As to the P5020 being SO much worse than the G4, I really don't think that is the case

    I don't think "same performance realm" means "much worse".

    > It has a clock speed advantage

    Yes, that's what takemehomegrandma says makes up for its lack of AltiVec.

    > AltiVec [is] only a small group of routines

    Yes, but definitely of significant advantage for tasks like video decoding.

    > I don't like the idea of wasting a huge part of my cpu power on video decoding

    I'd love to waste the power of my 2nd CPU on video decoding. Currently, I'm wasting it on doing nothing at all ;-)
  • »25.04.17 - 00:36
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    >I'd love to waste the power of my 2nd CPU on video decoding. Currently, I'm wasting it on doing nothing at all ;-)

    That's not what I meant and you know it. ;-)

    But we do have many systems now with an extra cpu to spare (even my G4 system has one completely unused), so why not?
    If we are not implementing SMP, they are just sitting there.

    Oh, another advantage of the P50X0 cpu?
    DDR3 and much higher memory bandwidth, that will add to the performance advantage over G4s and help close the gap with G5s (the best G5s only offer DDR2).

    And its a pity we may never see a PPC implementation designed for us with a cpu that supports DDR4.
    "Never attribute to malice what can more readily explained by incompetence"
  • »25.04.17 - 02:05
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Zylesea
    Posts: 2057 from 2003/6/4
    Quote:

    Andreas_Wolf schrieb:

    > AltiVec [is] only a small group of routines

    Yes, but definitely of significant advantage for tasks like video decoding.



    IIRC MPlaywer with Altivec gains about 15% speed over the non Altivec version.

    Btw.: A shame the 864x Pegasos" didn't made it: Not the major bottle neck of existing G4 systems (slow bus), but with low wattage and optionally with two cores. An 8610 (single core, 1.3GHz, NEC "redtail" (IIRC) board) was even shown to decode FullHD. Chance missed and gone many years ago. And since these chances are long gone I massively advocate since a couple of years the ISA switch to x64. PPC is dead, I cried some tears for it, but eventually you need to look forward then. MorphOS on i7 or Ryzen.. stop it, i am wetting my pants!
    --
    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
  • »25.04.17 - 08:16
    Profile Visit Website
  • Yokemate of Keyboards
    Yokemate of Keyboards
    takemehomegrandma
    Posts: 2720 from 2003/2/24
    Quote:

    Jim wrote:

    As to AltiVec, that's only a small group of routines


    On MorphOS it's used more than you perhaps may think, but for stuff being discussed here (video encoding/decoding, rendering, etc) it is a big and important performance booster.


    Quote:

    and the cpu still has a standard fpu.

    Oh, another advantage of the P50X0 cpu?
    DDR3 and much higher memory bandwidth, that will add to the performance advantage over G4s and help close the gap with G5s (the best G5s only offer DDR2).


    If I interpret Andreas Wolf's benchmarks above, it says that the X5000 has a slight upper hand for most (non-SIMD) integer operations. For floating point and memory speed, it seems to be the other way round. And SIMD, Floating Point and memory are kind of essential for many day to day things, like codecs encoding/decoding and other "multimedia" situations, not to mention rendering, image processing, etc.

    And then the G5 has the benefit of much higher clock frequencies on top of it all, so...

    You may of course like/want an X5000 for several reasons, and you may even prefer it over a G5 system for some reasons as well. But I don't see how the X5000 wins over G5's in performance, I don't see how it would be a "G5 killer".

    And then you factor in the price in the equation...

    Looking at it through 2017-glasses, I'd say the X5000 is about on par with all other PPC stuff from more than a decade ago. Who cares what is slightly faster or slower in one or another area of measurement, when the situation in total is so off the charts behind? The rest of the world moved several dimensions ahead since these performance levels at all were relevant in desktop computing.

    Here we are now, looking at how G4's and X1000 and X5000 can decode 720p x.264 streams, and the G5 can do it in 1080p. While the rest of the world is at 4k and x.265. My phone films in 4k. All films of my children, family activities etc. You know. It's everyday stuff now, not science fiction. My next monitor will be 4k. 720p isn't really funny then.

    Of course there are better/smarter ways to decode video streams, by letting the GPU offload the CPU for example. But the need for CPU grunt power doesn't go away. It's still needed for a lot of other things in everyday computing. PPC is hopelessly left dimensions behind.

    In my eyes, the move to x64 is urgent.
    MorphOS is Amiga done right! :-)
    MorphOS NG will be AROS done right! :-)
  • »25.04.17 - 09:34
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    I've never disagreed with either of you two on that point, and the fact is WE are apparently moving to X64.

    When and how long it will take to complete is anyone's guess, but that journey has started, so we are one up on the competition.

    And we have, or will have, systems that represent about the best that we can get with this ISA.
    Two things I'd still like to see happen, the T2080 laptop and a port of MorphOS to PCIe PowerMac G5s.

    One thing I see NO point in what so ever is Tabor. Why would I want to invest any money in that much of a downgrade?

    And yes, the cpu in the X5000 isn't anything but a lateral move, at least in terms of power. Its really the other elements of the e5500 and e6500 cored cpus Soc that drive my interest. PCIe, higher bandwidth memory, (somewhat) faster SATA, etc.

    As to video decoding, if we aren't going to adopt gpu assisted decoding (which would vary somewhat with each gpu), there are decoding ICs available for both PCI and PCIe applications that drivers could be written for.

    I just don't see the point in using 70-100% of our available cpu power on one task.
    Not when the demands on a Linux, Windows, or MacOS system are so much lower.
    "Never attribute to malice what can more readily explained by incompetence"
  • »25.04.17 - 10:28
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    >>> another advantage of the P50X0 cpu? DDR3 and much higher memory bandwidth,
    >>> that will [...] help close the gap with G5s (the best G5s only offer DDR2).

    >> If I interpret Andreas Wolf's benchmarks above, it says that the X5000 has
    >> a slight upper hand for most (non-SIMD) integer operations. For floating
    >> point and memory speed, it seems to be the other way round.

    > Its really the other elements of the e5500 and e6500 cored cpus Soc that drive
    > my interest. PCIe, higher bandwidth memory, (somewhat) faster SATA, etc.

    Regarding memory performance, my G5 clearly destroys the X5000 as can be seen from the benchmarks in this thread, despite what's written on spec papers.

    > there are decoding ICs available for both PCI and PCIe applications that drivers
    > could be written for.

    Indeed, but I'd still favour an ASMP solution to keep my 2nd CPU from feeling bored. After all, it is unremovably sitting there constantly drawing current and creating heat. I'd consider using a dedicated expansion card for video decoding (or transcoding) an option as soon as MorphOS supports switching off (or power saving modes for) the 2nd G5 CPU.

    > I just don't see the point in using 70-100% of our available cpu power on one task.

    I don't have a problem with that, especially if it's spare CPU power anyway.
  • »25.04.17 - 12:48
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    koszer
    Posts: 1250 from 2004/2/8
    From: Poland
    Quote:

    Andreas_Wolf wrote:
    Indeed, but I'd still favour an ASMP solution to keep my 2nd CPU from feeling bored. After all, it is unremovably sitting there constantly drawing current and creating heat.


    Well, they say that you can safely pull it out (as long as CPU in socket A stays in place)...
  • »25.04.17 - 13:34
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    >>> Leaman Computing Ltd. ("AmigaKit") is not A-Eon Technology Ltd.

    >> "In order to bundle forces, A-EON Technology considers acquiring AmigaKit."
    >> http://translate.google.com/translate?sl=de&tl=en&u=http://www.amiga-news.de/de/news/AN-2017-01-00060-DE.html

    > you [...] liked to claim they were 2 completely unrelated companies ;-)

    Your recollection is flawed. As even your quote shows, what I've been claiming is that Leaman Computing Ltd. ("AmigaKit") and A-Eon Technology Ltd. are two separate companies, which is still true as far as I know and will remain true until one has acquired the other. As to relatedness of the companies, I've never grown tired of making clear that Matthew Leaman is the sole managing director of both companies, and that he is also the main shareholder (50%) of A-Eon and the sole owner of AmigaKit.
    The concept that two companies owned and/or directed by the same person(s) doesn't mean they're not separate companies shouldn't be unknown to you.

    Btw, as recently as 5 days ago Leaman has incorporated a new company called "Amiga Technology Ltd". He is listed as the owner and sole managing director:

    https://www.companysearchesmadesimple.com/company/uk/10732120/amiga-technology-ltd/

    Maybe this is supposed to be the new company merging A-Eon and Leaman Computing?
  • »25.04.17 - 14:13
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 12179 from 2003/5/22
    From: Germany
    >> my 2nd CPU [...] is unremovably sitting there

    > they say that you can safely pull it out (as long as CPU in socket A stays in place)

    Last I remember reading was that it was possible but that afterwards the system fans will run 100% all the time. Or would fan recalibration help in this case?
  • »25.04.17 - 14:41
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    Intuition
    Posts: 1110 from 2013/5/24
    From: Nederland
    Quote:

    Andreas_Wolf wrote:
    >> my 2nd CPU [...] is unremovably sitting there

    > they say that you can safely pull it out (as long as CPU in socket A stays in place)

    Last I remember reading was that it was possible but that afterwards the system fans will run 100% all the time. Or would fan recalibration help in this case?


    setenv boot-args cpus=1
    reset-all

    From open firmware doesn't help?

    Edit: Answering myself from 4 years ago.

    https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=9353&post_id=102605&viewmode=flat&sortorder=1&showonepost=1

    I think dementia might be setting in early lol

    [ Edited by Intuition 25.04.2017 - 17:58 ]
    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
  • »25.04.17 - 16:49
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    >I don't have a problem with that, especially if it's spare CPU power anyway.

    ;-) Me either, IF its the spare cpu.
    If its the primary cpu, its f'ing irritating. ;-)

    Besides, those discreet ICs with their fixed algorithms lack flexibility and programability.

    Oh, and thanks for the word on memory bandwidth, although that does make me want a PCIe G5 even more, since it ought to offer even higher memory bandwidth than our current systems.

    You guys keep talking, and you'll talk me out of the X5000.
    Which would be a shame as we still don't have anything, outside of the SAM460, that will take newer video cards.

    Then again, its a future collectors item, so...
    "Never attribute to malice what can more readily explained by incompetence"
  • »25.04.17 - 21:02
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    amigadave
    Posts: 2795 from 2006/3/21
    From: Northern Calif...
    Quote:

    Jim wrote:
    Then again, its a future collectors item, so...


    That was part of my thinking regarding my purchase of the X1000, but I was also incorrect in thinking that the X1000 would possibly/probably be the last PPC AmigaOne computer ever designed and manufactured (when I was first thinking about my X1000 purchase. Things changed by the time I actually took delivery of my X1000, and Trevor had informed me about the future X5000 plans, even suggesting to me that I wait, and get the X5000 instead). The X1000 still has the distinction of being the only (to my knowledge) desktop computer produced for sale to the general public which uses the infamous PA6T CPU, which is not outstanding in its performance when compared to much older G4 and G5 CPUs, but it is a fairly low power usage PPC CPU, and has a few nice features. The X1000 was just about a decade or so too late in being manufactured, to make it interesting to anyone else in the computer world.
    MorphOS - The best Next Gen Amiga choice.
  • »26.04.17 - 01:23
    Profile
  • Jim
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Jim
    Posts: 4977 from 2009/1/28
    From: Delaware, USA
    A quick update:

    I traded some messages with Wiktor today, but he said he has really been too busy to get much use out of of the X5000, and he was thinking of selling it.
    He has not tried too display HD video with it.

    I asked him to give it a shot, if he had the time.

    Maybe we should have just asked for further PPC Mac support...

    I'm still planning on buying the X5000, but part of that will primarily be to own a piece of history.
    A 64bit Qorlq based AmigaOne, that runs both PPC NG Amiga operating systems. Now that Freescale is a rapidly fading memory...
    "Never attribute to malice what can more readily explained by incompetence"
  • »26.04.17 - 16:27
    Profile