Open Power
  • Order of the Butterfly
    Order of the Butterfly
    ppcamiga1
    Posts: 202 from 2015/8/23
    It is obvious that everything below amiga gui and graphics should be dropped.
    Replacing old api with new not compatible with today's standards is dumb idea.
    I still hope MOS Team will made something usefull not shit like AROS.
  • »02.05.19 - 09:58
    Profile
  • Paladin of the Pegasos
    Paladin of the Pegasos
    redrumloa
    Posts: 1398 from 2003/4/13
    Quote:

    ernsteiswuerfel wrote:
    Are you sure that UAE even overtook Amiga-m86k? Did you ever set full chipset emulation in WinUAE? :-D


    Yes, yes it has. It doesn't have the nostalgia feel of using an original keyboard and looking at a crt monitor with scanlines, but for functionality and speed it easily overtook Amiga-68k. Heck, I have an Amiga 500 with Vampire v2. It is great! It cannot run Payback as well or as fast as WinUAE. Even if you had an A4000T with CSPPC and Voodoo3 3000, it would not run Payback as fast as WinUAE.

    In WinUAE you can dial in any settings needed. You can literally run any software ever made for the Amiga. Real hardware is not as fortunate. Sideways and backwards compatibility was never an Amiga strong suit.

    Quote:

    And I think the G5 has already been topped by the X5000 in at least some aspects, even if it is still faster when Altivec is used.


    It has a lower power draw, that's about it. Even if it did come even with my G5 2.7 one day, it would still cost literally 25x times as much as I paid for my G5 with MorphOS license.
  • »02.05.19 - 18:01
    Profile
  • Moderator
    Kronos
    Posts: 1982 from 2003/2/24
    Quote:

    redrumloa wrote:


    It has a lower power draw, that's about it.


    Better (possible) GPUs are also a thing.

    At least as long as G5s are restricted to AGP.
    --------------------- May the 4th be with you ------------------
    Mother Russia dance of the Zar, don't you know how lucky you are
  • »02.05.19 - 18:53
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    >> I think the G5 has already been topped by the X5000 in at least
    >> some aspects, even if it is still faster when Altivec is used.

    > It has a lower power draw, that's about it.

    https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=11137&start=603

    With a 2.7 GHz G5 instead of my 2.3 GHz G5, the index of the G5 would increase from 1.00 to 1.17. This would result in the 2.7 GHz G5 being slower than the 2.0 GHz P5020 in 33% of cases (non-SIMD integer operations).
  • »04.05.19 - 02:00
    Profile
  • Order of the Butterfly
    Order of the Butterfly
    ernsteiswuerfel
    Posts: 370 from 2015/6/18
    From: Funeralopolis
    Quote:

    redrumloa schrieb:
    Yes, yes it has. It doesn't have the nostalgia feel of using an original keyboard and looking at a crt monitor with scanlines, but for functionality and speed it easily overtook Amiga-68k. Heck, I have an Amiga 500 with Vampire v2. It is great! It cannot run Payback as well or as fast as WinUAE. Even if you had an A4000T with CSPPC and Voodoo3 3000, it would not run Payback as fast as WinUAE.

    In WinUAE you can dial in any settings needed. You can literally run any software ever made for the Amiga. Real hardware is not as fortunate. Sideways and backwards compatibility was never an Amiga strong suit.

    Of course I do know and use WinUAE from time to time. Everybody does. ;-) My point is not much about nostalgia here, it's about WinUAE slowing down a lot if you use Full Chipset Emulation and some cycle exact timings and/or MMU emulation. The config imitating my A1200/030 sometimes is actually slower than the real A1200.

    But that's ok as I don't use it regularly. Software/Games which depend mostly on the CPU and/or use a RTG screen will undoubtedly get a great speed boost from WinUAE. That of course is a big plus for anybody using such software, but does not matter much for me, 'cause WinUAE can get slow in my use cases and I am too lazy to switch configs and reboot. ;-)
    Talos II. [Gentoo Linux] | PMac G5 11,2. PMac G4 3,6. PBook G4 5,8. [MorphOS 3.13 / Void Linux / Gentoo Linux] | A1200. ACA-1233, Indivision AGA Mk2. [Amiga OS 3.2]
  • »05.05.19 - 11:43
    Profile
  • Priest of the Order of the Butterfly
    Priest of the Order of the Butterfly
    KennyR
    Posts: 747 from 2003/3/4
    From: #AmigaZeux, Gu...
    WinUAE isn't perfect but it covers just about anything you can do on the Amiga that you can't do elsewhere. That's what it was designed for.

    Playing later Amiga games like Napalm, Payback (with caveats) Foundation or Genetic Species will run faster than any real Amiga system (won't comment on Vampire as I haven't seen it in action). And whdload works on basic 040+MMU emulation. You can play any Amiga game you like, from ADF for WHDLoad images: I don't know if there's a single one that is incompatible with WinUAE.

    Playing Doom or Quake on it is iffy, but you can play doom on Windows. Same with watching video or browsing.

    When phase5 decided to go the PPC route, there still was a genuine advantage in being able to mix old Amiga software with new, even doing things like using PPC-native shared libraries with old 68k apps which AROS for example could not do. Not to mention that you'd lose most of the interest if you decided to just launch a new non-backwards-compatible OS.

    Today, that need is long gone. Insisting on PPC is still chasing that late 90s Amiga that doesn't even work in this day and age. There's simply no need for it any more.

    [ Edited by KennyR 05.05.2019 - 21:54 ]
  • »05.05.19 - 21:50
    Profile
  • Order of the Butterfly
    Order of the Butterfly
    ppcamiga1
    Posts: 202 from 2015/8/23
    From developer POV ppc is interesting because it is not a pc.
    commodity hardware? No problem but first AROS/MOS/ or any other amiga like solution must be not a shit compared to Windows/macOS/Linux.
  • »05.05.19 - 22:30
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    Update:

    > "IBM is also contributing a newly developed softcore to the community.
    > In a very short time, an IBM engineer was able to develop a softcore on
    > the POWER ISA, and get it up and running on a Xilinx FPGA."

    https://github.com/antonblanchard/microwatt

    The PPC64LE Microwatt should be more complete than the unfinished PPC32 zCore Pippo at least.

    "Microwatt [...] does not currently implement any floating-point or vector instructions, hypervisor mode, big-endian mode, 32-bit mode, or the HPT/SLB MMU facilities."
    https://lists.ozlabs.org/pipermail/linuxppc-dev/2020-May/210525.html


    Edit: Chiselwatt by the same author, but written in Chisel
    Edit2: added quote

    [ Edited by Andreas_Wolf 12.05.2020 - 00:58 ]
  • »22.08.19 - 09:29
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    Amiga, AmigaOS and Hyperion mentioned during OpenPOWER Summit NA 2019:

    https://www.youtube.com/watch?v=bwXX4fqsPrQ

    timestamps:
    - 0:24
    - 2:02 (only on slide)
    - 25:07
    - 27:36

    Obviously, he means to say "A-Eon" when he mentions "Hyperion" who "manufacture PowerPC boards" and "ported AMDGPU to AmigaOS" :-)
  • »06.09.19 - 16:25
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    Addendum:

    > Power ISA 3.0b published in March 2017:
    > https://openpowerfoundation.org/?resource_lib=power-isa-version-3-0
    >
    > Significant quote from Power ISA 3.0 spec: [...]
    > Updated quote from Power ISA 3.0b spec: [...]
    >
    > Funny thing to say that Power ISA 3.0 "continues" or "focuses" the
    > integration of PowerPC ISA and Book E ISA, when in fact it essentially
    > *reverts* this integration by eliminating the parts that Book E ISA
    > had brought into it.

    At OpenPOWER Summit EU 2019, IBM director and OpenPOWER president Mendy Furmanek completely misrepresented the fate of the Book E ISA ("PowerPC Embedded") components within OpenPOWER Foundation's Power ISA 3.0+:

    https://www.youtube.com/watch?v=ufBtrGJVF6g (5:40-6:40)

    Not only does she unjustifiably give credit to the OpenPOWER Foundation for the integration of PowerPC ISA and Book E ISA when in fact it was Power.org that accomplished this 7 years earlier in 2006, she also fully ignores that what OpenPOWER actually did 4 years ago with the release of its Power ISA 3.0 was essentially reverting this integration by eliminating the former Book E ISA components.
  • »06.11.19 - 13:03
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    Update:

    > Significant quote from Power ISA 3.0 spec:
    > "In 2006, Freescale and IBM collaborated on the creation of the Power ISA Version
    > 2.03, which represented the reunification of the architecture by combining Book E
    > content with the more general purpose PowerPC Version 2.02. Power ISA Version
    > 3.0 continues this integration by providing a single Book III for all Power
    > implementations, and also by eliminating optional architecture categories to
    > ensure increased application portability between Power processors."
    >
    > Updated quote from Power ISA 3.0b spec:
    > "In 2006, Freescale and IBM collaborated on the creation of the Power ISA Version
    > 2.03, which represented the reunification of the architecture by combining Book E
    > content with the more general purpose PowerPC Version 2.02. The resulting
    > Architecture included environment-specific privileged architecture optimizations
    > (two Book IIIs) and optional application-specific facilities (categories) as
    > extensions to a pervasive base architecture. Power ISA Version 3.0 B focuses this
    > integration by choosing a single Book III and a set of widely used categories to
    > become part of the base architecture for all forward-looking Power
    > implementations. All other optional architecture categories have been eliminated
    > to ensure increased application portability between Power processors. Legacy
    > embedded applications that require the eliminated material will continue
    > to use V. 2.07B."
    >
    > Funny thing to say that Power ISA 3.0 "continues" or "focuses" the integration of
    > PowerPC ISA and Book E ISA, when in fact it essentially *reverts* this integration
    > by eliminating the parts that Book E ISA had brought into it.

    Slightly updated quote from brand-new Power ISA 3.1 (POWER10) spec:
    "In 2006, Freescale and IBM collaborated on the creation of the Power ISA Version 2.03, which represented the reunification of the architecture by combining Book E content with the more general purpose PowerPC Version 2.02. The resulting architecture included environment-specific privileged architecture optimizations (two Book IIIs) and optional application-specific facilities (categories) as extensions to a pervasive base architecture. In support of the OpenPOWER Foundation’s standardization of server architecture, Power ISA Version 3.0 streamlined this integration by choosing a single Book III and a set of widely used categories to become part of the base architecture for all forward-looking Power implementations. All other optional architecture categories were eliminated to ensure increased application portability between Power processors. Legacy embedded applications that require the eliminated material will continue to use V. 2.07B."

    I'm not sure that changing the wording from "focuses" to "streamlined" eliminates the misrepresentation of the facts.

    Also interesting from the 3.1 spec:
    "Power ISA Version 3.0C took the first step in reintroducing optionality into the architecture as the Power ISA moves to an “open” model governed by the OpenPOWER Foundation. Material later in the preface identifies compliancy subsets of the architecture and the optional features which they comprise. [...] Change bars have been included in the body of this document to indicate changes from the Power ISA Version 3.0C. [...] Version 3.1 of the Power ISA [...] is intended to supersede and replace version 3.0C. [...] This version was created by making miscellaneous corrections and by applying the following requests for change (RFCs) to Power ISA Version 3.0C. Change bars in this summary of changes indicate changes relative to v3.0C."

    Power ISA "3.0C"? Never heard of that. Does anybody have a link to that "3.0C" spec?
  • »11.05.20 - 20:19
    Profile
  • Yokemate of Keyboards
    Yokemate of Keyboards
    Andreas_Wolf
    Posts: 10977 from 2003/5/22
    From: Germany
    Addendum:

    > Interesting new project called Libre-SOC: [...]
    > Details: https://libre-soc.org/openpower/

    New information from the company behind Libre-SOC:

    https://systemeslibres.org
    https://systemeslibres.org/updates/just-incorporated/
    https://systemeslibres.org/updates/code_to_tapeout/

    Schedule:
    - 2020 "shuttle smoketest" tapeout of 180nm (TSMC) embedded-class single-core chip, OOO, pipelined
    - 2021 commercial single-core or quad-core (information are conflicting here) chip with GPU
    - 2022 commercial quad-core chip with GPU (5 GFLOPS), VPU, DDR SDRAM controller, SERDES, UART, LPC, JTAG, GPIO, 3 watts power consumption

    Sponsors:
    - current: NLNet, RaptorCS
    - previous: Purism

    Other interesting statements:
    "RaptorCS is taking a serious look at our future line up of fully POWER compliant 4-core chips."
    "Libre-SOC was originally targeting RISCV, but RaptorCS expressed interest in a POWER SOC, so we pivoted to IBM’s Open POWER ISA."

    I guess Raptor is considering this Power SoC to replace the current ARM/ColdFire-based ASPEED BMC, hence the requirement for LPC.
  • »29.05.20 - 09:55
    Profile