1. Forum
  2. Usenet
  3. COMP.ARCH

  • Re: IBM architectural goals, Byte Addressability And Beyond

    From John Levine@21:1/5 to All on Thu May 30 15:06:35 2024
    According to Michael S <already5chosen@yahoo.com>:
    Note that the feature was introduced in Znext (2012). That it is
    still there must indicate that it gets some usage.

    Not necessarily.
    After feature was given publicly documented opcode it's very hard to
    remove it.
    Naturally, I don't know if this particular feature got publicly
    documented opcode and don't know where too look.

    They never remove anything from the architecture. If you're asking
    about decimal FP, it's still there, both scalar and vector.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Michael S@21:1/5 to John Levine on Thu May 30 18:12:28 2024
    On Thu, 30 May 2024 15:06:35 -0000 (UTC)
    John Levine <johnl@taugh.com> wrote:

    According to Michael S <already5chosen@yahoo.com>:
    Note that the feature was introduced in Znext (2012). That it is
    still there must indicate that it gets some usage.

    Not necessarily.
    After feature was given publicly documented opcode it's very hard to
    remove it.
    Naturally, I don't know if this particular feature got publicly
    documented opcode and don't know where too look.

    They never remove anything from the architecture. If you're asking
    about decimal FP, it's still there, both scalar and vector.

    I was talking/asking about CU14. I never had doubts about decimal
    stuff.
    You did mentioned that CU14 is a public instruction in your other
    post. It means that it can't be removed even if later proven useless.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Levine@21:1/5 to All on Thu May 30 15:37:52 2024
    According to Michael S <already5chosen@yahoo.com>:
    They never remove anything from the architecture. If you're asking
    about decimal FP, it's still there, both scalar and vector.

    I was talking/asking about CU14. I never had doubts about decimal
    stuff.
    You did mentioned that CU14 is a public instruction in your other
    post. It means that it can't be removed even if later proven useless.

    They added two instructions to convert between UTF-8 and UTF-16 in S/390, then the other six among utf-8/16/32 in the initial zseries.

    The only thing IBM has ever removed from the instruction set visible
    to applications was the ASCII mode bit from S/360 to S/370, keeping in
    mind that what they called ASCII was not what anyone else called
    ASCII. Everything else is still there.

    The cost of trying to track down and patch all of the customer dusty
    decks that might use obscure instructions is high, while the cost of
    leaving them in the instruction set is low since it's all vertical
    microcode anyway.

    I've seen a presentation that said the original implementation of
    decimal FP was all in millicode, while later models added hardware.
    Evidently the wall street types really like the consistent decimal
    rounding.



    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Thomas Koenig@21:1/5 to John Levine on Thu May 30 15:57:51 2024
    John Levine <johnl@taugh.com> schrieb:
    According to Michael S <already5chosen@yahoo.com>:
    Note that the feature was introduced in Znext (2012). That it is
    still there must indicate that it gets some usage.

    Not necessarily.
    After feature was given publicly documented opcode it's very hard to
    remove it.
    Naturally, I don't know if this particular feature got publicly
    documented opcode and don't know where too look.

    They never remove anything from the architecture. If you're asking
    about decimal FP, it's still there, both scalar and vector.

    I don't think the vector instructions of the /390 are in there
    any more. But those were markeded as extensions, so maybe that
    doesn't count. (And they were very underwhelming, so...)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Levine@21:1/5 to All on Thu May 30 16:43:47 2024
    According to Thomas Koenig <tkoenig@netcologne.de>:
    They never remove anything from the architecture. If you're asking
    about decimal FP, it's still there, both scalar and vector.

    I don't think the vector instructions of the /390 are in there
    any more. But those were markeded as extensions, so maybe that
    doesn't count. (And they were very underwhelming, so...)

    You're right on both counts. The Vector Facility was an optional
    add-on that wasn't in the S/390 POO. Now z has a differently
    underwhelming vector facility. For packed decimal, even.

    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)