Re: [UNSURE] Re: [UNSURE] Re: [UNSURE] Re: Hardware TPU? - VMS

This is a discussion on Re: [UNSURE] Re: [UNSURE] Re: [UNSURE] Re: Hardware TPU? - VMS ; Bill Gunshannon wrote: > In article , > "Stuart Brook" writes: > >>On 14 Mar 2007 at 10:26, Zane H. Healy wrote: >> >> >>>>I've read somewhere that this was a microcode option for PDP11/23-11/24 >>>>machines, and it's an optional ...

+ Reply to Thread
Results 1 to 8 of 8

Thread: Re: [UNSURE] Re: [UNSURE] Re: [UNSURE] Re: Hardware TPU?

  1. Re: [UNSURE] Re: [UNSURE] Re: [UNSURE] Re: Hardware TPU?

    Bill Gunshannon wrote:
    > In article <45F7F998.10227.FB9267@stuart.brook.ca>,
    > "Stuart Brook" writes:
    >
    >>On 14 Mar 2007 at 10:26, Zane H. Healy wrote:
    >>
    >>
    >>>>I've read somewhere that this was a microcode option for PDP11/23-11/24
    >>>>machines, and it's an optional set of cards (KE44-A) for 11/44s. (I have an
    >>>>11/44, but no CIS.) -- Ian
    >>>
    >>>I know it's an option on the /23+ CPU card, as well as one or two other
    >>>options. They are chips that plug in. My /23+ has it. My /44 on the other
    >>>hand is lacking it, and I believe the FPU.

    >>
    >>There certainly wasn't a lot of software that actually
    >>supported it.

    >
    > I thought just the COBOL and Dibol compilers which is to say just
    > locally written applications. (And maybe some of the stuff in the
    > Software Sourcebook.) Just out of curiosity, did you have to build
    > specific programs for computers with and without CIS? If so, because
    > of their rarity, I would imagine very few people bothered outside of
    > their own shop.


    I believe that you told the compiler to generate code for CIS, just as
    you can tell some compilers wether or not to generate code for FPP, FIS
    or EIS.

    So yes, you'd need to generate a separate image for those who had CIS
    and for those who hadn't.

    Johnny


  2. CIS (Was: Re: Hardware TPU?)

    For those still curious, I dug out my RSTS product manual last night
    and read thru it. At least under RSTS COBOL seems to be the only
    language that mentioned using CIS. DIBOL does not mention it and
    being as they thought it important enough a distinction to mention
    it for COBOL I am sure they would have if DIBOL also supported it.

    bill

    --
    Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
    bill@cs.scranton.edu | and a sheep voting on what's for dinner.
    University of Scranton |
    Scranton, Pennsylvania | #include

  3. Re: [UNSURE] Re: [UNSURE] Re: Hardware TPU?

    On 14 Mar, 19:33, Johnny Billquist wrote:
    > Bill Gunshannon wrote:
    > > In article <45F7F998.10227.FB9...@stuart.brook.ca>,
    > > "Stuart Brook" writes:

    >
    > >>On 14 Mar 2007 at 10:26, Zane H. Healy wrote:

    >
    > >>>>I've read somewhere that this was a microcode option for PDP11/23-11/24
    > >>>>machines, and it's an optional set of cards (KE44-A) for 11/44s. (I have an
    > >>>>11/44, but no CIS.) -- Ian

    >
    > >>>I know it's an option on the /23+ CPU card, as well as one or two other
    > >>>options. They are chips that plug in. My /23+ has it. My /44 on the other
    > >>>hand is lacking it, and I believe the FPU.

    >
    > >>There certainly wasn't a lot of software that actually
    > >>supported it.

    >
    > > I thought just the COBOL and Dibol compilers which is to say just
    > > locally written applications. (And maybe some of the stuff in the
    > > Software Sourcebook.) Just out of curiosity, did you have to build
    > > specific programs for computers with and without CIS? If so, because
    > > of their rarity, I would imagine very few people bothered outside of
    > > their own shop.

    >
    > I believe that you told the compiler to generate code for CIS, just as
    > you can tell some compilers wether or not to generate code for FPP, FIS
    > or EIS.
    >
    > So yes, you'd need to generate a separate image for those who had CIS
    > and for those who hadn't.
    >
    > Johnny


    I persuaded my bosses to buy a CIS board for our overworked 11/44. I
    wrote CIS instructions into Bonner RUNOFF, and also replacement
    strxxx()
    functions for C (maybe I did Coral too), but I didn't have the courage
    to modify RSX itself. As most of the machine's workload was scientific
    number-crunching (yes we had an FPU also), any improvement was
    insignificant.

    Chris


  4. Re: CIS (Was: Re: Hardware TPU?)

    On 16 Mar 2007 13:12:10 GMT in alt.sys.pdp11, bill@cs.uofs.edu (Bill
    Gunshannon) wrote:

    >For those still curious, I dug out my RSTS product manual last night
    >and read thru it. At least under RSTS COBOL seems to be the only
    >language that mentioned using CIS. DIBOL does not mention it and
    >being as they thought it important enough a distinction to mention
    >it for COBOL I am sure they would have if DIBOL also supported it.


    Are you really surprised that the existence of the Commercial
    Instruction Set option was only assumed by COBOL? AFAICT it was only
    bought by customers who ran COBOL.

    --
    Thanks. Take care, Brian Inglis Calgary, Alberta, Canada

    Brian.Inglis@CSi.com (Brian[dot]Inglis{at}SystematicSW[dot]ab[dot]ca)
    fake address use address above to reply

  5. Re: CIS (Was: Re: Hardware TPU?)

    In article ,
    Brian Inglis writes:
    > On 16 Mar 2007 13:12:10 GMT in alt.sys.pdp11, bill@cs.uofs.edu (Bill
    > Gunshannon) wrote:
    >
    >>For those still curious, I dug out my RSTS product manual last night
    >>and read thru it. At least under RSTS COBOL seems to be the only
    >>language that mentioned using CIS. DIBOL does not mention it and
    >>being as they thought it important enough a distinction to mention
    >>it for COBOL I am sure they would have if DIBOL also supported it.

    >
    > Are you really surprised that the existence of the Commercial
    > Instruction Set option was only assumed by COBOL? AFAICT it was only
    > bought by customers who ran COBOL.


    It seems to me that you may have cause and effect backwards. I would
    assume it was bought by people using COBOL because it would make their
    COBOL run faster. If it had been used by BasicPlus, Dibol, or anything
    else others would likely have bought it, too. The only exception being
    Fortran as I can't really think of any way it would have improved on
    the things Fortran was used for. But I am sure that if support had
    existed people other than COBOL users would have bought it.

    bill


    --
    Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
    bill@cs.scranton.edu | and a sheep voting on what's for dinner.
    University of Scranton |
    Scranton, Pennsylvania | #include

  6. Re: CIS (Was: Re: Hardware TPU?)

    On 22 Mar 2007 12:24:52 GMT in alt.sys.pdp11, bill@cs.uofs.edu (Bill
    Gunshannon) wrote:

    >In article ,
    > Brian Inglis writes:
    >> On 16 Mar 2007 13:12:10 GMT in alt.sys.pdp11, bill@cs.uofs.edu (Bill
    >> Gunshannon) wrote:
    >>
    >>>For those still curious, I dug out my RSTS product manual last night
    >>>and read thru it. At least under RSTS COBOL seems to be the only
    >>>language that mentioned using CIS. DIBOL does not mention it and
    >>>being as they thought it important enough a distinction to mention
    >>>it for COBOL I am sure they would have if DIBOL also supported it.

    >>
    >> Are you really surprised that the existence of the Commercial
    >> Instruction Set option was only assumed by COBOL? AFAICT it was only
    >> bought by customers who ran COBOL.

    >
    >It seems to me that you may have cause and effect backwards. I would
    >assume it was bought by people using COBOL because it would make their
    >COBOL run faster. If it had been used by BasicPlus, Dibol, or anything
    >else others would likely have bought it, too. The only exception being
    >Fortran as I can't really think of any way it would have improved on
    >the things Fortran was used for. But I am sure that if support had
    >existed people other than COBOL users would have bought it.


    (Packed) Decimal arithmetic has never had much support other than COBOL
    and probably also RPG.

    --
    Thanks. Take care, Brian Inglis Calgary, Alberta, Canada

    Brian.Inglis@CSi.com (Brian[dot]Inglis{at}SystematicSW[dot]ab[dot]ca)
    fake address use address above to reply

  7. Re: CIS (Was: Re: Hardware TPU?)

    Brian Inglis wrote:

    (snip)

    > (Packed) Decimal arithmetic has never had much support other than COBOL
    > and probably also RPG.


    and PL/I, which may have borrowed it from COBOL.

    Though I have known IBM PL/I S/370 implementations that did
    FIXED DECIMAL in binary, IBM traditionally does it in packed
    decimal.

    -- glen


  8. Re: CIS (Was: Re: Hardware TPU?)

    On Mon, 26 Mar 2007 10:07:03 -0800 in alt.sys.pdp11, glen herrmannsfeldt
    wrote:

    >Brian Inglis wrote:
    >
    >(snip)
    >
    >> (Packed) Decimal arithmetic has never had much support other than COBOL
    >> and probably also RPG.

    >
    >and PL/I, which may have borrowed it from COBOL.
    >
    >Though I have known IBM PL/I S/370 implementations that did
    >FIXED DECIMAL in binary, IBM traditionally does it in packed
    >decimal.


    Forgot totally about PL/I, as I never used DECIMAL, also assembler, on
    any machine supporting decimal arithmetic.

    --
    Thanks. Take care, Brian Inglis Calgary, Alberta, Canada

    Brian.Inglis@CSi.com (Brian[dot]Inglis{at}SystematicSW[dot]ab[dot]ca)
    fake address use address above to reply

+ Reply to Thread