How to enter tiny Binary Floating-Point value

[John Ganci]

The latest Principles of Operation, SA22-7832-13, Chapter 19, describes the binary floating-point instructions. Page 19-2 describes the 4 BFP data formats: tiny, short, long, and extended. The tiny format is new. John Ehrman's assembler book and the latest HLASM Language Reference manual I could find (SC26-4940-09), only discuss BFP short, long and extended. Is there a newer Language Reference manual that describes how to enter tiny BFP values?

I also did not see any instructions in the Principles of Operation that use tiny BFP. Did I miss something?

Regards, John Ganci

[Seymour J Metz]

Did you look in chapter 26?

________________________________________
From: IBM Mainframe Assembler List <ASSEMBL...@LISTSERV.UGA.EDU> on behalf of John Ganci <00000da23edbcde...@LISTSERV.UGA.EDU>
Sent: Thursday, July 27, 2023 12:35 PM
To: ASSEMBL...@LISTSERV.UGA.EDU
Subject: How to enter tiny Binary Floating-Point value

[John Ganci]

Nope. Oops. Thanks.

[Seymour J Metz]

That still leave the question of HLASM syntax for DC. I would hope that it's been extended to handle tiny floating point.


--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3

________________________________________
From: IBM Mainframe Assembler List [ASSEMBL...@LISTSERV.UGA.EDU] on behalf of John Ganci [00000da23edbcde...@LISTSERV.UGA.EDU]

Sent: Thursday, July 27, 2023 2:13 PM
To: ASSEMBL...@LISTSERV.UGA.EDU
Subject: Re: How to enter tiny Binary Floating-Point value

[Jonathan Scott]

Shmuel (Seymour J.) Metz writes:
> That still leave the question of HLASM syntax for DC. I would
> hope that it's been extended to handle tiny floating point.

This has not been done so far, as the architects said it should
not be needed, as it was not for general programming purposes.

When the architects added the BFP tiny format (that is, the IEEE
754 half-precision binary floating-point format: binary16) we
asked whether this format should defined in Assembler, and we
were told not to bother, as it was primarily for use as a data
format for NNTP input which would normally be generated by a
compiler or obtained from other sources.

If someone submits a requirement showing a strong justification
for making the format available in Assembler, I think that our
existing IEEE floating point conversion routine should already
be able to handle it, so the main things to be resolved would be
what to call the DC type and what type attribute it should have
(given that all upper case letters and some other characters
have already been used).

Jonathan Scott, HLASM
IBM Hursley, UK

[Paul Gilmartin]

On 8/2/23 09:52:47, Jonathan Scott wrote:
> Shmuel (Seymour J.) Metz writes:
>> That still leave the question of HLASM syntax for DC. I would
>> hope that it's been extended to handle tiny floating point.
>
> This has not been done so far, as the architects said it should
> not be needed, as it was not for general programming purposes.

Are there hardware instructions supporting the format?
Could it be generated with a library macro?

> ... it was primarily for use as a data
> format for NNTP input ...


Ah, Usenet. I remember it fondly. Does LISTSERV rely on NNTP?

>
> ... the main things to be resolved would be

> what to call the DC type and what type attribute it should have
> (given that all upper case letters and some other characters
> have already been used).

HLASM appears to have made a bad design choice in making the
case attribute case-insensitive.

--
gil

[Farley, Peter]

<Grin> I believe Jonathan was referring to “NNP-Data-Type-1 Format”, not the old “Network News Transfer Protocol”. Too many acronyms . . . 😊

Peter

From: IBM Mainframe Assembler List <ASSEMBL...@LISTSERV.UGA.EDU> On Behalf Of Paul Gilmartin
Sent: Wednesday, August 2, 2023 12:14 PM
To: ASSEMBL...@LISTSERV.UGA.EDU
Subject: Re: How to enter tiny Binary Floating-Point value

This message and any attachments are intended only for the use of the addressee and may contain information that is privileged and confidential. If the reader of the message is not the intended recipient or an authorized representative of the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail and delete the message and any attachments from your system.

[Jonathan Scott]

Ref: Your note of Wed, 2 Aug 2023 16:55:30 +0000

> <Grin> I believe Jonathan was referring to “NNP-Data-Type-1 Format”, not
> the old “Network News Transfer Protocol”. Too many acronyms . . . 😊
>
> Peter

Should have been NNPA. I'm currently off sick and thought
I had recovered enough to answer that pending question, but
perhaps my brain still needs more rest!

[Ze'ev Atlas]

If you give themthe tool, the will find a use.
For completeness, and for those who really push th limits, how hard is to create that data type?ZA

Sent from Yahoo Mail on Android

[John Ganci]

The reason for the original post comes via Don Higgins' z390 assembler, linker and emulator project -- see http://z390.info or https://github.com/z390development/z390. Some current work is adding new z16 instructions to the z390 assembler and emulator. We noticed the introduction of tiny BFP in the latest Principles of Operation so decided to add some support for it, thus wondering how HLASM handles it.

Not really a justification for making the format available in HLASM. Thanks for the replies from Seymour and Jonathan.

Regards, John Ganci