Read Alpha F0 reg. from MACRO32 ? (need to call math$random)

[John Simmonds]

Is there a technique to get at the FP registers
from MACRO32 source ?

Thanks very much!
John H Simmonds.

[Volker Halle]

John,

here is a pointer to an answer from Ask The Wizard:

http://h71000.www7.hp.com/wizard/wiz_2553.html

... the Macro32 compiler has no idea how to access F0 nor any of the other floating point registers ...

Volker.

[kari salminen · Accepted solution]

You can do it with a small trick in MACRO-32.

See,

http://h71000.www7.hp.com/freeware/freeware80/probe/src/DMS.MAR

has nice routine AXP_FIX_R0 how to move F0 in R0.

It's actually MACRO-64 embedded into MACRO-32 code.

Kari Salminen

[John Simmonds]

Volker, Thanks for confirming what I knew :)
I was wondering whether there were some
unsupported things like asm {} ;) but was
good to learn that the MACRO-64 compiler
is free too.

Kari, Now that's more like it !! Lovely
stuff !!

Thanks to you both.

[John Reagan]

BTW, that Macro-32 hack is just plain nasty. It assumes that the PSECT that contains the hand-crafted instructions will always be after the PSECT that ends with the label for AXP_FIX_R0. Make sure you use sufficient linker magic to keep them together. Also, don't even think about doing this in I64.

[John Simmonds]

John,

Thanks for your wise comments .. I meant
to bracket my (wicked) delight at seeing that
horrible hack with a "tounge firmly in cheek"
indicator.

But I did use it .. (and modified the machine
code a little to do an integer scaling with
the FP number) .. simply because the target
machine (in Production) had no license for
any compiler and I had no approval to install
the Macro-64 compiler.

Do you know any way just using the Macro-32
compiler ?

Best wishes,
John.

[John Reagan]

No, the Macro-32 compiler has absolutely no knowledge of the Alpha (or I64) floating register set. We don't use 'em at all.

You need the jacket routine to move F0 into R0. No way around that.