Experimental Physics and
Industrial Control System

[email protected] (Jeff Hill) · Sat, 11 Mar 95 10:59:50 MST

Nick,

> OK, OK, all you users of CA on VMS ;-). Have you fixed this one before I
> embark on a fault finding exercise?
> 

Maybe I should contact Intel and ask for a copy of their floating 
point test :-}.

VAX D_FLOATING <=> IEEE double floating point conversion were recently 
added to CA for EPICS release 3.12. I just looked at the code and it 
does appear at first glance to be forcing the sign to zero when the 
IEEE exponent is zero. I am forcing IEEE exponent values 
outside the range of the VAX D_floating to have the proper value.
I also appear to be not allowing IEEE exponent values at the the 
D_FLOATING range limit to convert into a zero D_FLOATING exponent 
with a potentially negative sign.

Part of the problem here is that the VAC C compiler defaults to
D_FLOATING when G_FLOATING is a much better match for IEEE double
precision.

Nick, could you send the binary value of the IEEE value that
generated this exception? This can be obtained in the debugger
by examining the structure pointed to by pIEEE after the
exception occurs. You will need to compile CA for debugging.

I will look at the code in more detail on Monday.

Also is there a strong interest in retaining the 3.11 capability
for CA events to preempt the IPL0 thread via VMS ASTs for EPICS 3.12? 
If so I will take another look at this on Monday also.

Jeff

______________________________________________________________________
Jeffrey O. Hill			Internet	[email protected]
LANL MS H820			Voice		505 665 1831
Los Alamos, NM 87545 USA 	FAX		505 665 5107

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System