Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: VAX <--> IEEE FP

> So my question for the DEC hive mind is: When we get a VAX FP software compl$

> A- Convert the in-register VAX format to IEEE.
> B- Then use the IEEE soft float library to execute the operation.
> C- Then convert the in-register IEEE format back to VAX.

> ?

I don't have enough details in my head to be certain, but it seems to
me that, for each conversion step (A and C, above), you have to either
lose some range or lose some precision, depending on how many exponent
vs mantissa bits the formats have.  VAX FP is specified to have fairly
tight error bounds, error bounds I'm not sure you can preserve through
the above process.

It probably will be good enough for almost all purposes; most code
doesn't care about an error bound changing from, say, .5 ULP to 2 ULP
(correspnoding to two bits lost/gained.  Other code may care about the
range being reduced by a factor of (say) 4.  Seriously careful numeric
code is likely to care about both.

Only you can decide whether that's reason enough to bother fully
simulating VAX FP.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B

Home | Main Index | Thread Index | Old Index