jds@odin.m2c.org (John D. Sutter) (07/29/88)
We are currently porting to the Sun3 a set of tools that we support
inside the M2C consortium. Currently RNL(V5) and Spice (3b1)
are completely right on the Sun. I believe I have resolved all the
null pointer problems and a good bit of plain nonsense in the code.
It seems someone may have already addressed these problems, if so I
would appreciate hearing from them.
The problems:
rnl (V5):
Rnl gives considerably different timing values, but the logic
seems correct. The format is not an issue, just the time values:
VAX
out.1=0 @ 1408.4
out.2=0 @ 1411.5
out.3=1 @ 1418.1
Sun
out.1=0 @ 1.435300e+03
out.2=0 @ 1.437200e+03
out.3=1 @ 1.563300e+03
spice (3b1)
Spice3 doesn't converge as well on the Sun as the Vax. An exampl
is 'mosamp' given as an example. This deck converges on the Vax,
but not the Sun. The incidence of non-convergence on the Sun is
low, but it happens. If you have already found a solution, I'd
like to hear about it.
-- john
------------------------------------------------------------------
John Sutter, Mass. Microelectronics Center (M2C)
(w) 75 North Drive, Westborough, MA 01581, USA +1 (508) 870-0312
(h) 516 Dutton Rd., Sudbury, MA 01776, USA +1 (508) 443-0515
UUCP: harvard!m2c!jds INTERNET: jds@m2c.org NSS 29355