Re: [gnugo-devel] Go as SPEC benchmark?

[Arend Bayer]

Jim Koehler wrote:

>   I haven't detected a lot of interest in adapting Go for cpu2004, but I

Hmm, the discussion on the list stalled. I don't know if this means that
nobody else is interested; but even if only Gunnar and I are
interested,

The work we would need to do is, as far as I can see:
1. Sort out the 64-bit issue with the cache (if this doesn't work out
we could disable it).
2. Audit for remaining floating point problems. (I believe there could
still be some in readconnect.c.)
3. get rid of GCC macros

I think if we use a gtp script as work load, which includes a query for
all move counters at the end, then we can say with confidence that
either there were no problems if the output is as predicted. (And this
would be a serious check for platform independance, much better than
our trial-and-error attempts for regression inconsistencies prior to 3.2.)

And then of course bother with potentially upcoming...

> thought I might take a stab at getting it off the ground.  I built it
> for an Itanium-based HU-UX system, but the executable aborts right
> away.

...problems :-)

Arend