[Simh] O2 optimization on Linux

[dott.Piergiorgio d' Errico]

Il 15/03/2012 17:44, Bob Supnik ha scritto:

> Here's the "uses unsigned" status on the other simulators that I've
> written:
>
> uses unsigned - GRI, HP (as rewritten by Dave Bryan), 1620, 7094,
> Interdata 16b/32b, LGP, Alpha (unreleased), Sigma (unreleased)
> uses signed - PDP1, PDP18b, PDP9, PDP10, PDP11, VAX, Nova, 1401, H316
>
> Of the "signed" simulators, I think only the VAX is at risk. Every other
> one is doing zero-extended 6b, 12b, 16b, 18b, or 24b arithmetic in a 32b
> container, or 36b arithmetic in a 64b container, and integer overflow
> should not occur.

I can confirm, because in my fooling around an hypothetical 6006 (an 
unusual pet project of mine...) I haven't found signed/unsigned issues.

(side note, I'm waiting for 3.9.0 and the corrected LGP simulator for 
really starting my planned work on bendix G-15 simulation...)

Best regards from Italy,
dott. Piergiorgio.