Re: +-Inf and NaN
Stavros Macrakis <[email protected]> Wed, 21 Feb 2024 14:55:58 -0500
| Newsgroups | gmane.comp.mathematics.maxima.general,gmane.lisp.gcl.devel |
|---|---|
| Message-ID | <CACLVabX6xzy3OxM5AshBWehV4Ks8ygqGVUmsTraAR=BAN6jqOQ@mail.gmail.com> |
On Wed, Feb 21, 2024 at 2:06 PM Henry Baker <[email protected]> wrote: > Re: hardware: > > Hmmm... In just the Intel processor category, there are probably > millions of combinations > of HW specs, so 'close to the HW' is exceedingly ill-defined. E.g., I > understand that Windows-11 > will no longer boot on Intel HW that doesn't have a popcount instruction > ?!? > ... Modulo bugs (have there been any recent ones?), I believe that the IEEE 754 floating point arithmetic is consistent to the bit across processors. There might be differences in how signaling NaNs behave, but that's not relevant here. _______________________________________________ Maxima-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/maxima-discuss