polynomial arithmetic over ginac

[Parisse Bernard]

> Hmm, just don't know whether I'm kicking a dead horse here.  It's an old
> thread.  But really, I've been hacking on CLN/GiNaC support for ARM
> lately and they seems to fit well into quite old ARM computers.  Many
> PDAs are running on such things.  So, the portability is there and the
> machines are getting faster than what we see here:
> 
> timings snipped

>From my own testing with giac/xcas on skiffcluster5 (Compaq cluster),
it seems that you can expect about the speed of a Pentium 60Mhz (well
I hope the speed of skiffcluster5 is about the same as the speed
of an iPAQ). The problem is more size (display size
and flash ROM size) than speed. Some examples (without optimization
flag for compilation)

factor     PIII-500  skiffcluster5   HP49
x^10-1     too fast     0.1s          17.0s
x^100-1       0.12s     1.82s         not tested
x^202+x^101+1 4.07s     52.07s