I was just watching google desktop -- that's supposed to index in background
using up 75% of my cpu(s) (3 out of 4) and wanted to try something
cpu-intensive to see if it would back-off (it doesn't).
I decided maybe factoring a large int might do the trick. So I popped into my
"cygwin/win32" window and started trying numbers with factor.
I found that the version on cygwin/win32 would take a maximum of 20 digits.
12345678901234567890 worked, but
123456789012345678901 fails with a "is too large" error.
It looks like it handles up to 2**64-1.
I thought maybe a machine with a 64-bit word size might give more, so I
tried it on a linux running x86-64. But was surprised when it was limited
to 2**63-1!
Should they be the same on a 32-bit platform (same type of processor)
as one running in 64 bits?
Would it completely kill performance to use the arbitrary precision lib like
'bc' appears to use?
Might be an interesting way to rate processor speeds -- I note
that
12345678901234567891 (substituting a 1 for the 0, in a 20-digit num on
cygwin)
takes a "while" (but can't compare it on my 64-bit
machine
because the number is "too big").
I'm guessing the cygwin version uses 64 bit-unsigned? Whereas the version on
the
64-bit machine is running in signed? Not sure why there'd be a difference...
except that the 64-bit version is V6.9, while the Cygwin ver is V6.10.
Was that changed between the deltas?
Curious,
-linda
_______________________________________________
Bug-coreutils mailing list
Bug-coreutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-coreutils
