t...@gmplib.org (Torbjörn Granlund) writes:
> I believe we already use clock_gettime(2) directly on GNU/Linux, which
> is the underlying syscall that clock(3) uses in newer GNU libc versions.
Last time I looked, one had to run speed with -o cycles-broken to force
using clock_gettime.
Maybe it wo
paul zimmermann writes:
I don't know how to check which function is used. With gmp-6.1.0 on my
workstation I get:
zimmerma@tomate:/tmp/gmp-6.1.0/tune$ ./speed
...
CPU cycle counter, supplemented by microsecond getrusage()
Gnuplot home page http://www.gnuplot.info/
Quickplot home
> since version 2.18 of the glibc, the clock() function is much more precise.
> For example on my machine it has a resolution of 1 micro-seconds whereas
> getrusage() has a resolution of 4ms only, as demonstrated by the program
> below:
>
> I believe we already use clock_gettime(2) direc
paul zimmermann writes:
since version 2.18 of the glibc, the clock() function is much more precise.
For example on my machine it has a resolution of 1 micro-seconds whereas
getrusage() has a resolution of 4ms only, as demonstrated by the program
below:
I believe we already use clock_ge
