Re: Please test pari-2.11.3 prerelease 1
James Wanless <[email protected]>
| Newsgroups | gmane.comp.mathematics.pari.devel |
|---|---|
| Message-ID | <[email protected]> |
2.11.3pre1 (nbthreads = 4) 1:62004 2:58477 3:60179 4:59802 5:59267 6:59135 7:60054 8:60752 9:57857 10:61302 2.11.2 (threading engine : single) 1:28196 2:27404 3:26215 4:27252 5:27068 6:27716 7:27074 8:26912 9:27414 10:26007 J On 17 Feb 2020, at 22:43, Bill Allombert <[email protected]> wrote: > On Mon, Feb 17, 2020 at 07:47:37PM +0000, James Wanless wrote: >> Hi Bill >> >> Great, thanks! Fixed. >> >> [Though I ought to just mention there’s possibly a speed regression??? >> On macports version of single-threaded 2.11: >> ? ## >> *** last result computed in 24,431 ms. >> On this latest binary (4 threads) : >> ? ## >> *** last result computed in 57,177 ms. ] > > Could you check using the real time instead of the CPU time ? > > p=nextprime(10^500); > for(i=1,10,setrand(i);t=getwalltime();C=primecert(p);print(i,":",getwalltime()-t)) > > Cheers, > Bill. >