CPU handling on CRAN

[fontikar]

Currently our package using too CPU testing

According to CRAN testing, our package is currently failing this criterion one of the many testing servers….
Flavor: r-devel-linux-x86_64-debian-gcc
Check: tests, Result: NOTE
Running 'testthat.R' [937s/57s]
Running R code in 'testthat.R' had CPU time 16.5 times elapsed time

I don't understand how this happens as I don't believe we have any parallel processing in the package (especially in tests...)

Updating it to best practices is quite a big job but probably ensures APCalign won't fail in the future.

Will branch and see how long this will take to implement

[dfalster]

Crikey. How long does it take on the other architectures?

[fontikar]

Thats the thing, its not it other servers and its a bit of a rabbit hole to dig to try reproduce. I think I will do some digging if no luck, I'll post to posit community then to R dev mailing list (terrifying but probably necessary)

[fontikar]

OK looks like I can replicate here with checkhelper

In the check_output, you will get the same outputs, in the same format as used by CRAN, for the
pre-test of incoming packages.

testing now...

[wcornwell]

we do have low level parallel processing via the stringdist library and openMP

see https://www.rdocumentation.org/packages/stringdist/versions/0.9.1/topics/stringdist-parallelization

[wcornwell]

i think

options('sd_num_thread'=1)

in the test part will probably handle it?

[fontikar]

Oooh this has been addressed! in #254 its currently failing for other reasons that I was unable to reproduce anywhere. Sorry folks... closing this one though!