Le dimanche 26 février 2023, 21:37:51 UTC Axel Beckert a écrit : > Control: tag -1 + confirmed > > Hi Andreas, > > Andreas Beckmann wrote: > > Checking intel-mkl (pre-built binaries in non-free) with lintian is very > > slow. A full build (i.e. source+all+any) on amd64 takes nearly 18 hours > > to check with > > lintian -i -E -L ">=pedantic" -v intel-mkl_2020.4.304-4_amd64.changes > > on my local machine (8 cores, 64 GB, tmpfs on /tmp, storage on nvme). > > Thanks for that bug report. Knowing such examples to test is always > good to know. Performance is one thing we managed to get much better > in the past ¾ year or so, especially thanks to Bastien. But there is > obviously (and not only known since this bug report) place for > improvement. :-) Please note that cruft source could be run in parallel if needed. I am not sufficiently fluent in perl but if it exists a await stuff (like in python), or a green thread we could easily increase the performance here. Bastien > > > I don't know where all the time is spent ... > > Well, 549 MB of source package... :-) > > I'll check if --perf-debug brings up anything helpful. Well, yes, but > I would have gathered that otherwise, too: > > Warning in processable ../libmkl-threading-dev_2020.4.304-4_amd64.deb: > MLDBM error: Second level tie failed, "No space left on device" at > /usr/share/perl5/MLDBM.pm line 144. > > So, new try. > > From the output I already see that these checks took quite a while > (that's where output stopped for quite a while: > > Running check: debian/control/field/rules-requires-root on source:intel-mkl_2020.4.304-4 ... > > Running check: documentation/manual on binary:libmkl-meta-threading_2020.4.304-4_amd64 ... > > Anyway, will wait what I get at the end. Everything with "done (0." > can be ignored. :-) > > Anyway, tagging as confirmed as I can already see now that it will > take a while. :-) > > Regards, Axel >
Attachment:
signature.asc
Description: This is a digitally signed message part.