[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#994426: octave-sparsersb: flaky autopkgtest on ci.d.n armhf worker



On 20210915@22:12, Paul Gevers wrote:
> Source: octave-sparsersb
> Version: 1.0.8-3
> Severity: serious
> X-Debbugs-CC: debian-ci@lists.debian.org
> Tags: sid bookworm
> User: debian-ci@lists.debian.org
> Usertags: flaky timesout
> 
> Dear maintainer(s),
> 
> I looked at the results of the autopkgtest of you package on armhf
> because with a recent upload of glibc the autopkgtest of
> octave-sparsersb fails in testing. In august 2021, we replaced the host
> we were using for armhf testing for a new one. Since then, the
> autopkgtest has been failing most of the times, due to it timing out. We
> should get this fixed. Do you have any ideas what could be the root
> cause of this?
> 
> Ironically, the passing tests (since the new host) report a segfault
> and: Summary: 0 tests, 0 passed, 0 known failures, 0 skipped
> 
> Don't hesitate to contact us at debian-ci@lists.debian.org if you need
> help debugging this issue.
> 
> Paul
> 
> https://ci.debian.net/packages/o/octave-sparsersb/testing/armhf/
> 
> https://ci.debian.net/data/autopkgtest/testing/armhf/o/octave-sparsersb/14991286/log.gz
> 
> ***** test
>  a=sparsersb(sprand(100,100,0.4));
>  nrhs=1;
>  maxr=1;
>  tmax=1;
>  tn=2;
>  sf=1;
>  o=sparsersb(a,"autotune","n",nrhs,maxr,tmax,tn,sf);
>  assert(o==a)
> librsb error:The requested feature (e.g.:blocking) is not available
> because it was opted out or not configured at built time.
> librsb error:The requested feature (e.g.:blocking) is not available
> because it was opted out or not configured at built time.
> !!!!! test failed
> index (_,0,_): subscripts must be either integers 1 to (2^31)-1 or logicals
> ...

Dear Paul,

The problem seems librsb-sided.

Assuming librsb fails at detecting cache memory size, maybe can you
first export e.g.
  RSB_USER_SET_MEM_HIERARCHY_INFO="L2:4/64/512K,L1:8/64/32K" 
and see if this avoids that?
Same with 
  OMP_NUM_THREADS=1
explicitly?

Maybe you send me outputs of  `rsbench -C` and `rsbench -I` ?

If I had a guest account to such an armhf machine it would be a bit
easier.

Cheers,
Michele


Reply to: