Hi Martin and Tomas,
Thanks for your reasoned replies. It seems that improving this is going to
take more effort in pinning down exactly what is appropriate than I
anticipated.
Sorry if the intention was to keep the initial discussion of this off the
list, I didn't mean to cause offence by
> Tomas Kalibera
> on Fri, 5 Jan 2018 00:41:47 +0100 writes:
> In practical terms, failing tests are not preventing anyone from using
> an optimized BLAS/LAPACK implementation they trust. Building R with
> dynamically linked BLAS on Unix is
In practical terms, failing tests are not preventing anyone from using
an optimized BLAS/LAPACK implementation they trust. Building R with
dynamically linked BLAS on Unix is supported, documented and easy for
anyone who builds R from source. It is also how Debian/Ubuntu R packages
are built
Hi Tomas,
Thanks for your reply.
I find your response curious, however. Surely the identical() test is
simply incorrect when catering for possibly different BLAS
implementations? Or is it the case that conformant BLAS implementations
all produce bit-identical results, which seems unlikely?
We build R with dynamically linked BLAS and LAPACK libraries, in order
to use the AMD Core Math Library (ACML) multi-threaded implementation
of these routines on our 64 core servers. This is great, and our
users really appreciate it.
However, when building like this, make check fails on the