| 1 |
Just an update: gmsh build with USE="-blas" |
| 2 |
Denis |
| 3 |
|
| 4 |
|
| 5 |
On Wed, Aug 31, 2011 at 01:14, denis cohen <denis.cohen@×××××.com> wrote: |
| 6 |
> Following Michael Mol's advice I masked these in /etc/portage/package.mask: |
| 7 |
> |
| 8 |
>>=virtual/blas-2.0 |
| 9 |
>>=virtual/cblas-2.0 |
| 10 |
>>=virtual/lapack-3.1 |
| 11 |
>>=sci-libs/cblas-reference-20030223-r5 |
| 12 |
>>=sci-libs/lapack-reference-3.2.1 |
| 13 |
>>=sci-libs/gsl-1.14 |
| 14 |
> |
| 15 |
> and did an "emerge -uD world". |
| 16 |
> |
| 17 |
> All blocks were gone and after the emerge I got gsl working again. |
| 18 |
> I am also not sure what use is eselect-cblas since eselect blas works |
| 19 |
> without it. |
| 20 |
> |
| 21 |
> gmsh still won't emerge with these types of errors: |
| 22 |
> ... |
| 23 |
> [100%] Building CXX object CMakeFiles/gmsh.dir/contrib/bamg/bamglib/Meshio.cpp.o |
| 24 |
> [100%] Building CXX object |
| 25 |
> CMakeFiles/gmsh.dir/contrib/bamg/bamglib/QuadTree.cpp.o |
| 26 |
> [100%] Building CXX object |
| 27 |
> CMakeFiles/gmsh.dir/contrib/bamg/bamglib/SetOfE4.cpp.o |
| 28 |
> Linking CXX executable gmsh |
| 29 |
> CMakeFiles/gmsh.dir/Numeric/fullMatrix.cpp.o: In function |
| 30 |
> `fullMatrix<double>::invertInPlace()': |
| 31 |
> fullMatrix.cpp:(.text+0xcc): undefined reference to `dgesv_' |
| 32 |
> CMakeFiles/gmsh.dir/Numeric/fullMatrix.cpp.o: In function |
| 33 |
> `fullMatrix<double>::gemm(fullMatrix<double> const&, |
| 34 |
> fullMatrix<double> const&, double, double)': |
| 35 |
> fullMatrix.cpp:(.text+0x1e9): undefined reference to `dgemm_' |
| 36 |
> CMakeFiles/gmsh.dir/Numeric/fullMatrix.cpp.o: In function |
| 37 |
> `fullVector<double>::axpy(fullVector<double> const&, double)': |
| 38 |
> fullMatrix.cpp:(.text+0x260): undefined reference to `daxpy_' |
| 39 |
> ... |
| 40 |
> |
| 41 |
> So making some progress. |
| 42 |
> |
| 43 |
> Denis |
| 44 |
> |
| 45 |
> |
| 46 |
> |
| 47 |
> On Wed, Aug 31, 2011 at 01:02, Alex Schuster <wonko@×××××××××.org> wrote: |
| 48 |
>> denis cohen writes: |
| 49 |
>> |
| 50 |
>>> I've unmerged cblas* blas* gsl, synced, and "emerge -uDN world "also |
| 51 |
>>> without success (also tried cblas-external USE flag but could not see |
| 52 |
>>> it with eix). |
| 53 |
>> |
| 54 |
>> What's your problem with the world update? Adding --tree to the emerge |
| 55 |
>> command might show what pulls in what. I had to unmerge clapack and lapack- |
| 56 |
>> reference yesterday, but I cannot remember the details - it all started with |
| 57 |
>> digikam not building because some blas stuff was missing. But finally the |
| 58 |
>> @world update was possible. And both got re-installed. |
| 59 |
>> |
| 60 |
>>> eselect cblas also gave no results. |
| 61 |
>> |
| 62 |
>> Here it is set to the only option gsl. |
| 63 |
>> |
| 64 |
>>> One problem I may have is with eselect. eselect-cblas has been |
| 65 |
>>> unmerged but trying to re-emerge it gives this block. |
| 66 |
>> |
| 67 |
>> I get the very same blocker as you. eselect cblas is working, though. So is |
| 68 |
>> eselect-cblas being needed at all? |
| 69 |
>> |
| 70 |
>> Wonko |
| 71 |
>> |
| 72 |
>> |
| 73 |
> |
Archives