| Summary: | dev-python/numpy-1.3.0 fails to emerge with MKL and USE="lapack" | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Flo Gravo <even.more.spam.for.me> |
| Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
| Status: | RESOLVED TEST-REQUEST | ||
| Severity: | normal | CC: | josemariafg, nicolas.pinto, python, weigt.mathias |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | /var/tmp/portage/dev-python/numpy-1.3.0/temp/build.log | ||
|
Description
Flo Gravo
2009-06-08 14:16:31 UTC
Created attachment 193898 [details]
/var/tmp/portage/dev-python/numpy-1.3.0/temp/build.log
(In reply to comment #1) > Created an attachment (id=193898) [edit] > /var/tmp/portage/dev-python/numpy-1.3.0/temp/build.log > Fails here too (1.3.0-r1 also9), without MKL!!!! and USE="lapack". Will this get now some attention? (In reply to comment #2) > (In reply to comment #1) > > Created an attachment (id=193898) [edit] > > /var/tmp/portage/dev-python/numpy-1.3.0/temp/build.log > > > > Fails here too (1.3.0-r1 also9), without MKL!!!! and USE="lapack". Please forget my comment as I'm obviously an idiot: Someone changed my CFLAGS (to something weird containing -ffast-math) after changing back and recompiling python numpy compiles fine (with and without USE="lapack"). *** Bug 321737 has been marked as a duplicate of this bug. *** Bump, please. Still true with numpy 1.5.0-r2 and 1.5.1. I must uninstall mkl for numpy to build. I would like to be able to use mkl with numpy. Works her, could please attach /usr/lib64/pkgconfig/blas.pc? On a Gentoo x86 installation with Intel mkl 10.0.5.025 I have got numpy emerged with: LIBRARY_PATH=/opt/intel/mkl/10.0.5.025/lib/32 emerge -avq1 numpy so the gcc finds in the linking stage the libraries pointed by /usr/lib/libblas.so, which is a symlink to /opt/intel/mkl/10.0.5.025/lib/32/libmkl.so, which is indeed an ld script: GROUP (libmkl_intel.so libmkl_intel_thread.so libmkl_core.so) What about latest numpy and mkl? |