Dirk Eddelbuettel <edd <at> debian.org> writes:> One last follow-up. Debian's Atlas maintainer has started to work in the > OpenGoto packaging. There will be proper Goto packages now that the license > is better, but it may take a while. It took years to get Atlas packaged > 'right' so don;t expect anything too soon. The gotoblas2-helper package is > still useful for the 'here and now'.Unfortunately it is broken :( Link not found at ./getGotoBLAS2 line 83
On 9 May 2012 at 17:41, Sergei Petrov wrote: | Dirk Eddelbuettel <edd <at> debian.org> writes: | | > One last follow-up. Debian's Atlas maintainer has started to work in the | > OpenGoto packaging. There will be proper Goto packages now that the license | > is better, but it may take a while. It took years to get Atlas packaged | > 'right' so don;t expect anything too soon. The gotoblas2-helper package is | > still useful for the 'here and now'. | | Unfortunately it is broken :( | | Link not found at ./getGotoBLAS2 line 83 I recommend using the (already packaged) libopenblas-base / libopenblas-dev package. In casual benchmarking it outperformed the gotoblas2 package it was originally based on. Dirk -- R/Finance 2012 Conference on May 11 and 12, 2012 at UIC in Chicago, IL See agenda, registration details and more at http://www.RinFinance.com
2012/5/9 Dirk Eddelbuettel <edd at debian.org>:> I never ever play by hand with update-alternatives to override (and I can > simply install or remove as needed to get my preferred choice working), and > guess what, nothing is borked on my system. > > I recommend you simply purge the related BLAS packages, install again and > enjoy the (very well working) defaults.Removing atlas* completely repaired lapack. Thank you!
Maybe Matching Threads
- GotoBLAS2 breaks lapack
- Compiling R-2.11.1 patched on Ubuntu 10.04 with Optimized GotoBLAS2 libraries
- Function nlme::lme in Ubuntu (but not Win or OS X): "Non-positive definite approximate variance-covariance"
- SurviveGotoBLAS2 for Win64 (human sacrifice release)
- Lapack or Blas crashing R when using "large" matrices (Ubuntu 11.04)