floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Re-enable DAXPY microkernels for x86_64 

as the inaccuracies seen in the original testcase for #1332 appear to be due to an artefact that amplifies the very small rounding differences between FMA and discrete multiply+add
This commit is contained in:
Martin Kroeker 2018-03-04 19:37:03 +01:00 committed by GitHub
parent 0ab5bf1746
commit 81215711a2
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 0 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
kernel/x86_64/daxpy.c
View File

@ -33,17 +33,12 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "daxpy_microk_nehalem-2.c"
#elif defined(BULLDOZER)
#include "daxpy_microk_bulldozer-2.c"
/*
these appear to be broken, see issue 1332
#elif defined(STEAMROLLER) || defined(EXCAVATOR)
#include "daxpy_microk_steamroller-2.c"
#elif defined(PILEDRIVER)
#include "daxpy_microk_piledriver-2.c"
#elif defined(HASWELL) || defined(ZEN)
#include "daxpy_microk_haswell-2.c"
*/
#elif defined(HASWELL) || defined(ZEN) || defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
#include "daxpy_microk_sandy-2.c"
#elif defined(SANDYBRIDGE)
#include "daxpy_microk_sandy-2.c"
#endif