szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-379315463
@pengzhao-intel thanks for the update. @ashokei let us know how things
progress and how this change impact
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-379156918
Unless intel provides a static library for libiomp5 built with -fPIC flag,
or drop the dependency on it, I
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-379156409
@pengzhao-intel any update?
This is an au
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-356790792
@rsdubtso Hi. I can verify that mxnet pip package is linked dynamically to
`libiomp5.so`. We include `libi
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-356771456
@ykim362 could you elaborate? Currently I'm using '${ORIGIN}' in rpath for
mkl builds already, and shippin
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-356478981
@ykim362 our goal is to be able to provide good out-of-the-box experience
for MKL-enabled mxnet. Though th
szha commented on issue #8532: mxnet-mkl (v0.12.0) crash when using
(conda-installed) numpy with MKL
URL:
https://github.com/apache/incubator-mxnet/issues/8532#issuecomment-341834234
@ykim362 what's your recommendation on this? MKL2018 doesn't distribute .a,
which means that I can't hide
