It reports that gemm is not implemented. gemm is provided by blas libraries, e.g., openblas or cublas.
Do you have openblas? How did you set the compilation options, e.g., USE_CUDA? On Wed, Jul 13, 2016 at 1:58 PM, Erhard M. Dinhobl jun. < [email protected]> wrote: > Hi > > After compiling when I run ./bin/test_singa I get: > > [F d0712 t15:08:08 p28750:312 > /home/erhard/apache-singa/src/core/tensor/./tensor_math.h:323] GEMM Not > Implemented > > And it writes a Memory dump > > -- > > master-studios.net - Intelligent Software / Consulting / Research > > +43 (0)699 / 102 69 056, Vienna, Austria
