As reported previously, I was able to build a debug cn.exe with ACML 5.3.1 under gcc 4.8.2 (Ubuntu 14.04).
I now built a debug cn.exe with MKL from Intel Composer XE 2015.1.133. I had to replace line 3550 of CPUMatrix.cpp with:
The resulting cn.exe ran and produced a log file. The speed on Demos/Simple was now several times faster than with ACML but still slower than the Windows version, by a factor of over 20.
- //[email removed] fixed this
- #ifndef USE_MKL
- long int info;
- int info;