Weird Performance

1. If you see serious performance loss (extremely low performance),

   probably you created too many threads or process. Basically GotoBLAS

   assumes that available cores that you specify are exclusively for

   BLAS computation. Even one small thread/process conflicts with BLAS

   threads, performance will become worse. 

   The best solution is to reduce your number of threads or insert

   some synchronization mechanism and suspend your threads until BLAS

   operation is finished.

2. Simlar problem may happen under virtual machine. If supervisor

   allocates different cores for each scheduling, BLAS performnace

   will be bad. This is because BLAS also utilizes all cache,

   unexpected re-schedule for different core may result of heavy

   performance loss.

Anyway, if you see any weird performance loss, it means your code or

algorithm is not optimal.