cuda - Number of cores on NVIDIA K20 GPU in Mathematica -
i using nvidia k20 gpu in mathematica. whitepaper k20 states following specs:
- 2496 single-precision cores (192 per multiprocessor x 13 multiprocessors)
- 832 double-precision cores (64 per multiprocessor x 13 multiprocessors)
- 416 special function units (32 per multiprocessor x 13 multiprocessors)
however, when launch mathematica , @ specs of gpu after loading opencl or cuda, says have 416 cores (i have attached screenshot of mathematica commands , reported specs). somehow related number of double-precision cores or special function units? mathematica reports these specs before tell kind of code write (e.g. double precision, transcendental functions).
it more mathematica not yet know kepler architecture yet, , translates multiprocessor count core count incorrectly.
since "cores" marketing invention anyway, can safely ignore number. mathematica still use entire gpu (although kepler-aware version might still bring speedup through other optimizations).
Comments
Post a Comment