zeratax / yacx

Yet Another CudaExecutor - wrapper to easily compile and execute cuda kernels
https://zeratax.github.io/yacx
MIT License
8 stars 4 forks source link

Ram usage benchmark #133

Closed LukasSiefke closed 4 years ago

LukasSiefke commented 4 years ago

Benchmark-Method can take lots of Ram when use lots of different dataSizes because the Java-garbage-collector do not collect unnecessary KernelArgs fast enough. So after every benchmark for every dataSize the to the KernelArg correspondings C-Objects should be destroyed manually.