time/batch about same regardless of gpuid

karpathy / char-rnn

Multi-layer Recurrent Neural Networks (LSTM, GRU, RNN) for character-level language models in Torch

11.53k stars 2.58k forks source link

time/batch about same regardless of gpuid #156

Closed robinmeier closed 8 years ago

robinmeier commented 8 years ago

finally managed to install cutorch and tried running the test:

th train.lua

my time/batch is about 0.2s regardless of if i use gpuid -1, gpuid 0, gpuid 1.... how can i check if lua is really using my gpu? and shouldn't it be faster with one of above options?