GeForce GTX TITAN X Pascal is 200% better than its predecessor in terms of performance
Even though the new NVIDIA GPU, GeForce GTX TITAN X, will become available on the market tomorrow, some users have already laid their hands on this device and have conducted several synthetic tests that demonstrate this chip’s performance. The screenshots below display the benchmarks used for the comparison of the two NVIDIA TITAN X chips – we must not forget that the latest GPU bears the name of its predecessor but the difference between the two is significant – the new model uses the Pascal architecture, and not Maxwell.
This is why the new GPU, along with the updated software package cuDNN (CUDA Deep Neural Network), has demonstrated a difference of 200% in one of the synthetic tests, which says a lot about the chip’s raw performance. It is very likely that these results will appeal to those of you who are developers or work with large image files and videos.
Furthermore, there are some differences in terms of specs – the new TITAN X has 16% more CUDA cores and 42% higher base core clock. The image above displays the following results: an increase from 74% to 91% in Alexnet, from 76% to 200% in OverFear, from 74% to 84% in Inception, and from 91% to 98% in VCG. With this improvement of performance in mind, we can’t wait to see how TITAN X runs some games, although this is not entirely its purpose.
You can also browse through all currently available NVIDIA GPUs here: http://amzn.to/2a5jYXC