GPU computing with Cuda
We are accelerating our development with GPU computing.
The latest workstation has five P100 Nvidia GPUs and three GTX 1080 Ti GPUs.
AI needs plenty of time to train.
We are parallelizing the algorithms on multi-servers for efficiency.
Gonsalves Lab on Sophia University Ichigaya Campus (209A).
Each of the desktop machines is taylor made either with Intel i7 or i9 CPUs and GPUs connected in SLI mode. The systems are water-cooled for efficiency and noise control.