Hands-On GPU Computing with Python

Tesla platforms

Built primarily for GPGPU and stream processing, Tesla GPUs were originally launched in 2007, based on the Tesla microarchitecture named after the legendary electrical engineer and scientist, Nikola Tesla. Though the Tesla C-class GPUs were released with dual-link DVI outputs, the general norm was not having any output ports.

The Tesla V100 belongs to the NVIDIA Volta architecture. It comes in 32 GB and 16 GB HBM2 memory versions with a 1,380 MHz boost clock speed, 640 tensor cores, and 5,120 CUDA cores. As they are purely made with the purpose of accelerating computational applications, such GPUs are most suitable when displaying requirements are nil and computational tasks are the only priority. So, Tesla GPUs are specifically ...