We’re getting 90 TFLOPS on AWS V100 for deep learning (using dlib not Tensorflow), so I can’t understand this comment:
> With TensorCores the Titan V has a new shiny deep learning feature, but at the same time, its cost/performance ratio is abysmal.
> With TensorCores the Titan V has a new shiny deep learning feature, but at the same time, its cost/performance ratio is abysmal.
Is it really X times faster to justify X times price tag?
Do we need significant software changes to take advantage of the new power? Are the TFLOPs somehow not directly comparable?