I've been oogling the Mac Pro from Apple with loaded specs. Check it here if unfamiliar.

I'm curious to hear anyone's thoughts of the computer for deep learning/machine learning applications vs cloud based computing. Obviously, a flexible solution on GCP or AWS has the ability to scale and therefore produce results faster. However, I'm mostly interested to hear thoughts on the economics of purchasing the machine vs renting a cloud based machine.

If we assumed something like a 5 year lifetime of the Apple Mac Pro for this type of computing (before the chips/processors become immensely slower than the newest chips used in the cloud), that would be like $1000 per year amortized using the Mac Pro. Does it make more sense to rent from AWS/GCP for \$1000 per year as a renter instead? I don't have much experience in the cost of AWS/GCP so I'm looking to here an answer from anyone well versed in cloud computing.


The Mac Pro uses AMD GPUs. These don't support CUDA, but instead support the OpenCL framework.

TensorFlow, the most popular deep learning library, uses CUDA to run on GPUs, although OpenCL support is in the works. That said, one of the main approaches to OpenCL support, SYCL, isn't planning to support OSX:

We have no plans to support OSX in near future. I know sad face.

The current advice is to avoid AMD GPUs:

Currently, if you want to do DL and want to avoid major headaches, choose Nvidia.

| improve this answer | |
  • $\begingroup$ Thank you so much for this advice! VERY useful information! $\endgroup$ – CaptainPlanet Jul 15 '18 at 20:03

Not the answer you're looking for? Browse other questions tagged or ask your own question.