BrunoScience / BrunoScience.github.io

The best that most of us can hope to achieve in science is simply to misunderstand at a deeper level.
https://BrunoScience.github.io
Apache License 2.0
0 stars 1 forks source link

Rentable Cloud Compute Partners and NVIDIA vGPU or Launchpad #20

Open MarkBruns opened 1 year ago

MarkBruns commented 1 year ago

TPUs are probably not worth learning ... it's like TensorFlow vs PyTorch ... nobody can base a career or a business on expecting continually Google support ... same is true for GCP; which may explain why Oracle Cloud, Azure and AWS are not especially afraid of GCP ... it's just like any other acquisition because Googlers lose interest after a few years ... but having said that, Google will have some TEMPORARILY very interesting offerings -- even if one should not expect the Googlers to be available or the product to be well-supported ... but that doesn't make it a bad deal -- it's just not going to be the ONLY game in town ... https://www.cloudskillsboost.google/ ... https://codelabs.developers.google.com/ ... https://www.kaggle.com/

NVIDIA has a strong interest in supporting their products for a longer-time horizon... that includes a degree of backward compatibility after new generations of more powerful hardware come along ... https://www.nvidia.com/en-us/data-center/virtualization/it-management/#product ... https://www.nvidia.com/en-us/launchpad/ ... of course, what's really interesting is the NVIDIA partner ecosystem ... a LOT of companies are exploring different ways to add value to NVIDIA's GPUs ... https://www.nvidia.com/en-us/about-nvidia/partners/partner-locator/

https://www.paperspace.com/gradient/pricing

https://lambdalabs.com/service/gpu-cloud#pricing

https://www.coreweave.com/gpu-cloud-pricing

https://saturncloud.io/ ... provider of resources to hackathons, eg https://openproblems.bio/neurips_docs/about/explore/

https://www.cirrascale.com/cirrascale-cloud-pricing.php

https://www.ironorbit.com/

https://www.def.com/about.php