GPU to 72 status...

[bayanne]

Originally Posted by chalsall

Thank you for that. Let me please rephrase my statement...

The GPU72 et al swarms are currently focusing on a bit of clean-up. We're effectively up to at least 77 "bits" all the way up to ~125M.

Please let us know if anyone has any questions, comments, etc et al.

Seriously.

I am using TF on colab to do 33m exponents from 76 to 77. Could I ask if I could be allocated some exponents to 78, unless you wish me to continue with what I am doing.

[Uncwilly]

Originally Posted by bayanne

I am using TF on colab to do 33m exponents from 76 to 77.

Do you mean 332M? Because if you are and are using the notebook access keys that are available on GPU72, all you have to do is switch the work preference in your notebook settings.
Here are my 3 different keys with 3 different levels. 76->77, 77->78, 78->79

[bayanne]

Originally Posted by Uncwilly

Do you mean 332M? Because if you are and are using the notebook access keys that are available on GPU72, all you have to do is switch the work preference in your notebook settings.
Here are my 3 different keys with 3 different levels. 76->77, 77->78, 78->79

Not seen that before, so I have amended to the 77-78 setting, thanks

[bayanne]

How many clients per access key is it possible to run, without upsetting Colaboratory in any way?