pwr | : zap : The interactive CLI for npm

Unless you do something with the vaiable timer (such print it out) then the compiler can tell that you don't need it and saves time by not really accessing it.

If the reason that you are using it is to waste some time, try declaring it volatile int timer.

Also, note that you are wasting time between setting and clearing PA5, but not between clearing and setting. Maybe the jump instruction to go back to the start of the loop will waste some time, but maybe this will not be where you expect it.

Extract the substring between two strings:

You have to swap out keys in the dictionary object, since there is no such thing as "changing" an immutable key. One way to do that is to make an independent list:

You can use re module to parse the string (regex101):

You can calculate the sample size for a given power with the same pwr.t.test function.

From help(pwr.t.test):

Exactly one of the parameters 'd','n','power' and 'sig.level' must be passed as NULL, and that parameter is determined from the others.

There are more efficient calculations that could be done with base rle or data.table, but this tidyverse approach is legible to me and should be adequate for data that isn't extremely large. In my testing, the code below for 100,000 observations took about 5 seconds to run.

My approach is to make a function that finds the cumulative min for any given window size (using slider::slide_index so that the window is defined by time, not a fixed number of observations), and then output the max of these. Then I iteratively feed a range of window sizes into purrr::map_dbl to get the max_min for that size.

The approach is correct, as long as the GPUs are on the same host. The TensorFlow manual has examples how the tf.distribute.MirroredStrategy can be used with keras to train the MNIST set.

No, there are multiple strategies that can be used to acheive the workload distribution. For example, the tf.distribute.MultiWorkerMirroredStrategy can also be used to distribute the work on multiple devices trough multiple workers.

The TF documentation explains the strategies, the limitations associated with the strategies and provides some examples to help kick-start the work.

According to the issue from github, the ValueError: SyncOnReadVariable does not support 'assign_add' ..... is a bug in TensorFlow which is fixed in TF 2.4

You can try to upgrade the tensorflow libraries by

Thanks to @HernánAlarcón suggestion, I tried like this and it worked like charm:

As @talonmies pointed out, I was misunderstanding the versioning system. However, because it's a System76 machine, it was also confounding because System76 uses their own Nvidia driver, and it's not straightforward to install Cuda 11 and Cudnn. I'm posting the answer in case anyone else runs into problems with System76.

First, DO NOT use the System76 install for Cuda and Cudnn. They have their own versions (on their website) so as to be compatible with their Nvidia driver, but they will not work (they are version 10, and TF 2.2+ requires 11). Also, most general Cuda guides will tell you to uninstall/install the Nvida driver first so as to have a clean install, but DO NOT do this if you have a System76 system. Just leave the System76 driver alone. Also, if you have any previous Cuda/Cudnn remove/uninstall all of it.

Go to Nvidia and get their latest Cuda and Cudnn. I used