unifi | Ubiquiti Unifi Go API | REST library

Raku doesn't have an ordered message queue. It has an unordered list of things that needs doing.

Your GTX770 GPU is a "Kepler" architecture compute capability 3.0 device. These devices were deprecated during the CUDA 10 release cycle and support for them dropped from CUDA 11.0 onwards

The CUDA 10.2 release is the last toolkit with support for compute 3.0 devices. You will not be able to make CUDA 11.0 or newer work with your GPU. The query and bandwidth tests use APIs which don't attempt to run code on your GPU, that is why they work where any other example will not work.

1. For the table problem, I changed any
into $'\r\n' accordingly.
2. When collecting info using the GH-CLI, it will come back as JSON and as such, you need to use jq to query and get the value itself. On the same note, the value will be with double quotes and therefore I used substitution operators as such:

The APIs are a bit confusing right now. But the community is working on removing this legacy and unifying the APIs.

In the early days, Flink started as a batch processor with a streaming runtime under the hood. So the DataSet API with ExecutionEnvironment was exposed for batch processing.

(The DataSet API is reaching end-of-life and will be deprecated soon.)

Later, Flink exposed the streaming runtime via DataStream API with StreamExecutionEnvironment. This is one of the main APIs today. Its vision is to work on unbounded and bounded streams. Since batch processing is only a special case of streaming, it can be categorized under bounded stream processing.

The Table API with TableEnvironment was the first API that fully unified the both worlds in one API. The DataStream API is currently receiving a batch execution mode for faster processing of bounded streams. All of that is already present in Table API.

Keep plot as it is after setting groupnorm='percent', but only display individual values in hoverinfo.

1. Include a duplicate column in your df for value with df['actual] = df['value']
2. Include hover_data = ['value', 'actual'] in px.area()
3. Change fig.update_traces(hovertemplate='%{y}') to fig.update_traces(hovertemplate='%{customdata}')

The reason why you'll have to include a duplicate column is that px.area() automatically calculates percentages for the column assigned to y in px.area(df, x='date', y='value'.... When setting hover_data = ['value', 'actual'], no calculations are done for actual which is later accesible in the hoverinfor through fig.update_traces(hovertemplate='%{customdata}').

If you drop fig.update_traces(hovertemplate = ...) from your setup, you'll get the following hoverinfo which may also be of interest:

In this case it might make more sense to change value to percent, which after all is displayed:

However, if I understand it correctly, in order to put the image on the GPU and have it processed in parallel, I have to copy it to unified memory or some other GPU-accessible memory

You understand correctly.

I am wondering whether there is a more efficient way to copy an image (i.e. a 1D or 2D array) on the GPU that does not have an overhead

There isn't. Data in host system memory must pass over the PCIE bus to get to GPU memory. This is bound by the PCIE bus bandwidth (~12GB/s for PCIE Gen3) and also has some "fixed overhead" associated with it, at least on the order of a few microseconds per transfer, such that very small transfers appear to be worse off from a performance (bytes/s) perspective.

such that the parallelization is useless.

If the only operation you want to perform is take an image and multiply each pixel by a value, and the image is not already on the GPU for some reason, nobody in their right mind would use a GPU for that (except maybe for learning purposes). You need to find more involved work for the GPU to do, before the performance starts to become interesting

The operation is quite simple

That's generally not a good indicator for a performance benefit from GPU acceleration.

You can use modes, like this (Added some elements to make the wanted stages clear):

I found an answer to my problem!

To set an entry point to a .py file on a pex file you use the -c command!

e.g. pex . -r requirements.txt -c main.py -o test.pex

The official documentation is really complicated and and clunky!

Here is a blog that in great detail describes the steps you need to take to succesfully create a pex file!

https://www.shearn89.com/2021/04/15/pex-file-creation

This helped me a lot and I hope this helps you as well!

This answer led me to the solution.

X509Certificate2.PrivateKey was throwing a NotSupportedException because I was using ECD as the signature algorithm. I switched to RSA and now it properly sends the certificate.

What's strange is that tracing showed no issue with the ECD cert, and it was able to successfully identify the private key. I have no idea why this is the case -- it sounds like a bug to me. Nevertheless, RSA will work for my use case and I am tired of debugging this.