Speed up data analytics with GPUs quiz

1. What open-source library does cuDF accelerate? (Select all that apply)

   Choose as many answers as you see fit.

   pandas

   Polars

   Apache Spark

   Oracle

2. Before importing cuDF in a Colab Enterprise notebook, what configuration step must you take regarding the Runtime?

   Increase the disk size to 500GB or more

   Select a Runtime specifically equipped with a GPU (e.g. L4)

   Enable "High-RAM" mode on a standard CPU runtime

   Mmount Google Drive to store the temporary CUDA kernels

3. While attempting to run cuDF in Colab Enterprise, you receive an error indicating "no CUDA-capable device is detected," what is the most likely cause?

   The dataset is too small to require a GPU

   Google Cloud billing is currently disabled

   You have not installed standard pandas

   The notebook runtime does not have a GPU

4. GPU-accelerating pandas requires zero code changes when using NVIDIA cuDF

   True

   False

5. When using Colab Enterprise with NVIDIA cuDF, where does df.read_csv('/content/data.csv') primarily load the data to?

   Google Cloud Storage buckets

   The virtual machine's CPU RAM

   The attached GPU's VRAM

   Your local laptop's hard drive

6. Which of the following are valid methods to activate cuDF pandas acceleration? (Select all that apply)

   Choose as many answers as you see fit.

   Using the magic command %load_ext cudf.pandas in a notebook

   Renaming your .py file to .cu_py

   Adding import cudf.pandas; cudf.pandas.install() prior to importing pandas

   Running a script via python -m cudf.pandas myscript.py

7. When using cudf.pandas, what happens if a specific pandas function is not yet GPU-accelerated?

   Graceful fall back to CPU execution

   NotImplementedError is raised and execution is halted

   Function is ignored and execution continues

Submit answers