Arnab Phani

Thanks for this extension @Shafaq-Siddiqi. This crashes because FP32 and INT32 are missing in the exact serialized size calculation logic (`FrameBlock.java::getExactSerializedSize()`). I will investigate it further.

Nope. It is a duplicate. Safe to close.

I will look into this @BACtaki. Meanwhile, please check why the tests are failing.

Can you please execute the `GPUFullReuseTest` and `GPULineageCacheEvictionTest` tests under the Lineage package as well with the updated CUDA version? You need to uncomment the code in `TestUtils.java:isGPUAvailable()` to trigger...

There are two versions of T2 (KDD dataset) -- one with UDF-based scaling and another with scaling outside of transformencode. The UDF task implementation is still in the experimental phase....

Thanks @louislepage, for starting the project. Please add the Apache license to the new file, DenseBlockTrueBool.java. Looking forward to seeing how you use this new denseblock for the binary matrix...

Thanks for the commits @louislepage. I have a few questions regarding the experiment setup. How are you running and benchmarking the experiments? Are you running a dml script with time()?...

Measuring time in Java is fine. However, I recommend using `System.nanoTime()` instead. Use boolean/bitset inputs of different sizes.

> Can you restart the pipeline once? I think a couple of errors are due to issues with spark and caching and might not be directly code related. > >...

Thanks @louislepage. The plots conclude this task. Thanks for your contribution. 👍🏽 If there are any scripts that are needed to reproduce the plots, please share those with us.