TileDN arrays: Dense and sparse arrays

https://docs.tiledb.io/docs/write-global-one-write

https://docs.tiledb.io/docs/sparse-write-global-one-write

The array sizes of both examples are trivial (i.e., 4 * 4), you should increase it to larger values so that the physical memory wouldn’t hold it. We’d like to know the times taken for the following arrays in TileDB:

(i) 100% cells are non-empty values;

(ii) only the first 10% of consecutive cells are non-empty values;

(iii) only the first 50% of consecutive cells are non-empty values;

(iv) only the last 10% of consecutive cells are non-empty values; and

(v) 10% of random cells are non-empty values. (for simplicity, you can distribute the non-empty cells evenly within the array)

In the report, I expect to see the following:

1) How the experiment is setup (i.e., system specification, experiment parameters, etc.). Hint: check some published papers and see how they present this

2) Results comparison, ideally a plot from Excel or Gnuplot

3) Explain the results, e.g., why filling up 10% random cells takes longer (or shorter) time than 10% consecutive cells.

We can install TileDB with the following document attached

It should be done on Linux Virtual Machine(Ubuntu) with the Username as “ananth”