Increase the default value for MinPartitionSize variable

[dchigarev]

'32' seems like a very small value for such a variable.

I personally doubt that we would actually need 4 partitions for a dataframe with the shape of (nrows=256, columns=4) as the overhead of the execution engine will likely overdo all the benefits we can get from parallelization on such small data:

modin/modin/config/envvars.py

Lines 602 to 611 in d6a6cf3

	class MinPartitionSize(EnvironmentVariable, type=int):
	"""
	Minimum number of rows/columns in a single pandas partition split.
	Once a partition for a pandas dataframe has more than this many elements,
	Modin adds another partition.
	"""
	varname = "MODIN_MIN_PARTITION_SIZE"
	default = 32

We should run some tests and reconsider the default value for this

[anmyachev]

@dchigarev it seems that at first it makes sense to increase the size only by row. However, now changing this parameter will also affect the data columns. Here, apparently, it will also be necessary to revise the approach of minimal square partitions towards rectangular ones (a set of values describing two dimensions).

UPD: #7284