this author has no idea about execution plan. He just checked statistical IO and did not analyse the execution plan.

Select min(field1) from partitionedtable1

The above statement on a partitioned table does not have any conditions and 100% will cause a clustered index scan on all partitions. That is why it is slow. What you need to do is to add condition to the query and also add appropriate index on field1(field1 can be in an index with multiple columns or included columns. The only thing you need to know is that field1 should be the first field in the index column list). The condition can be something like that field1 > -999999999(one value you know it is less than all field1 values).

In another word,

select min(field1)

from partitionedtable1

where field1 > -9999999999

Now, if you check the execution plan, it might become either index seek or clustered index seek(no appropriate index on field1) on all partitions. It should be much faster than before. Although this might still be slower than non-partitioned tables. Howver, when you write you query, if you can help optimzer to determine which partitions will be used at compiling time, you can achieve better query performance than non-partitioned tables.