I have to say that, for the most part, it's all relative. It truely depends on
the operations you're doing on the table.

if you have a table with 1.6 million rows and you try and insert to it, you should
see no hit in the speed. I've seen my server hit upwards of 200 inserts per second
with no negative impact.

On the other hand, if you're doing massive select left join on non indexed tables.
You will see it slow down as the size grows. Again though, it's relative. It's still
going to be faster then anything DBD::CSV could do. Though normal basic selects should be ok.

just as an example.

i have one table called "raw_test_results".

It's currently 5.8 gigs of data, and 1.8 gigs of indexs (7.6GB total size).
That's 43,973,656 rows x 36 cols.

I currenty insert to it at 64 inserts per second. However, crunching the numbers
down with a leftjoin takes well over a minute. When the project started it would take well under a second.