A big cache on an in-memory DB is a bit like insisting to sit inside a
row-boat while on a big ship. It has zero effect in helping you float
better - it's probably slightly worse even, considering the cache
computation cycles could have been avoided.
To get clarity, are you saying the 33% speedup is the gain of the
non-Indexed vs. Indexed table, or due to setting that cache size on the
already in-memory DB? (The latter would be worrying).
On 2017/12/18 4:48 PM, curmudgeon wrote:
You're definitely right about me wasting my time Simon. I loaded my entire
database (1 GB) into memory and set cache_size = -8192 for an 8 GB cache
size (using win64). I then ran my test (inserting the results of a query,
returning 111 million bigints, into a non-indexed single column table) and
there was no real difference. For lesser inserts (2 million) the speedup was
around 33% but would hardly be noticeable to the end user.
I daresay it would've been more noticeable if my laptop had a hard drive but
the moral of the story is get yourself an SSD and leave sqlite to take care
of the hard stuff.
--
Sent from: http://sqlite.1065341.n5.nabble.com/
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
