Hi Joe,
those numbers are your requirement? You can simulate that load quite easily with ham_bench (tools/ham_bench).
First create a database. You did not describe your keys and your records; the default key size is 16, the default record size is 1024. You can change that (including the key type) as described here:
https://github.com/cruppstahl/hamsterdb/wiki/Evaluating-and-Benchmarking
this inserts 50 GBs:
./ham_bench --stop-bytes=5000000000 --use-recovery --cache=1000000000
now say you want to open that file and perform 1 mio additional read/write operations, 50% read and 50% write:
./ham_bench --open --use-recovery --find-pct=50 --cache=1000000000
On my machine (a commodity system, but with an SSD) loading the database has on avg 65k ops/sec.
The second command has 260k inserts/sec and 100k find/sec.
Best regards
Christoph