Hacker News new | past | comments | ask | show | jobs | submit login

>(albeit waay slower than MySQL in the past).

That isn't even remotely close to true. Single threaded "insert 10,000 rows" benchmarks had postgresql slightly slower, not "waay slower". Concurrent access benchmarks have always had postgresql faster than mysql.




I'm old. When I say in the past, I mean pgsql 4 ;-)


I'm old. I know that there was no postgresql 4. The first release of postgresql was 6, and what I said was accurate as of the release of version 6, in 1997 or 1998.


[slow bow, with flamboyant hat salute] :)

I never expected the very geek joke to be understood. Very cool!

On a more serious note, I didn't mean to support the ever old myth that postgresql is slower than mysql. It hasn't been the case for over ten years, and all the while being a better database in every aspect.

However, go back enough and it was slower (than mysql) for real-world loads. You could get near enough if you disabled fsync, but then were negating many of the advantages of pgsql. Even then, in the old days, table based locking would kill a real database. Table based locking was solved early enough ('98? '99? around that time). Fsync was solved with WAL, in the performance effort of the early 7.x series. It was around late '2000 that I came back to postgresql for good. Until then, it was either mysql for a quick hack or oracle for anything serious. Since then, it's postgresql for everything (I don't deal with the scenarios where Oracle still leads).




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: