Benchmark of marshalling performance

[Sylvain Joyeux]

I had problems w.r.t. performance of Marshal.dump in one of my
applications, so I tested if it was *that* better to dump objects by sets
instead of one by one. The results are here:

http://www.jhaampe.org/software/marshal-perf-ruby

 

[Robert Dober]

I had problems w.r.t. performance of Marshal.dump in one of my
applications, so I tested if it was *that* better to dump objects by sets
instead of one by one. The results are here:

http://www.jhaampe.org/software/marshal-perf-ruby

Thanks for sharing this, it is also very comforting to know that for
an application that does few things but *fast* Ruby still seems fast
enough

Salut
Robert

 

[Sylvain Joyeux]

Thanks for sharing this, it is also very comforting to know that for

an application that does few things but *fast* Ruby still seems fast
enough

Lol. The only real problem I have with Ruby's performance is its GC ... I
have cycles that should be around 100ms and the GC will sometime take
ages.