[erlang-questions] Benchmarks

Jarrod Roberson <>
Tue Jan 12 22:08:28 CET 2010


On Tue, Jan 12, 2010 at 2:57 PM, Isaac Gouy <> wrote:

> On Tue, 12 Jan 2010, Henning Diedrich wrote:
>
> > "Performance Measurements of Threads in Java and Processes in Erlang"
>
> Does Jarrod Roberson's comment about those other benchmarks "they are
> irrelevant as none of those things are what Erlang is designed to do REALLY
> WELL" apply to this too?
>
> *irrelevant as none of those things are what Java is designed to do REALLY
> WELL*
>
>
Write some "benchmarks" that measure how long it takes to update a single
line of code in a running system.
Compare Erlang to C, Java and whatever else you want. I think the results
"will look somewhat bad" when comparing C or Java to Erlang. Now do the same
thing with distributed function calls across dozens of VM's and try
multi-threading those benchmarks in C or Java or doall the other things that
Erlang "does REALLY WELL". I am pretty confident that C and Java "will look
somewhat bad" as well.

I stand by my assertion that those benchmarks are not what Erlang was
designed to do REALLY WELL. And like Python and Lua, Erlang is designed to
do those things in C if needed very easily. Especially with the new NIF API.
That said, I do make my living writing Java code, but I don't think it is
appropriate for most of the things people try and do with it, Erlang is much
better suited for SOA than Java by a long shot.


More information about the erlang-questions mailing list