[erlang-questions] Shared memory and message passing

Ulf Wiger (TN/EAB) ulf.wiger@REDACTED
Thu Sep 7 15:43:22 CEST 2006


For those writing papers out there: it's ok to 
say that Erlang has been used to develop applications
with more than a million lines of Erlang code.

("Several hundred thousand lines of code" sounds to
me as "at least more than 200,000". That hardly even
registers as "large" in my book anymore. ;-)

Also, the number of processes can be way larger than
50,000. We consider 30-50,000 processes to be a
fairly normal number.

One could perhaps note while comparing erlang 
programs in different execution environments, that
just about all erlang programs today are written 
with the assumption that message passing involves
copying. Therefore, it should be expected that 
the amount of message passing - and the size of 
messages actually sent - be kept reasonably low.

Another thought (apologies if I just missed it in
my cursory read of the paper) is that program 
analysis would have difficulty handling common
constructs like

InterfaceModule:f(X) -> 
  ImplementationModule:f(X,...) ->
     ... gen_server:call(..., y(X))

when doing module-local analysis.

The hybrid model might reward more explicit use of 
message passing (a bit in line with Joe's !! syntax)
and also encourage people to use processes and 
message passing much more than today. 

I can see already how lifting the limit on how 
many processes one may use (from the stifling 32000
to millions), and also optimizing the runtime so that
spawning, linking and message passing are already 
quite cheap, is changing how people go about structuring
their programs.

BR,
Ulf W

> -----Original Message-----
> From: erlang-questions-bounces@REDACTED 
> [mailto:erlang-questions-bounces@REDACTED] On Behalf Of 
> David Hopwood
> Sent: den 6 september 2006 16:41
> To: Erlang/OTP discussions
> Subject: Re: [erlang-questions] Shared memory and message passing
> 
> Richard Carlsson wrote:
> > PS. Shameless plug: see the July issue of TOPLAS (vol 28, 
> issue 4) for 
> > all the gruelling details on the hybrid architecture.
> 
> On-line at <http://user.it.uu.se/~kostis/Papers/toplas06.pdf>.
> 
> --
> David Hopwood <david.nospam.hopwood@REDACTED>
> 
> 
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://www.erlang.org/mailman/listinfo/erlang-questions
> 




More information about the erlang-questions mailing list