New major release of Parany (v14.0.0)

Community
,
1 
# opam search parany
[...]
parany 14.0.0      Parallelize any computation

I switched back the runtime to fork+marshal.
As was the case before version 13.0.0.
Versions 13.* (up to 14.0.0 excluded) use ocaml5-threads to parallelize computations.

There is a branch GitHub - UnixJunkie/parany at ocaml5_threads
if anybody wants to maintain this backend.

The Parany.Parmap module has new functions for arrays:

Regards,
F.

4 Likes
2

Can you tell us more about your reason to switch back to fork+marshal?

3 Likes
3

I want optimal parallel performance for the kind of workloads I encounter daily at work.

4

I haven’t gone through the repository in great detail - curious if you have any numbers comparing fork+marshal against OCaml 5 domains?

5 Likes
5

The only public benchmark I have is this one:

My advice: all ocaml5 benchmarks should be ported to fork+marshal, in order to detect
when and by how much ocaml5-threads are behind in terms of parallel performance.

6

Are you able to release the dataset so that we can reproduce your results? Otherwise, I’m unable to see the utility of your advice.

It is certainly possible that fork+marshal is faster for certain programs. But a generalised advice of “port all your ocaml 5 programs to fork+Marshall” is not useful. The advice has to be more specifc such as “If your program has a pattern X, then fork+marshal may offer better performance”. In your case, the work is embarrassingly parallel, and I assume fork+marshal works better (I presume).

2 Likes
7

I thought there are plenty of ocaml5 parallel benchmarks out there.

Was it in there?

I guess some of the ocaml5-threads parallel programs in there could be ported to fork+marshal
and have some parallelization performance comparison.
Since I contributed the Gram matrix benchmark in there, I could have a look at this one
when I have some time.