[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CommsTime times?

I have some up-to-date benchmark information for the libcsp CSP-on-Posix-threads library. Not surprisingly, these
figures are less impressive than the more direct approaches for implementing CSP (notable Kroc, CCSP and CTC++). That is
to be expected because libcsp provides generality rather than performance.

The following benchmark run was on Solaris 2.7 on an Sparc Ultra 10 250MHz:
> time ./commstime 1000000 1
Finished 1000000 loops.
Time per loop = 645us
Context switch = 80.6us

real    10m45.464s
user    5m15.820s
sys     5m27.940s

The same machine running occam (Kroc 0.9b with Peter Welch's benchmark) for 10E7 iterations:
Last value received = 10000015
Time = 12959496 microsecs
Time per loop = 1295 nanosecs
Context switch = 161 nanosecs

The Kroc version is 500 times faster per loop.

Both cases are for the sequential delta, rather than the parallel delta.
Versions: libcsp v0.4 (which has an improved channel algorithm compared to the version reported at the CPA2000
conference in September).
Kroc 0.9b (Solaris) (the 1.2.3 version doesn't appear on the website yet - any chance of this happening soon?)


See http://www.beton.freeserve.co.uk/libcsp/
Libcsp is covered by the LGPL licence.

tel;pager:ICQ: 56840977
tel;cell:MSN/Hotmail: richardbeton@xxxxxxxxxxx
tel;fax:01794 833434
tel;work:01794 833458
org:Roke Manor Research Limited;Internet Technology & Networks
adr:;;Roke Manor: http://www.roke.co.uk/;;;SO51 0ZN;UK
title:Internet Consultant
note;quoted-printable:The information contained in this e-mail is confidential and must =0D=0Anot be passed to any third party without permission. This =0D=0Acommunication is for information only and shall not create =0D=0Aor change any contractual relationship. =0D=0A
fn:Rick Beton