Interesting benchmark performance (was RE: Pitiful benchmark perf ormance)

[Sean Hinde]

Hi,

> What would happen if you e.g. change INPUT_REDUCTIONS in erl_vm.h
> from (2* CONTEXT_REDS) to CONTEXT_REDS, doubling the I/O poll
> frequency?  (CONTEXT_REDS is set to 1000)
> 
> If basically all that happens in the erlang vm is that some 
> process is waiting for I/O, ports may not be polled with any
> predictably high frequency (then again, this is only guessing
> wildly. Perhaps it effectively gives the ports even higher
> priority...)

I didn't try this but did get the benchmark to spawn an additional process
to do this:

busy() ->
  busy().

The time to complete the benchmark came down to:

real    0m34.032s
user    0m22.190s
sys     0m11.640s

from:

real    0m45.845s
user    0m32.910s
sys     0m11.790s

while also completing some huge amount of additional work..

We have a number of machines working as pure mnesia servers running small
transactions - pretty much exactly what this benchmark is simulating. I will
do some more testing but it looks like in this sort of OLTP type
applications where there is little else going on in the emulator except -
read message from tcp/ip, do small work, wait for next message there might
be a useful boost in throughput available.

More thoughts anyone?

- Sean




NOTICE AND DISCLAIMER:
This email (including attachments) is confidential.  If you have received
this email in error please notify the sender immediately and delete this
email from your system without copying or disseminating it or placing any
reliance upon its contents.  We cannot accept liability for any breaches of
confidence arising through use of email.  Any opinions expressed in this
email (including attachments) are those of the author and do not necessarily
reflect our opinions.  We will not accept responsibility for any commitments
made by our employees outside the scope of our business.  We do not warrant
the accuracy or completeness of such information.