[Simh] 10ms delay on ethernet communications - the solution!

[Boucher, François]

Hi to all!

 

I did some more search and it appears that Mr Robert Supnik did modify in Oct 2006 the

files PDP11/pdp11_xq.c and pdp11_xu.c to synchronize the polling of the Ethernet interface with

the timer clock.

 

This modification makes the polling rate totally unused.  Even if one sets the polling rate to another value than the default of 100 polls per second, it stays at the default rate.  This change of behaviour is not documented.  Further, if one do a set xq poll=1300 and then a show xq, the output indicates correctly the new polling rate, but that rate is simply not taken into account, and simh continues to poll the reception of Ethernet packets at it's default rate of 100 polls per second.

 

To fall back to the previous setting, I did copy those two files from the sources of version 3.6-1 of simh, into the same files of version 3.7-3, and then recompile.

 

After that, I did some tests, and it appears that for 100Mbps Ethernet, the optimal polling rate on our setup is of 1300 polls per second.

When I do a monitor system, I obtain a 2300 buffered IO rate per second, from a 800 buffered I/O rate per second.

That modification and tuning of 1300 polls per second gave us a 300% increase in Ethernet throughput!

Also, I did see by running PT_VAX.EXE that there is a slight performance penalty when I rise at 1300 polls per second the rate, but that penalty is of the order of passing from 25.0 VUPS to 23.9 VUPS, for the performance gain of 300% Ethernet I/O.

 

So, for my part, this case is closed, but I would appreciate any comments about further optimisation of the Ethernet driving of simh.

 

Thanks,

 

Francois Boucher

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20080110/15a7f2f5/attachment-0003.html>