Log entry time 16:28:13 on November 14, 2013

Entry number 385866

This entry is a followup to: 385863

Here is some performance info about the HV software.  I don't know what
this means yet.  The software is identical, but with some small diffs.

Server = computer where LecroyHV_Shim runs, talks to HV via serial
Client = computer where hvg Java GUI runs.

intelha3 is on L-HRS.  Intel PC, RH 5.8
ahut1 is on R-HRS.  Old laptop with RH 5.10 recently installed
halladaq8 is on R-HRS  Intel PC (slightly different model from intelha3)

There are 2 HV crates on R-HRS.  The upper crate has 14 cards and
the lower crate has 2 cards.  The performance is the same for both: bad.
I've tried turning one crate off to avoid any interference.

For all configs, the low-level Javier C code answers rather quickly,
I don't see any diffs there. 

HV CONFIGURATIONS for tests

Trial   Server  Client   Works ?   Speeds (sec)
                                   Crate ON/OFF   change a HV

1.   hacweb5    adaql1       yes    5 - 30       5 - 10

2.   intelha3   adaql1       yes    30 - 60      5 - 20

3.   ahut1      adaql1       no     

4.   halladaq8  adaql1       no

5.   halladaq8  halladaq8    no (using localhost for telnet connection)

-----------------------------------------------------------------

The performance of config 1 is probably normal and looks ok to me.  

Config 2 is a little worse than 1.  Some optimization was done by
adjusting delay parameters, which helped, but it's still slow.

Configs 3 - 5 are a total disaster at the moment.

The symptom for configs 3 - 5 is what was reported in halog 385863 (linked).
The Java code times out, throws an exception, and freezes.  
I will try to debug that with Roman P.'s help.

I thought it might be an issue with the network on R-HRS, but I believe
config 5 exonerates the network because client and server are both
on halladaq8.  It's somehow a "local" problem affecting BOTH crates.