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

Re: BMI TS4 Tick Data Test



PureBytes Links

Trading Reference Links

On Saturday, June 24, 2000 5:08 PM Bob Fulks wrote:
>
>I think your conclusions are correct if your system is doing nothing else.
But if you are using the computer for anything else you risk getting buffer
overflows when the system is busy with other things.
>

Yes, a qualitative viewpoint indicates that a buffer can certainly help in
certain situations.  My point is that even if you can save 100% of the data
in the BMI data stream using a buffer you can't retrieve the 10% or so that
BMI had to trim from the initial data stream.  So, unless data losses above
2% are being experienced a buffer is probably not warranted.

Also, a buffer is a dual edged sword as it stores data for later retrieval
it is also adding lag.  I think you and I are pretty much in agreement that
you can have "too much" buffer.  Maybe two seconds lag is acceptable which
would indicate a maximum buffer size of 8K - anything more than that is
probably not beneficial.  Eventually, the PC must process the buffered data
and if a lot of data is being lost maybe a better solution is lightening the
PC load or increasing CPU speed/RAM size.  In addition to the BMI satellite
feed I subscribe to QCharts and a few months back I noticed that my data was
getting increasingly delayed (15 - 30 seconds).  The delay was determined to
be caused by the PC being too overloaded for the task and the data was
backing up in a software buffer.  The answer was to double both CPU speed
and RAM size.

Bob