GPSOut+PocketFMS on same comp cause stutter

[Tord Hoppe]

Hi Peter and the rest.

I´m using the GPSOut module along with the thingamajig that connects two COM ports so that I can run PocketFMS on the same computer as MSFS. Everything works just fine but there is one little thing that I find nagging.

Connecting PocketFMS to the GPS(Out) signal causes a stutter in MSFS everytime a GPS "packet" is sent. I´ve tried different values for Interval (if I remember correctly) and the stutter seems to correlate, i.e. Interval=1000 results in 1 second intervals between the stutter, Interval=500 in half second and so on.

Funnily this only occurs when I click on the PocketFMS button that tells PFMS to receive a GPS signal. I suspect that GPSOut constantly streams the "signal" and that indicates that it is PFMS that cause the stutter, perhaps an intermittent high processor load since the GPS "burst" means a map refresh in PFMS? Your thoughts? I´m going to experimen´t with different Interval and Baud rates in GPSOut to see what happens...

Another funny little thing is that regardless of what baud rate I´ve set in the GPSOut.ini PocketFMS detects something like 38400 (not sure, at work right now).

This is no big issue, just wanted to check if the expert has some thoughts as to what might cause the stutter.

[Pete Dowson]

I´m using the GPSOut module along with the thingamajig that connects two COM ports so that I can run PocketFMS on the same computer as MSFS.

I hope you are using a very powerful PC or one with HyperThreading, as running another separate graphics program alongside Fs in the same Pc is asking for jerky behaviour otherwise.

Everything works just fine but there is one little thing that I find nagging.

Connecting PocketFMS to the GPS(Out) signal causes a stutter in MSFS everytime a GPS "packet" is sent. I´ve tried different values for Interval (if I remember correctly) and the stutter seems to correlate, i.e. Interval=1000 results in 1 second intervals between the stutter, Interval=500 in half second and so on.

Sounds like the PocketFMS program is not so good at using overlapping I/O and is hogging the processor which reading each byte of the incoming message.

Funnily this only occurs when I click on the PocketFMS button that tells PFMS to receive a GPS signal. I suspect that GPSOut constantly streams the "signal" and that indicates that it is PFMS that cause the stutter, perhaps an intermittent high processor load since the GPS "burst" means a map refresh in PFMS?

It is either bad programming on the COM port end, as just mentioned, or, yes, a graphics thing. For the former try (1) the highest possible speed accepted in the PocketFMS settings and (2) the minimum number of NMEA sentences which work to give the data it needs.

For the latter I'm afraid I've no solution otherthat a very fast video card and dual processor or hyperthreading.

Regards

Pete

[Tord Hoppe]

Hi Peter and thanks for your reply. I´ve tried upping the transfer speeds and still get the same kind of stutter, albeit for shorter periods which seems to match the shorter data bursts. Since this does not occur when I instead feed my PDA via the USB port I think we can safely say that the graphics update in PocketFMS is the "culprit".

Hope you had a nice vacation, bye!

[Pete Dowson]

I´ve tried upping the transfer speeds and still get the same kind of stutter, albeit for shorter periods which seems to match the shorter data bursts. Since this does not occur when I instead feed my PDA via the USB port I think we can safely say that the graphics update in PocketFMS is the "culprit".

On the contrary, if the stutters are shorter when the transfer speed is higher, that indicates that the programming in PocketFMS is rather assuming no other processes are running when it is receiving data. It should be using overlapped I/O which allows the data to accumulate independently of program action instead of looping for each character without relinquishing control.

If the stutter was due to graphics updates then they would be the same no matter what speed of transfer were used, as the update of the graphics would be independent of that.

Regards,

Pete