Live View

Looking fwd to try and experiment :-)

Test does not work here (but I am sure I am doing something wrong)

Nothing is moving after er press start.

All network connection are established. Opus indicates green.
I created one default window on the server for the Clent so far.
C172 Left 90 degrees view (lookng out the left window woth VC visible)
If I start the flight (So no test running) Everything works fine, both server and Client are moving (but client with stutters as before of course)

Mm each time I restart FSX and Opus on client and server I have to start of with only one axis.
Sometimes only, sometimes only bank.
Thereafter everything works as it should and all axis can be tested simultanuously.

Anyway, I can set the update rate to max 1dot right of the default (middle)
Anything more and I have lag when stopping.

At this one dot right update rate I have light stutters all the time and severe stutters (attitude changes) every oh 10 sec or so.
Clients Spy window sais 100+ position updates.

Everything (sliders) is reduced as far as possible on the client.
I have tried unlimited and 30fp on the client, same result.
NvidiaInspector is set to default, so no creazy AA filtering etc on the client.

The server runs with unlimited frames and very high scenery settings.

I have corrected the problem with the tick boxes initialisation.

Firstly the one notch right of centre is fine, that's what I get as well, except my movement on the client is fairly smooth with no major stutters at all. It remains smooth with all three attitude boxes ticked.

I will post the beta including the client side LV test tomorrow (hopefully). That will definitely answer whether its your client or network causing the stutters, my money is on the client.

Stephen

The announcements sais version 2.82.1 with Client test is now available.

But when I download from the Beta link I get version 2.81.6

We will soon be deleting the option at the top of the page. The latest beta can only be downloaded using the Beta button and the zip option underneath. My fault, I keep forgetting about the hyperlink at the top of the page. That will still be downloading the previous beta most probably. Sorry about that ...

Regards
Stephen

Ok. Small stutters, but much better than I have seen so far.

At 30fps fixed small stutters but no abrubt attitude changes. Quite usuable!
The faster I go with the update rate, the faster the C172 Banks/pitches/yaws and the more difficult it becomes to detect stutters, but they are still there.

At unlimited fps the test is even smoother. almost no stutters!

60fps stutters worse than 30fps at default update rate (middle) but it does just as good as unlimited when I set the update rate one notch to the right of center.

So the conclusion is, my Client can run smooth.
Is it my network then?
Or my server?

Both server and client at unlimited framerate.

Server test results in stutters at client.
When the most severe attitude changes occur at the client, I can see a major leap from 55-60%CPU avarage usage to 100% in clients task manager (CPU 1 jumps from 30% to 100% while CPU 0 is allways at 100%.)

let me try putting FSX affinity to core 0 only on the client!

Well,...just flew around a bit.
Setting FSX to core 0 only on the clent did get rid of the most severe stutters associated with the sudden load increase on core 1.
But the other smaller (but still quite annoying) stutters are still there.

Yes Client test was smooth with no affinity applied.

I bought a 10/100/1000 network switch so that should not be a problem anymore either.
Ofcourse I understand very little of networking so determaining a problem there is verry difficult for me.

One more thing I can try is running the client without overclock in case it is not stable.

Not really. Have you tried bypassing the switch and cabling back to back with a cross over cable?

Stephen

No Difference with cross over cable between the two PCs direct.

Also unclocked or clocked Client does not matter.

I did another Client test and noticed the same CPU peaks (task manager) and stutters this time as during a flight.
Will do some more tests, but I really think you were right all along, and it is the dual core that is too old and slow.

So, since you have tried everything you can, and I am close to accepting this will not work,

I have a 1year old laptop I could try (it has HDMI out that I can connect to a monitor)

Do you think an Intel i5-2430@2.4Ghz, 4 GB DDR2 (I would be willing to invest into some more RAM) with an AMD Radeon HD6630m GPU would be up to the task?

ok.

thx for all the assistance!

PS if you ever make an upgrade for cockpit flight instruments, could you let me know (PM perhaps)?

Carrying out the client side LV Test on your server would at least tell you if your server spec is acceptable as a client, I believe that is quad core single threaded. Don't bother changing the FSX settings, just try it and see, of course it would be even better with the settings changed and all the other client performance related mods.

But again, regarding your instruments I don't mind giving that a go. You could at least use you dual core client for some instruments.

Stephen

Hello

OK, I still haven't finished with the client tests. I have prepared a modified FSXCLIENT_TEST.EXE program for you to try. This modified version disables the second IPC link, which in turn reduces the number of threads active and also allows Windows to service the I/O requests of just one IPC link. My hope is this will reduce the burden on the limited available cores and allow the client FSX to run more smoothly on your systems.

Ok test done on the client.

The first test seemed better as before.
exit FSX and OPUS.
Start FSX and OPUS
Tried a flight and it stuttered as allways. Best result at 60fps on server and unlimited fps on client.
Less than 60fps on server = less updates (spy window) = stutters
More than 60fps on server = lag on client.
Exit FSX and OPUS.
start FSX and OPS.
did a second test on client
more stutters than during first test.
only at default (middle) update rate smooth movement.

Also, at default update rate the aircraft moves gently (nice and slow).
One notch to the right and the aircraft moves quickly.
At 2 - 5 notches to the right I see no difference, it moves just as fast (and not smooth) as at 1 notch right!

Try just the client LV Test and compare with the previous version. Is the new FSXCLIENT_TEST program giving a smooth movement or no different to the previous version?

Stephen

Ok many thanks, it was worth a try to see if reducing the number of active threads an IPC network links would improve things. It was just one of these 'sleep on it ideas' that I hoped would lead the way to a solution for your system.

Stephen

Not really, but it will be FSX (or the a/c sim) trying to calculate the vertical speed based on the changing altitude parameter during the updates. Its one of these ambient and parameter calculations. So it could just depend on the frequency and smoothness of these calculations. In both cases, FSX is not 'flying' the aircraft just responding to the updates.

The aircraft sim can also makes a difference, some just can't let go even when the sim is placed in 'freeze' mode. The SF260 for instance will cause the ASI to fluctuate up and down, where as the Legacy will not interfere and the ASI is rock steady.

Stephen

Cheers, I will add the vertical speed sync and see if that helps your gauge read correctly.