Thair

EVGA GTX670 videocard

I finally decided to switch from the AMD platform to the Nvidia one. FSX was written for the Nvidia cards and the GTX670 is of the latest generation. Hence a reasonable performance gain can be expected. Today I picked up the ordered card and at the beginning of the afternoon I replaced the current card.

Before I removed the old card I ran two benchmarks to be able to see the difference in performance. First I ran 3DMark Vantage. I have used this one several times in the past time to measure the performance after I upgraded the computer in some way. I got a Pscore of 11.804 (GPU 12.726 and CPU 9.696).

Then I ran the FSXMark11 benchmark specifically made for FSX. In a standard flight a route is been flown in which a part of five minutes (300 seconds) is monitored. Each flight is logged for number of frames, average, minimum and maximum frames per second. The first measurement s not to be counted.

The results of the next five measurements went from 4.331-4.453 (number of frames), 14,437-14,860 (average), 10-12 (miniumum) and 18-20 (maximum). Not to good, but nevertheless acceptable and close to eachother.

After replacing the videocard, the removal of the old drivers and installing the drivers from the supplied CD (version 301.25) I ran 3DMark Vantage again. Now I got a Pscore of 24.980 (GPU 21.654 and CPU 46.324). The CPU score astonished me, it looks like the old videocards kept the CPU on a kind of a handbrake. Strange.

When running the FSXMark11 benchmark again I got a big disapointment. First I renamed the FSX.CFG file to a .7770 version so FSX could make a new (clean) version with the Nvidia card data. After that I altered the FSX.CFG file as mentioned in the file in Downloads. The results of this benchmark were: Number of frames 4.836-6.384, average 16.120-21.280, minimum always 0 and maximum 22-29.

These values have a wider spread than with the AMD card. What stands out is the ‘0’ value for the minimum. This can however be correct because every so often all the screens go black. The led on the TH2GO DE lights up red for about one second, then starts flashing green and after a few seconds lights up green constant (as it shoud be). After the screens go black it takes 7-20 seconds before they all (one by one) come back to life, It often happens that even the message ‘No signal’ is displayed as if the monitor is not connected.

The black screens don’t only occur when a key is pressed or the mouse is clicked, but also when an alarm sounds and even when I am doing nothing but watching the screens. The performance figures really look nice, but ‘flying’ with all those black screens definitely is not.

With the AMD card I was able to put the instruments on the 4th monitor, this isn’t possible with the Nvidia card. I can move the mousecursor across all screens, but the instruments all stay on the three screens behind the TH2GO DE. Everytime the screens go black, the mousecursor is placed in the middle of the 4th screen.

Tomorrow I will look further into this problem. The GTX670 card has four outputs. A maximum of three screens can act like the TH2GO DE does. The 4th screen can be seen as a separate screen. This all acts like the current (previous) setup does. I will look into this tomorrow to see If the black screens disappear and the 4th screen can be used normally.

Leave a Reply

Your email address will not be published. Required fields are marked *