VGA Charts - Question on Flight Sim X Results

[menetlaus]

Looking at the Flight Sim X - VGA charts, all the results from a Nvidia 7800GT to the 8800Ultra fall between 20 and 24 FPS regardless of the resolution (1024x768 to 1920x1200)

Can someone tell me why such a range of video cards gives such a close result on these charts, and failed to drop off much after a resolution increase?

The next two questions are from the 1024x768 chart (still FS:X): How did 16 video cards all fall within a 5% range of each other? Isn't the 8800 Ultra a lot more powerful than the 8600GTS - but the 8600 is ranked higher, what happened?

Thanks.

 

[mactronix]

Hi im not familliar with the game but it sounds to me like its very cpu based and not heavily dependant on the gpu.
As the same test bed ie same processor is used for all cards then it would make sense that this was what was going on.
If this is indeed the case then the question i would like to know yhe answer to is why the h*** are they using it to test graphics cards.
Sorry for the rant but maybe someone else knows if im guessing right or not.
Mactronix

 

[TheGreatGrapeApe]

FSX is very CPU-centric, and even driver efficiency centric, but that will likely change once DX10 arrives and changes the loading on the CPU/GPU. Early looks at the beta info look promising for geometry and texture workloads.

MacT, the reason it's there is because alot of people asked for it. It's not great for low end, but crank the resolution+AA and you see the difference where the 8600GTS-G8800GTX/Ultra diff is ~6fps ~(+35/-26% for the Ultra), but it's of course still very host-centric.

Right now it's not that benchmark showing some huge dramatic difference (other than across companies), but a benchmark that also show little to no difference is also important to show people just that, that in FSX it's still currently primarily about the rest of the system's performance. So someone gamine at 1280x1024 can pretty much see that spending $500+ on a new GTX/Ultra isn't going to do much for them over their previous 7600GT/7900GS or going to the X1950XTX from an X800XT won't yield much difference, and that going from that same X800XT to HD2900XT isn't about to make the performance jump 2 fold or something wishful like that, money might be better spent upgrading the CPU+MoBo+Memory.

I hope they keep it so that if the do a benchmark (on same config) once the DX10 patch comes out in the fall we can see the difference from that change.

I say it's still a valuable benchmark, but that's just me.

 

[mactronix]

Hm now you put it like that ape i can see the relivance.
The prob is though a lot of people dont know these things its like my pet gripe of them putting sm2 and non hddr cards in tests that support them, people who dont know the diff will get missled,i understand its not intentional and you cant cover every base.but still.......
Mactronix

 

[TheGreatGrapeApe]

Yeah I understand, but when you think of the absolutely minor numer of FSX tests out there this one gains a little more value. but I agree I hate when people crank Oblivion's other setting but then turn down or off foliage, textures, etc. WTF!?!

If you were to look at [H]'s FSX tests you'd almost get the impression that ther's a night and day difference between the cards, but look at it from the wider perspective I think you get a better idea of where the cards fit into that game, and that they don't play a prominent role. It's definitely the opposite of what most reviews do (try to sell a compelling reason to empty your pockets for sponsors), but I like that this one shows alot of people who would know no better, hey, don't expect miracles from the graphics cards, focus you attention on the rest of your system, or if you have a good CPU, maybe save your pennies for now.

Anywhoo, glad I was able to illustrate how some may find it useful, and I appreciate that at first blush it definitely looks like 'what's the point?'.

 

[mactronix]

Cheers ape and yea 100% with you on the impartiality of toms thats one thing no one can fault.
Mactronix

 

[No1sFanboy]

I don't know for certain but I would guess that the difference between cards is probably greater now since sp1 for FSX. Some of the cpu bottleneck is alleviated now that FSX uses more than one core.

That said my video card still runs cooler running FSX than most games. I hope grape is correct and FSX under dx10 finally gives my GPU some work to do.

 

[TheGreatGrapeApe]

Yeah if you read PTaylor's blog about the FSx development you can see the areas they're focusing on and the progress they're making.
No benchies or anything like that unfortunately though. But it's obvious Paullikes this test to as he linked to the last THG test and said it was pretty similar to what they experienced themselves.