NVidia 770 GTX to 970 FTW+ Upgrade
-
Thank you for the responses.
I will double check the NVidia control panel vsync setting and report back.
Morgan (Corsair?), I will then try a reinstall. Did that include removing Falcon from the registry or just re-running the 433 patches pointing at the OEM F4?
Is there something within Falcon or BMS software (config files) that latches onto the original graphics hardware and settings that doesn’t recognize the vCard update; of which a reinstall workaround bypasses? If so, this is a bug.
If the vsync isn’t the problem, I will try the reinstall and report back.
Any other ideas or similiar hardware upgrade experiences?
-
Is there something within Falcon or BMS software (config files) that latches onto the original graphics hardware and settings that doesn’t recognize the vCard update; of which a reinstall workaround bypasses? If so, this is a bug.
You can try deleting the dx9display.dsp under user/config, but I’m not sure it’s related.
Other than that, the sim EXE will create the DX HW Device reference every time it is loaded, so there is no such thing as “remembering” anything from previous runs…
-
if you try to reinstall, run First this https://www.sendspace.com/file/1rj2am
-
Confirmed both BMS Graphics and NVidia control panel 3D settings have VSYNC disabled.
Removed DX9DISPLAY.DSP and it created a new one (I used same settings as before).
Still same frame rates as before.It looks like ill have to consider trying the reinstall methodology next. (and save a copy of my current install to use later if no difference).
I see that this is a registry cleaner that removes the following:
[-HKEY_LOCAL_MACHINE\SOFTWARE\Benchmark Sims\Falcon BMS 4.33]
[-HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Benchmark Sims\Falcon BMS 4.33]This necessary?
Update:
as another test, I fired up BMS 432 with the same settings as with 433 (I have both installed), and yes there’s a significant increase from before by about 1.5 - 2x (comparing 770 with 970 gpus)
in cockpit: 75 fps
ext view: 95 fpsSo… before I remove the 433 bms registry, why would I be seeing significant gains in 432 and negligible in 433?
-
Time to get a bit more scientific about this…
BMS 432 and 433 are running same desktop NVidia settings, and as close to same in BMS Graphics as well.
Same single jet on the runway training TE scenario. This is the 970 running at 1920x1080x64-bit, no vsync, latest NVidia driverMy PC at rest (no BMS running).
With BMS 432 running
With BMS 433 running
So its clear that BMS432 is using 50% more CPU cycles, but at 32% its still not close to being CPU limited in this scenario. Note that some of the CPUs are running over 50% individually, but still not clipping at the max utilization when running 433 on the 970.
What still puzzles me why there isn’t a significant difference between the BMS 433 test for the 770 and 970. Something seems to be throttling back the 970 when run with 433. I should be getting around 100 fps instead of 60 - 65 fps.
-
Mobo? Ram?
Vram speeds and bw speeds comparing old and new vga? -
Asus 87Z
16gb corsair
http://www.hwcompare.com/18043/geforce-gtx-970-vs-geforce-gtx-770/The version of the 970 FTW+ (For The Win) I have is the max overclocked and handpicked components.
-
Well… you should do a SERIOUS comparison in Falcon BMS 4.33, meaning that you should take some reference mission, maybe even something with TGP and WPN running at the same time and run some VERY BASIC benchmark test:
Natural cockpit view - Note FPS
Orbit view - Note FPS
TGP working - Note FPS
TGP & WPN working together - Note FPSThen you should render some conclusions. From my experience judging new HW too fast doesn’t say anything really, especially when you compare 2 relatively strong cards with only 1 generation difference. I remember to have a similar feeling when I upgraded back then from GTX-260 to GTX-560TI (2 generations difference), I loaded a TE and said to myself: “That’s it??? 5-6 FPS difference?” but then after flying for some time I could see that suddenly a lot of high-GFX situations in the sim would suffer much less hit than before.
770 and 970 are both strong cards, I don’t expect to see THAT much of a difference between them. However there should be SOME difference, which only a serious benchmark on the same system can confirm.
100 FPS is nothing but a dummy number, I bet even with a 980TI I wouldn’t see 100 FPS in a given TE in cockpit view (I’m running with everything maxed out including multi-sampling).
Are u using all shaders? are u using multi-sampling? if yes on which quality? or if not are u using the Nvidia Built-in FSAA? There are many variables involved here to just say: “I expect 100 FPS”.
-
To put a fine point on my OP, I’m interested in what might be shared as to:
A. Are others with the 970 or 980 getting frame rates greater than what I reported.
So far we know some folks are, I’d like to know this and what their experiences are. Morgan suggests he was able to get over 100 FPS.B. Explain the significant gains with 432 compared to 433 in my post #8.
Which did show a significant performance gain in 432 but not in 433 for a very basic test with same PC, drivers, BMS and desktop graphics settings, and TE.In retrospect, maybe a poll would better serve on what FPS people are getting with these cards. Of course some test constraints would have to be posted to try to keep a relatively level field.
I want to keep my inquiry simple before it gets lost into the details…
Sure, there’s a lot one could do by diving down deep into this; however the fact remains that others are seeing greater than 100fps with similar rigs. Even Morgan posted likewise in this thread, so it’s not an unreasonable expectation to voice. If others with frame rates greater than what I’m seeing with the 970 could report in, we could get a clearer picture without just digging deeper with one rig. Wider and not only a deeper scan is valuable, and what I’m asking for by the posts.
I agree that the nvidia 970 is compatible to the 770 GTX in terms of video ram BW and texture fill rates (quantity), but the 970 is superior in VRAM capacity and AA (quality). I’m sure the 970 should be able to run same or higher AA settings with less of a performance hit than the 770. Choosing Bms 433 worst case FPS hit areas (like TPG, NVGs, MP campaign, etc) might show more of a difference between the cards as a stress test; but that’s not what I’m looking for currently. I’m looking as to why a 433 non-stress test doesn’t show a significant difference between the cards as did with 432.
Advertised fill rates are not the only or even biggest factor in determining frame rate performance. In fact, all things being equal (i.e. graphics Settings), game optimization can be a huge FPS impact. This might account for what I’m seeing or maybe not.
Cheers mate
-
Bump for the OP