1X 2080Ti vs. 2X 2080Ti vs. 1X TitanXP - P3D V4.3

[A340Bangla 40]

Gerard,

your visuals are similar to mine - 3 x 1920*1080, no cockpit visuals (I am still flying Project Magenta 😊). At the moment I am on a 6700K with 2 x 980ti's in SLI.

I am about to build a new system (have ordered a 8086K) and am torn between 2 x 1080Ti and 1 x 2018ti (which would be cheaper). Do you mind to let me know how the 2080ti is doing in your system compared to the 2 x1080ti's you had ?

[VHOJT 167]

On 10/29/2018 at 2:33 PM, A340Bangla said:

Gerard,

your visuals are similar to mine - 3 x 1920*1080, no cockpit visuals (I am still flying Project Magenta 😊). At the moment I am on a 6700K with 2 x 980ti's in SLI.

I am about to build a new system (have ordered a 8086K) and am torn between 2 x 1080Ti and 1 x 2018ti (which would be cheaper). Do you mind to let me know how the 2080ti is doing in your system compared to the 2 x1080ti's you had ?

@GSalden this would definitely be good to know 🙂

[GSalden 4,656]

3 hours ago, VHOJT said:

@GSalden this would definitely be good to know 🙂

The 2080Ti is performing exactly the same as the 2x 1080Ti SLI combo.

[pgde 571]

Gerald -- so was the upgrade of cards worth it in your opinion?

[mole_man99 0]

Getting similar FPS results on 4k monitor in all 3 settings scenarios with a GTX1080.

This is paired with 8700k at 4900 mhz

Seems only worth upgrading if your chosen flightsim starts to use the new features discussed. Stepchange in dynamic lighting sounds interesting.

Thanks so much for the thorough numbers.

[Guest]

Unfortunately I didn't have another spare 1080Ti or TitanXP to test in SLI mode so as to get a valid compare with 2X 2080Ti.

I'll see if I can find my recorded FSR file (not sure these can be used in a different system?) if anyone wants to really try to duplicate my exact test and not a "similar" test … I'll also need to post my add-on configurations including ENVSHADE/TEX and all those settings as shader enhancements do impact performance.

I'll add the info if someone is really wants to do an exact Apples to Apples compare of 2X 1080Ti and 2X 2080Ti … just let me know.

Cheers, Rob.

[A340Bangla 40]

11 hours ago, GSalden said:

The 2080Ti is performing exactly the same as the 2x 1080Ti SLI combo.

My 2080ti arrived yesterday. But thank you anway, means I have have probably made the right (cheaper) decision.

Plus, reading somewhere else it seems 1080ti's are getting rare.

[mole_man99 0]

As you used a default aircraft, the FSR file should be transferable. Happy to try if you find it. Your settings and flight could become a standard to check any hardware combo via other users, or real world improvements from future hardware.

May be an idea to run it with no add-ons to level the field. I had UTX active, and maybe it effects a frame or two with extra roads and lake shorelines.

I'll have to figure out data logging with afterburner.

[vortex681 806]

1 hour ago, mole_man99 said:

Your settings and flight could become a standard to check any hardware combo via other users, or real world improvements from future hardware.

I don't think that it would be that straightforward. The systems would have to have identical versions of the OS with any local tweaks, exactly the same background apps, the same antivirus, identical storage drives and so on.

[AlphaInfinity 13]

On 10/19/2018 at 3:55 PM, bbain1187 said:

With my 2080 Ti, I can now maintain 30fps in 4k at dynamic lighting-enabled airports with 4x SSAA set. I have jacked up all the GPU-driven graphics settings (shadows, textures) except for disabling dynamic reflections. GPU utilization was still just about 70% from what I recall sitting at the gate in FlyTampa Boston in the FSL A320. This is with a 6700K @4.7 and 16 gigs of RAM running at 3200. 

Ben 

EXCEPT at KSEA right? I am curious to know if your rig will maintain 30FPS at the toughest airport in the game. 

[bbain1187 118]

22 minutes ago, AlphaInfinity said:

EXCEPT at KSEA right? I am curious to know if your rig will maintain 30FPS at the toughest airport in the game. 

I took off at KSEA the other day in rain/overcast conditions and it was 30fps no problem. Depending on graphics settings and time of day, or if I use the new dynamic lighting at DD KSEA, I can get it to go below 30fps in some situations. Over the last couple of weeks I've created different graphics profiles for different situations in order to maintain 30fps without stutters. If I were flying into DD KSEA at night in rain with dynamic lighting in the FSL A320, I would use 8x MSAA and turn down my shadows settings to maintain 30fps. I'll post some screenshots in the next few days with different settings if I get a chance.

Ben 

Edited November 13, 2018 by bbain1187

[Ray Proudfoot 8,932]

19 hours ago, bbain1187 said:

I took off at KSEA the other day in rain/overcast conditions and it was 30fps no problem. Depending on graphics settings and time of day, or if I use the new dynamic lighting at DD KSEA, I can get it to go below 30fps in some situations. Over the last couple of weeks I've created different graphics profiles for different situations in order to maintain 30fps without stutters. If I were flying into DD KSEA at night in rain with dynamic lighting in the FSL A320, I would use 8x MSAA and turn down my shadows settings to maintain 30fps. I'll post some screenshots in the next few days with different settings if I get a chance.

Ben 

Without wishing to be disrespctful what you have posted above is totally meaningless without knowing your hardware. PLEASE put it in your signature.

[Bob Scott 7,716]

On 11/6/2018 at 7:23 AM, GSalden said:

The 2080Ti is performing exactly the same as the 2x 1080Ti SLI combo.

This right here has me intrigued.  A single card that performs at-par with my 2x SLI 1080Ti configuration.  I'd really like to get away from SLI if I can, but not at the expense of losing the performance I have now.  Seems maybe now I can.

So, I have an eVGA 2080Ti on the way.  Will take my two 1080Ti boards out and use them to upgrade my DCS/FSX/XP-maybe-someday box and my FPS/other game rig.

So I will have the opportunity to do some performance comparisons, first on 2x SLI 1080Ti, then on a single 1080Ti, then on a single 2080Ti.  If someone can lay out a few relatively straightforward test scenarios that do not require me to significantly reconfigure my system or spend a lot of time on it, I might do that while deconfiguring and reconfiguring the system.

Regards

[Guest]

46 minutes ago, w6kd said:

So I will have the opportunity to do some performance comparisons, first on 2x SLI 1080Ti, then on a single 1080Ti, then on a single 2080Ti.

Recently, I've had what I initially thought were some strange results with 2080Ti SLI in the FSL A320 at EDDK ... as in 8-12 FPS ... I removed one 2080Ti, DDU, manual drive install and re-tested and was around 26-30 FPS with a single 2080Ti. (keep in mind these are fairly maxed out settings including EDDK maxed out settings).  So I put my other 2080Ti in and back to SLI performance mode and check CPU usage at this scenario and sure enough, my poor CPU was pegged at 100% ... as soon as that happens then all bets are off on the performance front ... just can't feed the GPUs and performance takes a noise dive.  

After looking at my initial test spreadsheet (at the beginning of this thread), AVG CPU usage increases when SLI is active and the higher graphics settings ... about a 17% "relative" increase in CPU load.  Keep in mind I picked my test scenario at a location that I knew would NOT overload the CPU so I could get a valid idea of GPU(s) actual performance.  So this is something to take into consideration when going SLI ... it will be important to provide enough processing headroom for the CPU to work with SLI ... so one will either need to decrease CPU loads in order to realize the benefits OR increase CPU performance (in my case I would probably need to increase CPU performance by about 17-20%).

So my conclusion here is that increasing CPU load too much with NVLink/SLI setup can actually become a performance detriment ... it's the balancing acting.

Cheers, Rob. 

[Bob Scott 7,716]

15 minutes ago, Rob Ainscough said:

Recently, I've had what I initially thought were some strange results with 2080Ti SLI in the FSL A320 at EDDK ... as in 8-12 FPS ... I removed one 2080Ti, DDU, manual drive install and re-tested and was around 26-30 FPS with a single 2080Ti. (keep in mind these are fairly maxed out settings including EDDK maxed out settings).  So I put my other 2080Ti in and back to SLI performance mode and check CPU usage at this scenario and sure enough, my poor CPU was pegged at 100% ... as soon as that happens then all bets are off on the performance front ... just can't feed the GPUs and performance takes a noise dive.

One thing I'm wondering a bit about is whether a pair of 2080Ti GPUs on a Z390/Z370 board would/could have issues saturating the eight PCIe lanes to each GPU.  Your 7900X isn't so constrained...I think you still get 16x on both GPUs there, correct me if I'm wrong.  But an 8-core 9900K just might start to max out the busses with 2X 2080Ti at 8x/8x.

Something to keep an eye on with your 9900K experimentation, anyway.

Cheers