December 11, 201312 yr Yeah, SLI is more for computational speed rather than storage (VRAM) of things that need to be computationalized (I made up that word :Nerd: but you get the point). So VRAM does not speed up frame rates. Multiple GPUs do that. However if you are using up your VRAM things will slow down. SLI also needs a good CPU to feed both of those cards. SO if you compare both cards not maxed out VRAM the 780s will always have faster frame rates than the 680 or 770 ever could. Throw in 780ti or 780s in SLI without maxed out vram and frame rates can really grow. But watch out - you also need a monitor that is able to rock all of those frames. Like a 144hz instead of the more popular 60hz. Rob"Life is 10% what happens to me and 90% of how I react to it"
December 11, 201312 yr Yeah, SLI is more for computational speed rather than storage (VRAM) of things that need to be computationalized (I made up that word :Nerd: but you get the point). So VRAM does not speed up frame rates. Multiple GPUs do that. However if you are using up your VRAM things will slow down. SLI also needs a good CPU to feed both of those cards. SO if you compare both cards not maxed out VRAM the 780s will always have faster frame rates than the 680 or 770 ever could. Throw in 780ti or 780s in SLI without maxed out vram and frame rates can really grow. But watch out - you also need a monitor that is able to rock all of those frames. Like a 144hz instead of the more popular 60hz. 120hz here, I would never go back.
December 11, 201312 yr I'm not sure I understand something very fundamental about monitors: Why would one need greater than a 60hz monitor? Wouldn't that really only come into play if we were allowing fps to get above 60? And why would you ever want fps up in that range anyway?
December 11, 201312 yr That is a good question my friend! Bottom line is really refresher rate - from 60-120fps on a 60Hz monitor will look the same. But once you go to 120-144hz the visual smoothness is increased. Rob"Life is 10% what happens to me and 90% of how I react to it"
December 11, 201312 yr Ok I try to refrain myself from going over geeked out on you guys.....but it is getting the best of me. If we are talking SLI - There are many other factors to consider on this also (note I am not speaking to FSX or P3D - but just the tech). If you are going to start running SLI you need to consider your PCIe on your motherboard also. Most boards, outside of higher end enthusiast boards, are PCIe 2.0 and can create a bottle neck. I run the newest motherboard released by Asus the Black Edition. You can read in my post that nobody (waaa) has even viewed :Doh: http://forum.avsim.net/topic/428384-asus-rampage-iv-black-edition/ Also our friend Rob Ainscough with the Titan can beat us all (VRAM wise) if he goes with an SLI Titans -god bless him if he can afford that- because he can then use all of that Vram on one board and have the computational speed dual GPUs. I am saying that even when I add my second 780 I will still only ever have 3 gigs of VRAM but he will always have 6. Rob"Life is 10% what happens to me and 90% of how I react to it"
December 11, 201312 yr Most boards, outside of higher end enthusiast boards, are PCIe 2.0 and can create a bottle neck. Completely disagree, and the first I ever heard of this. ASUS ROG STRIX Z390-E GAMING / i9-9900k @ 4.7 all cores w/ NOCTUA NH-D15S / 2080ti / 32GB G.Skill 3200 RIPJAWS / 1TB Evo SSD / 500GB Evo SSD / 2x 3TB HDD / CORSAIR CRYSTAL 570X / IPSG 850W 80+ PLATINUM / Dual 4k Monitors
December 11, 201312 yr Completely disagree, and the first I ever heard of this. That is completely fine by me buddy. Rob"Life is 10% what happens to me and 90% of how I react to it"
December 11, 201312 yr That is completely fine by me buddy. Good.. On average, with cards available now in 2-way SLI the gain of PCIE 3.0 is... A few percent tops.. There is no bottleneck.. 3-Way, 4-Way.. You may start seeing a problem. ASUS ROG STRIX Z390-E GAMING / i9-9900k @ 4.7 all cores w/ NOCTUA NH-D15S / 2080ti / 32GB G.Skill 3200 RIPJAWS / 1TB Evo SSD / 500GB Evo SSD / 2x 3TB HDD / CORSAIR CRYSTAL 570X / IPSG 850W 80+ PLATINUM / Dual 4k Monitors
December 11, 201312 yr 3-Way, 4-Way.. You may start seeing a problem. Now we agree. Rob"Life is 10% what happens to me and 90% of how I react to it"
December 11, 201312 yr Author But I wonder if faster is necessarily better for P3D if faced with the choice of a slightly slower card with twice as much memory?Just decided to buy the Titan second. For € 550 ( $ 750 ) It will be the EVGA Titan Superclocked. When it has arrived I can borrow a second one ( Gigabyte Titan ) to see what SLI does. If I will be getting the same extra performance in FSX with SLI as before I will buy that card too ( also € 550 ). Hopefully my 1000W Thermaltake ps will be strong enough. Just sold my 3 26" monitors and one of my 2 GTX680 cards so I can do this Then next weekend I will begin with installing P3D v2. Lucky me , Flight 1 is going to help me installing the Ifly Cockpit Builders Edition in P3D. 5950x3d 5.4-5.7 GHz - Asus ROG 870 Crosshair Apex - GSkill Neo 2x 24 Gb 6000 mhz / cas 26 - MSI RTX 5090 Gaming Trio OC - 1x SSD M2 6000 2TB - 1x SSD M2 2800/1800 1Tb - Corsair 5400 case - Corsair 360 liquid cooling set - 3x 75’ TCL tv. 13600 6 cores @ 5.1 GHz / 8 cores @ 4.0 GHz (hypterthreading on) - Asus ROG Strix Gaming D - GSkill Trident 4x Gb 3200 MHz cas 15 - Asus TUF RTX 4080 16 Gb - 1x SSD M2 2800/1800 2TB - 2x Sata 600 SSD 500 Mb - Corsair D4000 Airflow case - NXT Krajen Z63 AIO liquide cooling - FOV : 200 degrees My flightsim vids : https://www.youtube.com/user/fswidesim/videos?shelf_id=0&sort=dd&view=0
December 11, 201312 yr VRAM usage is based on the scene it's trying to render, that scene will change as you fly thru the world. Your P3DV2 setting will influence the VRAM usage, but it's a dynamic usage. If the scene needed to be rendered exceeds the amount of VRAM you have available, then the scene has to use system RAM (VAS) -- at which point you'll start moving to the limits of 32bit address space (OOM). Indeed--have a look. The first shot is w/ Shadows set at Ultra (all receive checked), w/ vegetation/autogen at Extreme. Note VRAM & GPU use. Now here's autogen/vegetation at Extreme, but Shadows still at Ultra: The 3rd shot is w/ Shadows at High, but vegetation & autogen at Normal: It's really setting Shadows at Ultra AND Extreme autogen/vegetation that really pushes GPU utilization. I quite frankly don't appreciate a visual difference when shadows are set to Ultra versus High. What is completely unclear to me still, and I think this can be sorted out by testing, how much of the total burden of the what FSX processes (i.e., all the non-DirectX work, autogen, mesh, LOD, etc) is done by the GPU in V2 versus how much of the total work by the GPU is dedicated to DirectX effects as it were: tessellation & shadowing for example. Noel System: 9900X3D Noctua NH-D15 G2, MSI Pro 650-P WiFi, G.SKILL 64GB (2 x 32GB) 288-Pin PC RAM DDR5 6000, WD NVMe 2Tb x 1, Sabrent NVMe 2Tb x 1, RTX 4090 FE, Corsair RM1000W PSU, Win11 Home, LG Ultra Curved Gsync Ultimate 3440x1440, Phanteks Enthoo Pro Case, TCA Boeing Edition Yoke & TQ, Cessna Trim Wheel, RTSS Framerate Limiter w/ Front Edge Sync. Aircraft used in MSFS 2024: Fenix A320, Aerosoft CRJ, FBW, WT 787X, I-Fly 737 MAX 8, Citation Longitude.
December 11, 201312 yr Indeed--have a look. The first shot is w/ Shadows set at Ultra (all receive checked). Note VRAM & GPU use. SFO Extreme Vegetation & Autogen Shadows Ultra .jpg The 2nd shot is w/ Shadows at High, but vegetation & autogen at Normal: SFO Normal Vegetation & Normal Autogen Shadows High .jpg It's really setting Shadows at Ultra that really pushes GPU utilization. I quite frankly don't appreciate a visual difference when shadows are set to Ultra versus High. Noel, What do you have ticked for shadows when you had them set to Ultra? I am wondering if it is a particular object that really causes the heavy usage i.e. terrain cast... Just curious. Bob Officially retired
December 11, 201312 yr Noel, What do you have ticked for shadows when you had them set to Ultra? I am wondering if it is a particular object that really causes the heavy usage i.e. terrain cast... Just curious. Bob I need to go back and run this test again as on the most extreme example (all shadows checked on the receive side, vehicle in/out receive and generate) I had ALL shadow boxes checked on both sides, but on the Normal object example I had all unchecked except vehicle both sides. But the bottom line is VRAM is utilized full under the right circumstances, and at one point I saw VRAM all the way to the top at 5900. This could be why I have yet to see an OOM. Noel System: 9900X3D Noctua NH-D15 G2, MSI Pro 650-P WiFi, G.SKILL 64GB (2 x 32GB) 288-Pin PC RAM DDR5 6000, WD NVMe 2Tb x 1, Sabrent NVMe 2Tb x 1, RTX 4090 FE, Corsair RM1000W PSU, Win11 Home, LG Ultra Curved Gsync Ultimate 3440x1440, Phanteks Enthoo Pro Case, TCA Boeing Edition Yoke & TQ, Cessna Trim Wheel, RTSS Framerate Limiter w/ Front Edge Sync. Aircraft used in MSFS 2024: Fenix A320, Aerosoft CRJ, FBW, WT 787X, I-Fly 737 MAX 8, Citation Longitude.
December 11, 201312 yr I need to go back and run this test again as on the most extreme example (all shadows checked on the receive side, vehicle in/out receive and generate) I had ALL shadow boxes checked on both sides, but on the Normal object example I had all unchecked except vehicle both sides. But the bottom line is VRAM is utilized full under the right circumstances, and at one point I saw VRAM all the way to the top at 5900. This could be why I have yet to see an OOM. The reason I ask is that I with default scenery everything maxed I had 3.2 used, but I only had vehicles checked. Bob Officially retired
December 11, 201312 yr I've reached 2.2 GB on my 780 with one 1080p monitor. Settings were high and flying around Edinburgh at very dense autogen. I don't think it's possible for me to reach the full 3 GB on one monitor at my settings, but I probably could with sliders full right around NYC.
Create an account or sign in to comment