manashttu

Last night I spent some time attempting to isolate the ideal settings in nvidia control panel which maximize FPS and maintain eye candy in DCS on a triple monitor setup. I've read/watched lots of guides, but many of them are old, and several have some findings inconsistent with my experience. For example, it is frequently said on game forums that anisotropic filtering settings have no or imperceptible effect on FPS on modern gpu's. A couple of days ago I inadvertently turned off AF in both DCS and nvidia C.P.---and saw a dramatic jump in FPS (and a rather blurry runway in front of me). Hence, my curiosity to test every setting in N.C.P. Short word on AF---indeed there is a difference in FPS from zero to 2 through 16 and each step of the way. Running latest open beta. System specs below. Also---after trying to cover every combination of settings in N.C.P.---here is what I arrived at for the best performance on a fixed set of in-game settings. I plan to fiddle with determining the best balance of in-game settings next. System specs: 8700K, 32gb DDR4 3000, RTX 3080. Running on triple 1440p monitors in Surround mode 7770x1440, using OpenTrack head tracking.

The TM Warthog stick (the magnetic base)---doesn't have twist. Candidly, if I were in the market for a stick to go with an x52 throttle (which is a good throttle)--I'd be looking at the VKB Gladiator line. I have had little luck with Thrustmaster quality---specifically including the warthog stick. The Gladiator is less expensive than the TM setup and, in my opinion, superior in quality to the TM unit. The Gladiator has twist, a really nice grip, and lots of buttons and an extra axis. I have owned a TM Warthog, VKB Gladiator MK.II (used alongside an x52 throttle), and now a VKB Gunfighter. I love the VKB sticks---wouldn't want the worry of what plastic connectors and internal parts will eventually or immediately randomly break on the TM.

Yes. You can use the x52 throttle standalone, as well as the TM stick standalone. They will show up as separate controllers, which is fine.

Check out Bankler's outstanding Case I practice missions. I expect you could use it for setting up the scenario you describe, and it has a detailed (and addictive) scoring system for the pattern and landing. https://forums.eagle.ru/forum/english/digital-combat-simulator/dcs-f-a-18c-hornet/missions-and-campaigns-af/222472-bankler-s-case-1-recovery-trainer

Here is a post I made regarding my experience upgrading gpu from 1080ti to 3080 on triple 1440p setup. I'm running an 8700k...so it doesn't compare an upgrade to cpu ... But it sounds like your FPS is not far off my 8700k when I was on a1080ti (which is still a beast of a gpu). https://forums.eagle.ru/forum/englis...d-triples-user

Here is a thread of my experience of 1080ti to 3080 on triple 1440p monitors (which is 20%+ more pixels than 4k). In short, the cpu will remain a bottleneck at times in DCS---but the 3080 enables the use of higher levels of antialiasing. On a single 1440p monitor, the 1080ti can handle DCS with antialiasing, too. https://forums.eagle.ru/forum/english/reality/pc-hardware-and-related-software/291266-rtx-3080-experience-so-far-2d-triples-user

I bet you've gotten your money's worth out of that 4790k. Great chip. I loved a 2600k for years until I couldn't resist upgrading to the 8700k...it was still quite strong. Intel chips have had legs for years due to only modest clock and IPC gains generation to generation since the 2600k. I didn't get into DCS until after the upgrade...so can't advise of the gain you might see from a 4790k. Adding the 3080 turned out to basically be an upgrade in MSAA and average FPS..but little benefit to minimum fps. ...Frankly, the sim was still beautiful with no MSAA and played very smoothly on the 1080ti. Now...I'm gonna keep the 3080...but I'd have been some kind of pissed if I had paid $1200+ for a 2080ti for even less gain ...

Think you're exactly right...a 1080ti will run 1440p single screen pretty well maxed out.

Managed to get my hands on an MSI Ventus RTX 3080 this week, and upgraded my 1080ti, which has served very well. Have seen a couple of threads on VR performance, but thought I'd share some initial experiences for 2d players like me. My machine is an 8700k clocked at 5.1, 32 GB 3200 DDR4, Z370 chipset motherboard. Triple 1440p monitors with a bezel-corrected resolution of 7770x1440. Thats's just over 11 million pixels, about 3 million more than 4k. I run a default Field of View in cockpit of 120, and play the open beta version of DCS. 1080ti did very well running this, always playable---but had to run with no MSAA and would run 45 to 65 fps in single player, depending of the map (Syria was tough, Nevada, pretty smooth). Textures all high, no heat blur, trees high, view distance Extreme, 1024 cockpits. Using the new MSI Afterburner plugin which monitors VRAM usage in addition to VRAM allocation, all 11GB of the 1080ti VRAM was often allocated (especially in Syria)---but usage topped out in the low 9GB range on Syria, 7gb range on other maps (PG, Caucasus, Nevada). With the 3080---I see similar minimum FPS. I think this shows that DCS still loads up CPU for a lot of the rendering of buildings and terrain features. At takeoff from Turkey in F14 Instant action, get 48 to 55 fps as wheels start rolling. Where the 3080 shows its power, though, is in the settings one can run. With 4x MSAA (remember---on 11 million pixels)---it runs 45 to 75 fps in single player and sustains 60+ fps except dipping into mid 50's over dense cities on the Syria map. The low of 45, like on the 1080ti, is pretty much only seen on a runway in the Syria map. The improvement in image quality at 1440p with 4x MSAA or 2xMSAA with MFAA enabled is beautiful, and to do this on triple 1440p is impressive. The VRAM usage is the same as on the 1080ti, and it doesn't appear to be affected at all by MSAA levels. Pretty sure 10gb VRAM is sufficient for now if it handles this high of resolution. The sum of the matter is this: it appears due to CPU bottleneck which is inherent in DCS, the raw min and avg fps jumping to 3080 is not jaw dropping. However, the headroom to pump antialiasing to high levels on 7770x1440 resolution is rather impressive--- If Syria is disregarded---(hopefully it will be tuned in early access)---the 3080 delivers 60+ sustained fps on 11 million pixels with 4x MSAA or 2xMSAA and MFAA.

Using the new Afterburner GPU.dll plugin, did some testing on my triple 1440p setup (7770x1440--that's 11+ million pixels...3 more than 4k) on my 1080ti. The standard VRAM monitor showed all 11gb of VRAM allocated in the F-14 Turkey takeoff in the Syria map. Using the process VRAM monitor, putting it on 2x MSAA, high textures, and Extreme draw distance, flying over the most intense part of the city---it's actually using 9,395 mb of VRAM. This is encouraging information---as I was eyeballing the 10gb 3080 (whenever it actually is available)---but have been concerned about VRAM. This pretty much worst case shows the 3080 should be fine and 3090 would be a bit of overkill. A 16gb 3080ti would be nice.

Thank you for posting the new AB build. This new functionality is great--- On my 1080ti, the standard VRAM monitor showed all 11gb of VRAM allocated in the F-14 Turkey takeoff in the Syria map. I'm running triple 1440p monitors (7770x1440). Using the process VRAM monitor, putting it on 2x MSAA, high textures, and Extreme draw distance, flying over the most intense part of the city---it's actually using 9,395 mb of VRAM. This is encouraging information---as I was eyeballing the 10gb 3080 (whenever it actually is available)---but have been concerned about VRAM. This pretty much worst case shows the 3080 should be fine and 3090 would be just total overkill.

All this is polishing a jewel. THANK YOU for this fantastic module----the cost per hour of enjoyment is in the pennies and rapidly declining. Don't let your acknowledgement of areas for improvement diminish the accomplishment in making this beautiful recreation of the Tomcat.

Folks, it will be worth the wait. I got one of K-51's early units back in January, and still amazed at the features and quality at the price.

The trees are fantastic---such an improvement over stock. THANK YOU!

This is a tremendous deal---especially with the extra pedal pads. Those take many hours to print, and the photos look like a quality 3d print at that. Have these pedals already---and they are excellent. Good luck with the sale---someone should jump on these.

Love these mission. These exercises are obsession-inducing and great fun. I have done dozens of patterns, traps, and crashes in the F-14 and Stennis. Then today switched to the Supercarrier. Got my best score ever (62)---then in studying the score, I discovered that the scoring seems to be missing the "time in groove" five point line item that is in the non-Supercarrier mission, as all the possible points that show in the score screen add up to just 70. Here is a screenshot https://drive.google.com/file/d/16lIenSYBeoCKOwGD7QbysyYzjVsC_VXl/view?usp=sharing If the score is 62 out of 70---I'm thrilled. EDIT: I went back and looked at the scores on the previous attempts in the same session--the had the "Groove time" score on the first two of my five attempts. Attempts 3, 4, 5 did not. Weird.

Thank you for this outstanding mission set. It is surprising how much enjoyment/frustration/reward can be gleaned from DCS without even loading a weapon thanks to exercises such as this. The breakdowns of the Case I procedure are perfect for making practice productive and repetition rewarding. After a dozen cycles in the F-14 (after a dozen before that in the stock training mission)--best score is 46. I see many hours of fun practice ahead.

Hello all. Prior a couple of open beta versions ago, when activating the FPS and other stats on screen (Ctrl Pause), included in the info was the current FOV. That is now gone. Does anyone know a setting to change to get that readout back? Thank you.

This is a bump for Lucky's listing. I bought the Warthog he had in the original listing---it arrived today after making a deal on Saturday. It was exactly as described and in excellent condition, well packaged. Somebody ought to grab those Crosswinds. That is a great deal on a nice set of pedals (I'd have bought them in a heartbeat if I didn't already have the same set under my desk).

I recently worked a deal with this seller for TM MFD's. He invoiced through PayPal, shipped through FedEx, and promptly sent a tracking number. Everything has gone smoothly, for those who have concerns. If I didn't already have a set of Crosswinds, I'd be all over this set.

Love this mission. Thank you for sharing it. It is a fun exercise that demonstrates inertia and the effects of speed and AoA on turning and roll rates. And it nicely shows off DCS graphics and weather effects. I'm using it over and over to tune my stick and throttle movements, and to learn smooth, purposeful control. Having a start-finish timing aspect makes it perpetually replayable, too. In my sim racing days, it was fun and interesting to take different cars back to back to the same track to learn the power band and handling differences. Your mission nicely makes this possible for planes. I made copies of your mission for the F-15C and other planes (even did the P-51D--took 13 1/2 minutes). The F-15C is much more agile than it feels, I think because it's so smooth. Tried to replace the F-16C with the F-14B. However, having followed the exact procedure that worked to substitute other planes, when the mission loads, it ends up just showing a map and won't enter the cockpit. Do you or others more knowledgeable with the mission editor know why this might be? Thanks again for the cool mission. EDIT: Actually, turns out I had a corrupt F-14B. Repair fixed it. Works great in your mission (though--what a handful).

Thank you for this label mod. It is a nice balance making up for the limitations of screen resolution without feeling like a cheat. I did tweak it a bit, making the black dot disappear within 3000m instead of the box. It was very simple to just replace the "square" with "" as the definition for aircraft_symbol_near.

I can also verify that the GrassMonkey wireless headtracker LED unit works with the TrackIR camera and software. I have both (one for kids' rig) and the wireless Puck reads on TrackIR just like the Pro. It attaches to headphones or a head clip by Velcro.

If you're open to a very nice alternative, take a look at this US maker of a head tracker. I bought one and it works very well. https://rdial60.wixsite.com/grassmonkeysim