TheFamilyMan

Complete agreement, been hoping to do that since the 3080 ti released (the 3080's 10GB just doesn't cut it for me). Knew when I built my system about a year and a half ago that I wanted to retire my gtx 1070 soon and almost bought a gtx 2080 S at that time, but its 8GB was a deal breaker plus the 30 series was just about to be released. I now kick myself for not getting that 2080 S for about $660...if only I had known what the gcard market soon would become I've been surprised, and mildly pleased, that my gtx 1070 handles my base settings well enough to drive a 2560x1440 panel at 30 fps, but yeah it's really holding back my MSFS performance but I'm not willing to spend $2000 or more to upgrade it. Seeing how LOD 6/6 really stresses my i7 10700 at 30 fps, I can't see how a 3080/3090 could get around that computation limitation, but dang, how I'd love to give it a try. Besides, 6/6 was simply a test I ran to humor myself.

Seeing that SU5 works so well for how I fly in MSFS 2020, just for kicks I set both terrain and object LOD settings to 6 (600 in-game, though only possible using the UserCfg.opt file) and did my KBUR to KSNA flight to see how it goes, in a Waco YMF-5 at 6k'. Holy Mother of Pearl...ground detail for as far that it's practical to see (truly puts 2/2 to shame)! But it's anything but smooth, using the sim's 30 fps vsync. Other than having terrain vector data, texture resolution, text synthesis, ambient occlusion set to Ultra, all my setting are set to High. See my computer specs in my signature. Moving my virtual cockpit view with TIR causes big FPS hits for 2 or so seconds, plus the autogen pop is painfully obvious, mainly in the distance. But once it settles out smoothness returns. The most interesting thing with 6/6, at least for me, is seeing its execution impact on my overclocked i7 10700. With the standard 2/2 LOD in SU5, no core even comes close to ever hitting 100% use, but with 6/6 every view change spikes about half the cores to nearly 100%, and in general the core utilization is about 30% elevated. Also noticed that some of my cores are hitting max temps of 84 c, though only for very short periods; the average in the upper 60s c. I see these temp spikes as a good thing, for it shows that MSFS utilizes AVX instructions (I imagine particularly FMA3 since it's tailored for linear algebra computations, plus that instruction cooks the cpu like no other). Of course, it's easy to blame my 6/6 performance on my (ancient) gtx 1070, but it was rarely pegging 100% and looked to be performing as well as when it's running 2/2. Clearly IMO it's the CPU that matters most for 6/6...oh well, so now back to the mundaness of 2/2 (or perhaps maybe a bit higher). Thanks for reading!

If it's not MSFS 2020: IL-2 Battle of Stalingrad (or Moscow, Kuban, Bodenplatte, aka BoX), or its WWI version "Flying Circus", flying online private server missions with my squad mates. We've been together for 15 years, started off right before "IL-2 1946" released. I also write BoX mission, using their mission editor's "Graphical Data Flow" programming language, which is similar to the programming language "LabView" that I programmed when I worked in aerospace at LM (recently retired). Pinball Arcade, which is a highly realistic pinball table sim. All of its table sims accurately mimic RL tables from the 70's to recent years. The tables are very accurately depicted, with excellent physics and visuals. Unfortunately its developer "Far Sight" no longer supports it, primarily because about 3 years ago Williams and Bally didn't renew Far Sight's license to sell their tables. Though not a sim, at times I spend a fair bit of effort editing HD footage shot while traveling into short films of our adventures, using Premiere Elements. It's nice to see my 16 virtual cores getting a worthy and useful workout when crunching H.256.

I posted earlier that my update took about 4 hours to complete, which included the region updates. One factor I have in my favor is that my system disk is M.2 NVMe and dedicated MSFS disk is PCIe x4 NVMe. Perhaps having crazy ripping fast storage helps a lot with this install. Other interesting facts are that the MSFS runtime threading behavior is far better balanced across all the core and temps are hotter. Nice that the threading is improved, and I'm guessing that the temp increase is due to increased utilization of AVX instructions. This is a good thing for these instructions greatly speed up computation, which is precisely what they are implemented to do. IMO it's downright stupid not use use them to their fullest. I really admire the FMA instructions, which implement the level 1 BLAS linear algebra operation of y = Ax + B in a single instruction. The latest versions of LINPACK is uses this instruction to the extreme, and running such a version of LINPACK heats your CPU to extremes as well.

Just a bit of sunshine for all the gloom others are experiencing. My update took nearly 4 hours to complete. Started it all by bringing up the MS store app and running their app update for MSFS. Then ran MSFS with no issues, once the updated completed installed all the feature updates. All this took me less that 4 hours, using Comcast cable ISP @ 100mb in the south San Francisco bay area, aka Silicon Valley. My standard mini-torture test flight of a Bonanza flight from KBUR to KSNA, mid day hitting an alt of 5k' going right over downtown LA with KLAX in sight, was absolutely stutter free from takeoff to landing. It has NEVER BEEN THIS WAY before! Before it was always stutters at takeoff and landing, and stone cold 1 second freeze ups around downtown LA and at a few other points along the way. Check my hardware setup in my sig. Also noticed that their CPU thread refactoring clearly shows, yay no more core 15 pegging 100%, and my ancient 1070 GTX never pegged 100% either (another first for this flight). IMO Asobo delivered, but this one test isn't everything (but for me it nearly is). As always, unfortunately YMMV.

I've been using TIR for at least 15 years and still love it. One thing to know about TIR is that if you will be wearing eyeglasses while using it, using the reflective clip that comes with the TIR can be problematic since at certain angles your eyeglasses will reflect IR back to the IR sensor. When this happens, it usually causes your view to wildly and suddenly shift, which is most irritating to say the least. The TrackClip Pro solves this problem, but it's a additional expense and another cable to deal with. There are 3rd party versions of this clip that are wireless but they're rather expensive. BTW, VR is (nearly) King. Looking to go that route someday...if only its resolution was better and that #!#@! sweet spot wasn't something to be bothered with.

One facet of il-2 Battle of Stalingrad (also called BoX by its followers), another really ripping great and currently developed CFS, are using its scripted missions for both single player and online multiplayer, which are created using its missions editor (I written a few myself). I can't imagine a CFS that'd be worth my time that does not have such combat specific mission capabilities, and I seriously doubt that MSFS would ever (or even could) go to the depths that BoX does in this regards. Sure, it's fun to shoot something down or blow it up flying an aircraft, but that capability alone gets old really fast IMO.

I agree that Boom is probably the worst name a company making supersonic aircraft should be named. People hated SS AC for this reason - BOOM! But there is hope! Lockheed Martin announced this program a few years ago to develop quiet SS aircraft. I posted this info back then, where I noted that fuel costs will make this tech impractical, for that's what basically did-in commercial SS travel over 40 years ago. But this new SS tech coupled with Prometheus tech fuels...Bring It On! In this light, perhaps Boom should consider renaming themselves Hush. I just hope that Prometheus tech fuels are truly feasible. You'd think that such carbon neutral fuel tech should be front and center of international attention for weaning humanity away from fossil fuel use.

Oh well, another year of waiting for a new gcard for my 1 year old 10700k build. I originally had hopes for the 20GB 3080 Ti, and even may have settled for 12GB, but that foreshortened vram coupled with a "starting at" $1,200 price and practically nonexistent availability is a deal-breaker. What remains to be seen is that will the high-end gcard market ever really recover from its current state of desperation (which unfortunately now also spans into everything needing silicon).

It probably won't go well. I originally installed the MS Store MSFS completely (base install and appdata) and successfully on a dedicated SATA SSD drive. Later I got a NVMe SSD to replace that SATA SSD. So I cloned that SATA drive to the NVMe drive, pulled the SATA drive and changed the NVMe drive's letter to that of the removed SATA drive. This did not work, for on launching MSFS it complained about missing files and then exited. What is even more frustrating is that when I reinstalled that SATA drive with a new drive letter, launching MSFS using my new NVMe install ran fine (that is until I reformatted that SATA drive). After wasting a better part of a day trying to get it to work I ended up reinstalling MSFS on the NVMe drive, but even that was now painful for my "original" install was now botched up and wouldn't uninstall properly for the new install to take place. Can't say what I did in the end to get the reinstall to work, but I swore I'd never again attempt to drive clone MSFS. My advice for a dedicated MSFS drive swap is to first completely uninstall from the old drive, then online reinstall completely to the new drive.

Eye candy is life IMO. Finally tried Bloom Off; no thanks man, I'll take the performance hit (as I do with ultra field of depth) though I can't say I see a noticeable performance difference with this stuff disabled, running 30 fps vsync on. For details see my specs in my sig.

Consider this somewhat ironic twist on this. What if these sightings are actually time travel probes sent back in time from our far distant future generations? The good news here is that the human race would be still going strong, the irony is that it's only from the human race. Or possibly this "going public" could be subterfuge to bamboozle the masses (also potential adversaries) about tech that really is in development. What is fascinating overall is that I remember reading back in the '70's about astronauts seeing crazy unexplainable stuff when on their missions, some of it caught on film too, so nothing really that new here. Edit: Though it's fun and romantic to consider space travel, I don't think we're worthy to go there until we first fix all the problems we have living here on earth.

Please let us know how you fixed this problem. Seems that MSFS uses a well hidden/cryptic disc/device identifier to link the install with the files it tucks away inside, unless it's changed from its default, the user's \AppData folder. I once had this problem due different circumstances, and ended up having to reinstalling MSFS to get it to work again. Note: the good ol' drive letter is NOT used by this linking mechanism.

Not sure your level of knowledge about overclocking, but here is a rather thorough OC guide that is generic and discusses enough details to give one a good idea of all the facets of overclocking. Even though it's for a 9th gen cpu, the concepts are identical. Above I mentioned using RealBench to stress test my overclock. My comment about my 5.0 OC hitting the low 80's is incorrect running RB, it's really the low 90's...huge difference and a bit nerve racking at that. My NH-D15S gets pushed to its limits around 220W.

In the last few days both Linus and JayzTwoCents have put out videos that the market for most things silicon, and especially gcards, won't normalize until a year from now. Right now industry is in a perfect storm due to multiple factors, one which is highly exasperating is that crypto mining is paying top-dollar to strip the market dry. Things have gotten so bad that GM has stopped production of some models of their cars mainly because they cannot get the computer chips needed for them. I have plans to get a 3080 ti when it releases, though I imagine I'll have a 2 second window to buy one the moment it goes for sale. But I bet that straight away it'll go for around $1,200 US; if so I guess my gtx 1070 gotta last me another a year (gee, that'll make it 6 years old).

Okay, I now understand this plot...my bad. The Y scale is percentage, not FPS! Why they decided to make a plot comparing two data sets where one set is base constant of the comparison, i.e. 100%...oh yes...marketing!

Is it really possible for MSFS to run at 100 fps? Never have I seen that sort of frame rate reported anywhere, perhaps they used all its settings at their minimum. Rather suspicious graph IMO.

Intel puts a time limit on how long a processor will run beyond the spec power 125W limit. So even though all cores of a 10900k at stock will run at 4.9, that'll last for only about a minute usually. After that the multiplier is lowered (41-43) until only 125W are being consumed. MCE disables this behavior, though that setting alone without hand tweaking voltages, i.e. leaving them at default, may lead to unexpectedly high voltages. Though technically not an OC in the sense of pushing up the clocks past turbo speeds, some hand tweaking, as is done when overclocking, is recommended.

Obvious, the biggest factor on how high a CPU can be OC'ed under a NH-D15 is the CPU chip sample quality that's being OC'ed. I think that's that only thing really hold back the OC of my 10700k, which is sitting under a NH-D15S. To hit a higher OC would require voltages which I would not be comfortable using anyways, though custom closed-loop water, or a crazy huge (read: loud) AIO would likely handle it.

Seems that the link given earlier is for an out of date version of RealBench. Here's the only place on the Asus site I've seen a link for its most current version via the Download RealBench button. https://rog.asus.com/rog-pro/realbench-v2-leaderboard/ This tool really kicks the snot out of your processor. Though all its components are realistic use apps, the manner which RB goes about running them is not, at least for what I do with my computer which include video editing and compression. This means to me that when using this tool if you're maintaining stability while maxing your computer's power consumption and controlling the associated temps, everything else you realistically do with your computer is a walk in the park. BTW, my i7 10700k's 5.0Ghz OC is taken to the wall by RB, with temps occasionally hitting into the low 80's (if it wasn't I'd go for 5.1).

The Asus Maximus bios ranks your cooler; my build's NH-D15s scores in the range of AIOs with a 63. This scoring is a component of their AI 5WO. One really nice bios feature is that you can set the max power, in watts, at which your cpu will automatically downclock, great for load level tuning your cpu with your cooler. Out of frustration on the lack of concise information when setting up my OC, I wrote this guide on using Adaptive Mode when overclocking a 10th gen cpu. This guide is a big departure from what was once done, possibly that's why NickN dismissed the info in it.

Well in that light, considering that the stock all core boost clock is 4.7, seeing a 4.8 OC isn't getting one much anyways. When I first saw what Intel was doing with their 10th gen, it seemed to me that they decided to "pre-OC" out of the box. I am just glad a little headroom was there for me to exploit. Bottom line if you want to OC it, top end cooling is a must. I could complain that my NH-D15s is a limiting factor, but I knew that before building my system. In the end my 7/24 OC is 4.9. Though that's a 2% loss as you mentioned, its lower voltage and temps are well worth the trade-off. My 5.0 OC is little more than 6% faster than stock, 5.2 is 10% faster though without HT (yet only 4% faster than 5.0). Most of this is peanuts when you think about it, almost not worth the bother. But heck, when you pay for the best you want the best out of it...right? (plus it's fun 😁)

Panic inducing was a bit of hyperbole, and even though you state 1.3v is withing Intel's voltage limit (whatever that really may be, for Intel has decided that the general public will not have access to the 10th gen CPU technical documentation), I would never recommend using that high a voltage, and I've yet to read anywhere about really needing that much even with massively OCed DDR4. BTW, NickN points out this fact in his guide. OK, yeah 5WO works, but as mentioned YMMV. I got meh 4.8Ghz all core out of it with my i7 10700k. Hand tuning got me 5.0Htz all core, and 5.2Ghz no HT.

Hence that info is not in the same paragraph. I guess I should have been a bit clearer on that point, but it's something people really got to know when using XMP. Seeing those voltages around 1.3v was panic inducing for me.

As mentioned it does, but using Offset voltage is a less than optimal way to go about it considering the alternatives available. I did and found it lacking. I consider the current Asus AI Overclock implementation as having potential but definitely in the realm of YMMV. What is really bad is using XMP sets SA and IO voltages, if they are left at their default settings, at nasty high, and considered by some dangerous, levels.