I’m still trying to get my own arms around this fell beast, but I noticed something like this too on my box.
Prior to yesterdays, “small” update, my intro animation graphics were somewhat jankey - less so after I tweaked my MoBo a bit, but still more like a flip-book than an animation.
After yesterday’s update, the intro animation was remarkably smooth. It’s only at the end where the float-plane is flying in and the “Accessibility”/“Start Flying” buttons pop-up that it becomes jittery. (then it settles down after about 15-or-so seconds.)
Selecting a flight, (DV-20 at UUMO, broad daylight with fantasy weather set), the initial animation prior to the “Let’s get this beast rolling!” button appears was jerky - now it’s smooth.
Frankly I don’t care if they fixed a cat’s hairball or a pimple on someone’s back - whatever it was it appears to have improved the graphic smoothness too.
It was mentioned somewhere from a Community Manager that the update addressed a regression caused from the last update.
“A regression” could mean a whole bunch of stuff and perhaps addressing this regression could inadvertently affect some other parts of the sim. Perhaps the dev team are still trying to identify how such “regression” could potentially be a cause for other issues. Just a suggestion though.
I can’t speak as a developer, (though I’ve done some light dev.), but as a card-carrying member of the Software Quality Masochists of America, I can’t tell you the number of times a fix over here solved a nasty problem way over there at the same time. In one case this bug was caused by not releasing a file-handle or a mutex quickly enough - and it turns out that the stuck file-handle/mutex was ultimately the root cause of that bug over there too.
One very, VERY[1] good Unix software dev once told me that successful software development was one part skill, one part luck, and one part magic!
When I say this guy was good, I mean good. So much so that he was on beer-drinking terms with Tim O’Reilly and had a number of O’Reilly books he’d written. A totally class act.
It seems possible that both the “luck” and “magic” parts played a role for some observed improvements here. I can’t personally comment as I’ve not fired up the sim in a few days.
I’m repeating myself, but it’s because I find it important to mention - I’d have no problem with the state of the game, that’s IF they would clearly state the game’s brutal VRAM usage in their HW requirements list.
They didn’t, which means it’s either a bug or they’re purposefully misleading people into buying the game.
That’s the thing - the game DOESN’T have “brutal” VRAM requirements - it’s clearly a bug when the game requests more VRAM than the system even has when first loading in (as some of the recent posters in this thread have experienced).
And for those of us not encountering the bug at all, I rarely exceed 12GB used out of 16GB available, even under demanding conditions of complex aircraft, complicated scenery and AI traffic.
So it’s a bug but it’s not universal, which makes Asobo’s job that much more difficult in pinning down the trigger(s) and conditions necessary for it to manifest.
If they’re going to market a game that half the known universe is eagerly waiting for, and they’re only marketing it to people who have $20,000+ to spend on a system that’ll run it, this makes less sense than. . . I honestly don’t know.
It makes no sense at all to market a game that no one can effectively use.
Even if you only bought the $70 version, planking down seventy simoleans for something you can’t use because you need a Lawrence-Livermore class supercomputer to run the graphics is a serious problem God help the poor folks who planked down the 200 smackers for the Sooper-Dooper version!
I can totally understand designing a game to be “future-proof” and designing some headroom into the graphical capabilities - but making it require something that may not exist yet, (outside of a Top Secret research facility), is just plain crazy.
I don’t think even MSobo is that crazy. My vote is for a VRAM malloc/free bug.
The intro animation is literally a Video file playing - if you look in resource monitor you can see it reading the video file over and over. I very much doubt the behaviour at that stage correlates to behaviour once the simulation is running.
Parts of the sim experience that were jerky and jittery are not smooth - (er) to the point that it’s a whole new ball-game.
I’m a happy bunny if it’s working better!
There’s also the “Doh!” factor - it runs like a greased pig within their test environment so they assumed figured they were golden.
This almost borders on deja vu, (or maybe PTSD ), because it reminds me so much of my SQA days before I retired. There’s always one more feature that marketing insists has to be in the release, (and the Marketing VP always seems to trump the Engineering/Dev VP), so the poor devs get it rammed down their throat, resources or not, time or not, bugs or not.
I’ve been there and seen the “blood, toil, tears, and sweat” flowing in the isles between the cubicles while dev and QA try to keep up with the endless demands of upper management, (who have zero clue), and marketing, (who have less clue than upper management). So, I never fault the devs - they’re doing the best they can with the crumbs and kopeks they’re given.
My wish:
For the bosses at MSobo to be featured in “Undercover Boss” so they get wise to the reality of what it’s like in the trenches. Their eyes would open so wide it would split their faces!
when it comes to a showdown between marketing and engineering, marketing will always win, they are usually super cool dudes, with a lot of talk, and more importantly, they talk the language of money
the engineer sometimes doesn’t even talk at all
im like that, (im not an engineer) but i understand that being a salesman is something you kind of were born to do, you have the art of convincing… me, i couldn’t sell a life vest to a drowning man… really, i would botch that sale
@DrVenkman3876@Salty9649@IndigoAsp237317@Rivanov I tried using the simulator without Frame Generation. It seems that the VRAM usage drastically went down with frame generation going down off. Perhaps something to test?
I have already tested this in response to someone else’s remarks about framegen. It makes anywhere from 100MB - 500MB difference in enabling or disabling it. A few days at in the Asobo 787 at FlyTampa’s KLAS, with the Strip lit up in the background in heavy clouds around sunset, disabling framegen made a 400MB difference.
Conversely, I tested the same aircraft at Dominick Designs’ KBNA in live overcast weather and it was only 100MB difference.
), because it reminds me so much of my SQA days before I retired. There’s always one more feature that marketing insists has to be in the release, (and the Marketing VP always seems to trump the Engineering/Dev VP), so the poor devs get it rammed down their throat, resources or not, time or not, bugs or not.