I not would say “less stress for the devs”.. merging branches can be very tricky and there is a effort to offer parallel branches
But it is a kind of a more dev friendly strategy ( I miss the correct words ) … ( and of course it is good for us ).
Users which choose the beta just know that something can works “not optimal”. The beta is the version if a feature x, a bugfix, etc. leave the producers test-stage and is in “theory” ready for delivery. That this is then in “real” often not the case we know, therefore the beta. But the feedback from the users is then more dev friendly like “nope, still not fixed.. xxx happens now…” , or “half fixed, but..” and devs have again time for a second try. All of the critism is more “human” and more “respectfull” ( at least what I know from other forums and what I have in real experienced ). And because very short beta cycles, often only 14days , it is much more effective and we get much quicker “help for specific problems”.
( A experimental beta is completely without warrenty and realy for pre-testing. )
And I mean a continues beta branch. There is no “in meantime” .. The beta branch exist for ever. At a point x (3..6monts) a good beta version is choosen and marked as “a stable release”. Users which only want “stable” get now these version. The beta users are on its own branch and it is nothing to do, no re-download, not necessarly jump to stable because in beta you have already newer features, etc…
I agree, like fbw has it I often choose the experimental because i want the new features. But i still can change back if it gets broken for me. That has mostly been stable anyway though. Only needed to change back one time.
Quickly switching between Stable (Release) and Beta branch?
Surely Asobo should be able to do this. To only download the Delta between Stable and Beta.
But will they do it?
Let me tell you another story (seems I am in story-telling mode today )
Some years ago I wrote a little tool to switch between Retail and Beta version for a racing sim I am part of the closed Beta-Testing Team. Will not tell the name, NDA, ya know.
Time needed:
From wake up on a sunday morning after a great evening in the local pub with live music and lot’s of head-banging and pogo-dancing with a slight hangover to brunch-time.
And another hour with 3 guys testing it.
Pros:
Switch within seconds
Different settings between Retail and Beta are possible
No loss of logbook
Cons:
You need space for 2 installations
Only works with Steam
In case there is demand for such a tool, I might implement a FS-2020 edition of it.
As written, it will only work with the Steam version of FS-2020!
Why not with any other version, ya might ask.
No X-Box at hand (No console will ever cross my door! This is no offence against any X-Box User. It’s an offence against the gaming industry in the past!)
Only have a license for the Steam version and can only test and debug on the Steam version.
With MSFS, you’ll only notice an improvement in performance with DLSS if you’re GPU limited. Most folks are CPU limited. Chances are, if you have a GPU that supports DLSS, you’re probably not GPU limited unless you have a lower tier card (2060) running 4K. You’ll then be able to run at 100% render scaling and DLSS will do the graphical magic to upscale to 4K properly. You’ll still be CPU limited, so don’t expect a 60+% frame rate increase.
No one knows the performance uplift but Asobo themselves. I won’t be throwing numbers on anything! Lets hang tight and see what the update brings! Just by lower the CPU intensive sliders and you won’t be CPU limited.
It also didnt help with all the development branches they had. I still think they have that issue as alot of issues people are having are repeat offenders.
I was speaking on personal experience of having been on 3 to 6 moth releases back when I was a young dev.
The trickyness increases with release cycle and merge times, especially with Git Repos.
Heck, sometimes when I do dev work these days, I can spend 1/2 to 3/4 project time trying to get those merges done and pull requests straightened out. I hate Git with a massive passion but it’s how the industry operates today.
Don’t get me wrong, I also appreciate Git. It’s just flawed on some levels (like non sensical naming conventions of commands) and how it’s often implemented.
But the more frequent the merging/PRs, testing cycles, etc., the easier it all becomes.
That’s true, but most of the CPU limitations in this game come from poorly optimized code, not from CPU specs being too low. That’s why I wrote that I hope it will be reworked with DX12 (which is one of the requirements for DLSS), so that the threads can use the available cores a bit more effectively.
I can only speak from my experience in other games, with my setup (10700K, RTX3700, 64GB RAM, 1440p resolution) I had huge fps gains using DLSS.
But yes, MSFS is playing on a different level with its poorly optimized engine.
You can’t find another sim/game with this kind of drawdistance. It needs to be rendered all around us equally much and then saved in caching. I have never experienced a more detailed world in a flight simulator than this. It also needs to do much more calculations because it simulates air and stuff that the plane use to fly in. It’s more complex than a simple game. And thats why it uses alot more cpu than gpu.
not sure what they use, but yes merging is not allways fun ..
“Small” features and short release-cylces will help to reduce horrible merging stories. E.g. merging 6 months of work sounds more difficult, as merge a “change” into main if a beta-phase show the developers that the “change” works fine. But yes, it can be still tricky , but hwo we say here: “If it were easy, anyone could do it.” )
do you know one single line of source-code which you can proof such statements ?
Optimizations are allways possible, and sometime necessary. But speaking about “poor”, I don’t know…
Yeah how can we actually see if it’s bad optimised? I feel this sim is much more optimised than any other flightsimulator ever. Name one flight simulator with the visual quality better than this and at the same time running on 60fps if we set it up propperly for that without glasscockpits and using default ga planes. Thats how i meassure optimization. It’s maybe a bit worse since release but there is much more 3rd party addons in use now than at release. Many state it’s worse performance using a complex newly released airliner whike testing.
In some ways its well optimised. The engine copes with a lot of load that would make other sims including XP11 weep. Just recently Aerosoft have been testing highly complex models and have been amazed at how well the engine copes with millions of polys whilst still maintaining a high frame rate.
But we also have some poor code in the sim that causes hardware bugs. Keyboard right arrow key set to increase heading but for example will behave as if they key is held down sometimes. You cannot plug in or pull a USB peripheral without the sim freaking out and either freezing whilst it tries to recover or crash entirely. USB is by its very nature designed for quick easy connection in a live environment and the sim cannot cope with it. That’s poor. Other sims dont bat an eyelid when you do this so it can be done.
For example, X-plane has a lot more trouble with a bit of complex scenery and a detailed aircraft than MSFS. The fps is much better with MSFS 2020 than with X-plane with the same rig.
I often choose the experimental because i want the new features. But i still can change back if it gets broken for me. That has mostly been stable anyway though. Only needed to change back one time.
)
..