Perhaps it maybe worth trying to get this answered in the next Dev Q and A session as there seems to be a lot of uncertainty and variance around the issue
I believe that the cache should be based on two simple issues:
1º - How much can you make available for the size of this cache;
2nd - How much you use the same flight area.
Based on this principle, everyone knows what to use. I turn it off myself because I’m always flying around the world. In this case, I have to assume that my performance may be a little lower, since I have no data stored in the SIM cache.
For me, this is a discussion that goes nowhere because we all have different flying experiences and needs. But understanding can help to elucidate some issues.
Hugs
That’s assuming the data cached would only be unique for that area and never re-used. Do they use some kind of “dedupe” compression technology. So an area may be similar to another area and they would pull that from the rolling cache. So what’s the purpose of rolling cache then? They should have just made the manual cache system better. Again, these are just curious questions. I wish Asobo would clear the air on how all these caches operate, so we are all not spinning our wheels.
Pre-release sources put the total data for the world scenery at a little better than 2PB.
Not sure how many of us have room for the required 2000 1TB SSDs in our MSFS rig. I suggest focusing on specific regions you want to fly in.
I just did some tests with manual cache and for me it’s much better with it switched on. Word of caution when testing… it appears that a restart is required for the newly captured rolling cache to have an impact on performance..
One thing to keep in mind is that its not all about internet speed with rolling cache and performance. Its also about server performance with Microsoft. I suspect a lot of variation in performance is the load variation on Azure servers. Whether you have rolling cache on or not will only have limited impact.
I have been tracking my internet demands with rolling cache and at different times of the day, with cache off, the internet loads seem to be different. And this has a direct impact on stuttering. I also think the cache gets corrupted periodically and clearing boosts performance. Then if you run a flight several times in a row, the cache gets more and more corrupted and starts impacting performance. This is all conjecture, but is the only thing I can guess at to explain my performance degradation over multiple duplicate flights.
Good point.
My research leads to one conclusion… it all depends on numerous factors. So there is no one answerto this seemingly simple question that I asked.
I fly many, many duplicate flights and see no performance degradation.
And I think the inconsistency in performance between systems point to it be an issue external to the various systems. That’s why I suspect at some point it’ll be revealed its a server-side issue or server load issue.
‘www.dictionary.com’
“Elucidate definition, to make lucid or clear; throw light upon; explain”
I was not familiar with this word.
Rolling cache -vs- manual cache -vs- no cache. I think if your in a position to run the FS at different times of the day and compare results, you’ll see as you get closer to primetime in your area, the worse your performance becomes. On occasion the scenery is degraded last night.
Case in point. I got a new X56 stick yesterday, when I first loaded the game I was getting 50FPS the scenenry was crisp and clear (this is on ultra settings). As the day progressed things were pretty steady all day long…until around 5:30-6:00 then things started getting really bad. I was flying in S. Arizona so all there is mountains and cactus. Earlier in the day, the cacti were everywhere, around the 5:00 hour, there were hardly any cacti to be viewed anywhere. As I launched I could see the cacti springing into existance like some fast forward nature flick. and about 20 min into the flight the fps just slowed to a crawl, it started at 30 FPS (was steady 50 earlier in the day) and just started getting slower and slower and slower, until I hit an astounding 1fps…
I took a 1/2 hour break and came back and things were back to around 30 fps. Left the machine on all night and Flew a quick flight and were back to 50…
Ive seen this consistently in this game. After the last xbox crash a few weeks ago, MS announced that their Azure servers were over loaded and that they were building a new data center due to be operational in 2 months (about a 1 1/2 months now).
So lets just wait another 2 months and see if it gets better (Im betting it will).
That is my experience also. I pointed out somewhere that the inconsistency in performance is most likely related server performance. Frankly, if that’s the case, MS should have known this would be an issue. Its not a great advertisement for everyone moving their systems to the Azure cloud.
The fact that you needed internet access to even play the game kind of made me hesitant to buy it back in Sept. but I still had some of that free Govt Money Left, and just being retired I thought WTH…and now to some extent Im starting to regret that decision.
The other thing that I noticed with playing multiuser on, was if there are people who have high pings in your area, it affects you as well presumably while it tries to mediate the ping difference.
Off. Also deleted my manual cache as I find myself taking hops from different areas all the time and my HD space as designated is getting a little on the low side.
I totally agree with what is said. What still do not know is why manual cache at all. Having downloaded 1TB of manual cache I still see scenery changing when flying so 1TB is only a V V small part of data needed as I suspected. As already stated we need the internet to use the SIM and my 40 Mb/s data line is adequate but not great it is the servers that slow everything down.
Wish that I knew what a X56 stick is?
I think it might be a Saitek throttle quadrant and joystick.
Logitech X56 warthog … suckers on the net are paying upward of $500 for them, we have a computer store here that sells em for there real price $240 + 20 tax 260 … and uncle joe paid for it bonus lol
I’m leaning towards “definitely OFF when tinkering with Graphics settings”
I need to experiment with this a little more but…
Was flying around Pikes Peak Colorado this evening at Ultra + Render scaling 150 and getting reasonable frame rates (no lower than 30) even when low level in a fast jet.
Wondered if I could tweak the FPS higher by reducing tree’s to medium and dropping the Render scaling to 100…
Immediately crushed my FPS to 12 FPS which did not recover.
About 2 more flights in same area plagued by low FPS issues from the start.
Deleting rolling cache and turning it off restored some of my performance.
I can’t draw any conclusions from this yet. It was already late and I didn’t perform enough testing of this scenario BUT it does sort of make sense that certain changes to graphics settings might conflict with assets already downloaded into the rolling cache but it was almost as if I had created my own “76T” anomaly out of an airfield/region that was previously working very nicely.
I flew around with everything on Ultra tonight on a i7 and 1080ti after 8 pm with no cache and had no problem getting 40 fps, with PG on 24 fps. When I turn on the cache, every so often I’ll get a burst of stutters and then Im good to go until the next trasmission of data and burst of stutters.
Personal experience today unless your play on a 28.8 baud modem your probably better with it off.