OldProf wrote:I'm pretty sure that loading times are more subjective than objective.
I ran tests on versions going back to TS2013. In each one, I made certain to have a fresh boot and start Steam first (online mode), and then check logmate for errors while loading, and each test was done multiple times (again, on a fresh reboot) to ensure I wasn't getting one-offs (cold boots wouldn't have changed things in any measurable form except in loading the OS). The specific procedure for each instance of TS:
- Wait for all processes to load in Windows
- Start Steam
- Start game
- Time menu loading
- Time editor loading
- Exit
- Reboot
- Code: Select all
________________________________________________________
TS2013: v27.5a No Cache Cached
========================================================
Icon click to Main Menu = 0:22.81 N/A
Edit Blank Scenario = 0:41.83 1:09.10
Scenario Exit to Drive mode = 0:03.76 0:03.76
________________________________________________________
________________________________________________________
TS2014: v42.9a No Cache Cached
========================================================
Icon click to Main Menu = 0:20.39 N/A
Edit Blank Scenario = 0:26.58 0:32.50
Scenario Exit to Drive mode = 0:02.85 0:02.46
________________________________________________________
________________________________________________________
TS2015: v52.2a No Cache Cached
========================================================
Icon click to Main Menu = 0:21.71 N/A
Edit Blank Scenario = 0:55.24 0:41.47
Scenario Exit to Drive mode = 0:08.93 0:09.07
________________________________________________________
________________________________________________________
TS2016: v53.9b No Cache Cached
========================================================
Icon click to Main Menu = 0:30.71 N/A
Edit Blank Scenario = 1:30.56 1:07.55
Scenario Exit to Drive mode = 0:11.85 0:13.29
________________________________________________________
TS2014 is the clear winner and that TS2016 falls flat on its face - and both use similar Flash loading screens - only TS2013 got confused about some of the newer Flash parameters, likely in the route loading ad-driven slide show). When the initial 2016 update was released, I felt that load times were markedly better than the previous 2015 version, but all that went down the toilet with the latest hotfix. Right now, TS2016 is more useful as a door stop.
Start, go make coffee, come back and load editors, go get your cup-o-joe, and hopefully the route would have been already loaded by your return.
Note: TS2013 appears to have had a bug with loading or processing blueprint cache. I ran the tests several more times to make certain I hadn't goofed, but it all checked out. Also note that TS2013 didn't support irregular assets blocks, which might be a factor in what's going on in the background. I included this version mainly as an illustration and to plot a trend, and also I didn't bother to test earlier versions (I remember them as a bit slow anyway).
As an aside, although the original Rail Simulator was a bit clunky, the loading times are phenomenal. Of course, it didn't have to load the newer shaders and the bloatware menus we have today, but it's a snappy chappy and FPS is through the roof. Main icon to start was under 2-seconds, and less than 15-seconds to enter edit mode from the main menu. Tehachapi runs, albeit with the obvious errors regarding special shaders and there's no super elevation or distant scenery, and the best part is that it doesn't have to contend with Steam. It does work, however.
All in all I say we're going downhill and at a pace most would hardly notice from version to version. Having the game run most all the time everyday, I tend to notice immediate changes, at least be suspicious of their existence, so I didn't feel I was blowing smoke with my original comments. Of course, it was only the tests that confirmed my belief that TS is turning into a bloatware nightmare. IMHO, if this trend continues, we're in big trouble with the UE4 version.
The big question though, does DTG even care?