List of Best (and Worst) Optimised Routes

[Kyriel]

Hi everyone.

The list doesn't exist - yet.

I need your help to put it together.

I think it will make an excellent resource when it comes to buying routes, especially for those of us who do not have nice computers.

If you have have a route which runs particularly well, or one which runs particularly badly, say so in the comments. Any detail would be appreciated, but a simple "This route is(n't) optimised" will suffice.

If enough people reply I will create a community guide to share.

PS If this list does exist and I just haven't found it, please tell me where it is haha

[Axe]

You would have to have a lot of decent results to extrapolate any meaningful data enough to be called a worthy list.

1 - All would have to be subject to the quality of the data but moreso the data provider themself.

2 - PC specs.

3 - Proper recording of some data parameters.

If no stringent (enough) guidelines are set you could get data from many varied sources enough to not be able to quantify the results and even trust them.

That is why even Amazon don't have any worthy stats and only ask for purchasers giving a number out of 10.

I would think more clearly about your data collection method and remove the chances of people who are not experienced giving you duff results. Perhaps list the/some routes and ask others to give a 1 to 10 rating. Yes it's just as meaningless and would not be representative at all in my opinion other than 2 or 3 standing out either end of the scale. You would be missing out personal unrequired comment but at least you might get a number...

Make things too complicated and you won't get many replies and make it too simple and your results are frankly meaningless.

Just my opinion. Good luck though :-)

[electronx]

Generally,
- overhead lines are for some reason resource intensive (especially in yards and big stations)
- several 2013-2014 runs bad, older runs well, many new ones run well (I have some super high density routes running well)

For example,
- Canadian Mountain Passes, Norfolk Souther Coal District (nice 3D routes) and Fort Kent & Eagle Like (ultra-high density 2D route) run just fine,
- Oxford-Paddington or Edinburgh-Glasgow just about cuts it (despite being old), old Woodhead (electric) is fine,
- Kings Cross - Peterborough and München (-Garmisch) is notably terrible.

Often it's not just about the route, but some trains. For example the Hitachi Class 801 is a true FPS killer.

[Axe]

^
I
I

See what I mean OP? Put that in a database.. LOL

[Kyriel]

Thanks for the responses, both of you. Yes Axe, I totally take your points about the quality of data. But I am of the opinion that information of the quality which electronx provided is much better than nothing at all. Another 10 or 20 people with a comment similar to electronx's, just naming a few routes which run well, ok, and poorly, and it would be possible to create a rough list of routes which people have said run well, ok, and poorly.

I'd appreciate it if we could avoid having a big discussion about data collection and power and robustness of analysis - that's my day job :P

[Kyriel]

Originally posted by electronx:

Generally,
- Kings Cross - Peterborough and München (-Garmisch) is notably terrible.

I find these to be amongst my worst routes, too.

[Axe]

Originally posted by Kyriel:

Thanks for the responses, both of you. Yes Axe, I totally take your points about the quality of data. But I am of the opinion that information of the quality which electronx provided is much better than nothing at all. Another 10 or 20 people with a comment similar to electronx's, just naming a few routes which run well, ok, and poorly, and it would be possible to create a rough list of routes which people have said run well, ok, and poorly.

I'd appreciate it if we could avoid having a big discussion about data collection and power and robustness of analysis - that's my day job :P

OK :-) We sometimes never know who we are talking to on forums so I hope it didn't sound patronising as far as I knew you could have been a youngster with a good idea and I was just suggesting some guidance ;-)

Yes, I agree about something rather than nothing is OK, but let's now hope others will follow the path of electronx for the purpose of the reasonable consistency you need :-)

Like I said and mean't - good luck as I think for some this data would be very informative indeed if held to reasonable scrutiny. Clearly in your case you are that man. All the best :-)

[Phase3]

There are probably too many variables to get meaningful results, even the size of the route and scenario could affect the score. You might need cluster (and/or Factor) analysis to disseminate the results.

I created a table of all of my routes at that time and logged, cpu usage min %, cpu peak usage, cpu average %, RAM usage GB (mean of 4 per route running 4 different types of scenarios (day, night, good weather, bad weather) for a route.
For example Just Trains Western mainLine, cpu usage min 8%, peak 38% Avge 27%, and RAM usage was 6.4GB (includes 2.6GB for the OS).
Windows base was: cpu min 1%, peak 5%, Avge 5%, RAM usage: 2.4GB.
The Woodhead (original) cpu min 11%, peak 30%, Avge 11%, RAM usage 4.1GB (includes 2.4GB for the OS).
None of the figures really affected the running/quality of the route, at the time I ran everything locked at 30 fps and it had to be almost stutter free. that was my interpretation of a good route!
This was running an i7 Haswell overclocked to 4.2GHz, (water cooled), 2100 RAM, 1920 x 1200 Monitor (60Hz), GTX 970 all on a gaming mothernoard.
Changing to a GSync monitor, with an i5 6th gen cpu @ 4.2GHz, GTX 1080 etc I get similar figures with slightly higher smooth frame rates.

[Woodo]

Most routes work well. Turning shadows all the way off has the biggest impact for me. FPS doubled in built up areas after that.

[IC125]

Originally posted by Woodo:

Most routes work well.

Again this depends heavily on a number of factors, namely the machine you play on. Some may be able to get good FPS in, say, Central London but I would get very low FPS purely due to my laptop. Most modern day routes have a higher standard of scenery too so may have lower FPS potential. Paddington on the old GWML route from Kuju days gives me good FPS due to its aging detail levels. It all does depend heavily.

[Felix.AVMP]

As this game is 32bit in its nature, it is possible to draw conclusions even without taking hardware differencies into account - there are always two limitations:

- usable amount of RAM (and this one is pretty hard)

- disc subsystem speed (or, if you prefer, distribution of the assets, lately, this is not an issue, as most recent payware routes from DTG are monolithic in nature, however, 3rd party routes, no matter whether payware of freeware use various sets of asstes and this affects their performace considerably).

Without much further ado - for example route Koblenz - Trier is in its default state pretty bad no matter on what computer, as this route even in rather basic scenarios practically overloads the game engine itself, which in turn renders some of the 3rd party/workshop scenariois on some of the computers unplayable - and what is really interesting, this is rather independent on actual hardware configuration (provided, that the hardware is at least good enough to run RW/TS).