v360 thread

[Redcoat_1812]

Numbers and letters indicate which train has that section of track reserved (0-9 for the first ten trains, then A-Z). "=" means it is an unreserved section, "/" and "\" are passing places/switches that the dispatcher has a special interest in, "?" is a reversing point and "@" is where that train currently is. I think I got that all right.

Thank you James, great stuff mate.

I noticed ( rightly or wrongly ), that the physics are a bit more realistic than MSTS, even at this early stage.

My player train in the pic was slogging up that gradient and ended up down to 17mph, before crawling over the top to the downhill section.

Exactly the same activity run in MSTS itself, it goes like a racing car.

Thanks for the info, James.

Mike.

[hnt4redoktober]

Out of curiosity,
How are you guys getting such amazing FPS with Open Rails, I have been getting 4-5 on routes like Seligman and CSX W&A and 11-13 on the latest version of monon. This is with Dynamic shadows and lights on, if that helps.

[muskokaandtahoe]

Out of curiosity,
How are you guys getting such amazing FPS with Open Rails, I have been getting 4-5 on routes like Seligman and CSX W&A and 11-13 on the latest version of monon. This is with Dynamic shadows and lights on, if that helps.

Your machine is bottlenecked somewhere and theirs isn't. Part of the answer can be in the OS -- WIN 7 does a better job w/ multi-core CPU's that do the previous OS's. Open Rails also expects the GPU to do much more work than does MSTS so if your GPU is weak and your CPU strong you can actually see worse performance in OR than MSTS. Motherboard GPU's are often pretty sucky so that too could be the problem (things I read say MB GPU's simply are not up to the task of delivering good frame rates for most games; I dunno if that's a fact but I do come across that assertion fairly often).

Does your version of OR have the extended f5 display - the one that shows thread workload (e.g., render, update, and loader)? If yes, what is the data there?

[hnt4redoktober]

It basically states that my laptop does not have the power to process OR. However, that being said I am truly impressed with OR. WOW what amazing results! Good luck to you guys, hopefully I can get a Desktop that can truly do it justice! Keep up the great and outstanding work with this!

[csantucci]

This is a message to the OR developer of the dispatcher.
I have found some AI behaviours that are strange to me, and will explain them here with the help of screenshots where also the Dispatcher Information can be seen. Trains of interest are
-train #5, a freight train that has 13:59:30 as starting time and that appears from the lower part of the images
- train #3, a DMU that as 14:00 as starting time and that appears from the upper part of the images.
In this first image

https://www.interazioni-educative.it/Varie/OR_Problem/OR1.jpg

on the Dispatcher Information it can be seen that the last part of the path of train #3 is reserved to train #5; this is correct.
The two trains aren't yet visible.

In image OR3

https://www.interazioni-educative.it/Varie/OR_Problem/OR3.jpg

train #5 passes by and train 3 can be hardly seen between train 5 and the covered platform. Situation is still acceptable.
In image OR4

https://www.interazioni-educative.it/Varie/OR_Problem/OR4.jpg

a problem can be seen. Train 5 is farther away and can no more be seen; on the Dispatcher Information for train 5 at the left of the @ characters and of the / character two track nodes still contain a 5, that is they are still reserved to train 5; these sections should be free. Being they reserved to train 5, train 3 is constrained to stop.
This can be better seen in image OR5

https://www.interazioni-educative.it/Varie/OR_Problem/OR5.jpg

where I ensure you that train 3 is stopped, even if it shows on the Dispatcher Information a speed of 1,2 mph (and this is another strange thing).
In the meantime the engine of train 5 has uncoupled from its wagons, and proceeded further, but:
the @ char has disappeared from the train 5 line, and its speed is increasing and increasing (at the moment it is 134,7 mph); pls. notice that the @ char hasn't disappeared after uncoupling, but when reaching the reverse point. As I already wrote in another post, this malfunction seems to appear only when the AI train is out of sight.
So I restarted again the activity, mantained the AI train in sight, and this latter problem did not appear, although train 3 still remained stopped because of the two nodes reserved to train 5.
Only some minutes later, when engine 5 almost reached the second reverse point, the two nodes wrongly reserved to train 5 were released, and train 3 could proceed. This can be seen in image OR6

https://www.interazioni-educative.it/Varie/OR_Problem/OR6.jpg

Hope this helps and keep the excellent work!

Carlo

[lindsayts]

Your machine is bottlenecked somewhere and theirs isn't. Part of the answer can be in the OS -- WIN 7 does a better job w/ multi-core CPU's that do the previous OS's. Open Rails also expects the GPU to do much more work than does MSTS so if your GPU is weak and your CPU strong you can actually see worse performance in OR than MSTS. Motherboard GPU's are often pretty sucky so that too could be the problem (things I read say MB GPU's simply are not up to the task of delivering good frame rates for most games; I dunno if that's a fact but I do come across that assertion fairly often).

Does your version of OR have the extended f5 display - the one that shows thread workload (e.g., render, update, and loader)? If yes, what is the data there?

I have used OR on three machines using XP service pack 2. Better frames can be had with 4 cores than with 2. It definitely goes OK as long as the graphics card is something like Nvidia GTX 260 or a Radeon 4870, The slowest system is an Intel D975XBX2 MB with and Nvidia GTX 460 and a Quad core Q6600, lowest frame rates around 70 to 80 FPS. I have tried an Radeon 3870 on this last system with only a slight increase in the process wait times (the frames rates did not alter).

My second system had an E6850 two core Duo core CPU in it, this was swapped last week with a Quad core Duo core Q8400 nothing else in the system was changed. The frames rates increased around 20 to 25 % and the process wait times dropped from an average of around 25 % down to an average of zero leaving OR perfectly smooth. The MB in this system is an Gigabyte P35 DSO3P, graphics card Radeon 4870.

As I understand it one of an on board GPU's main problems is it uses the main system memory as video ram which means the cpu's and the GPU has to fight to use the memory.

[sniper297]

"things I read say MB GPU's simply are not up to the task of delivering good frame rates for most games; I dunno if that's a fact but I do come across that assertion fairly often"

"As I understand it one of an on board GPU's main problems is it uses the main system memory as video ram which means the cpu's and the GPU has to fight to use the memory."

Correct. In addition to having to share RAM, the GPU needs to use dynamic RAM to emulate video RAM, and since video RAM is specifically optimized for processing pixels dynamic RAM never can do the job as quickly. A really really powerful high horsepower laptop can run the latest games with no noticeable difference, but given equal horsepower any desktop will outperform the equivalent laptop simply because the desktop don't need to compromise performance for the sake of power usage and portability - desktops don't need to get max hours out of a battery charge and you don't pick them up and carry them around with you. Want to play 3D video games, got a laptop, expect performance issues.

[conrail1]

I have used OR on three machines using XP service pack 2.

Is this 32-bit XP or 64-bit XP? If it is 32-bit XP why isn't it up to date with SP3?

lowest frame rates around 70 to 80 FPS

That doesn't tell us much, what route,what activity, what resolution and what AA was used.? How about some screen shots?

this was swapped last week with a Quad core Duo core Q8400

Intel's Core 2 Duo is a “dual core” not a quad core.

leaving OR perfectly smooth

I suggest you do a little more testing as v360 has introduced stutter with long train consists.

[conrail1]

desktops don't need to get max hours out of a battery charge and you don't pick them up and carry them around with you.

Nor does a laptop if you know how to properly set it up for running performance applications -

https://www.trainsim.com/vbts/showthread.php?296268-Laptop-performance-with-OR-v360

[NF1-800]

I think i need some help cause i`m pulling my hair out...
I have OS W. XP SP 2, Direct x9.0c, NET Framework3.5, XNA as well...but i still cannot run OR...the "Don`t send error" appears when i want to launch the game...
Any ideas?

[lindsayts]

Is this 32-bit XP or 64-bit XP? If it is 32-bit XP why isn't it up to date with SP3?

XP 32, its only SP2 as it works OK and the machine is not and never will be on the net or any network for that matter. I run Linux (mainly Debian testing) on my network here, XP is ONLY used to run MSTS and now OR.

That doesn't tell us much, what route,what activity, what resolution and what AA was used.? How about some screen shots?

SAR Mid North and Barossa Valley route, activity is afternoon Angaston Passenger train , This is a South Australian Route. The frame rate given is in an early part of the route where the Frame rate in MSTS is in the mid teens (13 to 15 FPS)

Intel's Core 2 Duo is a “dual core” not a quad core.

Core 2 Quad then............

I suggest you do a little more testing as v360 has introduced stutter with long train consists.

A train sim will never be all things to all people as it depends what individuals are interested in. A single desk top machine has a limited amount of processing power availible and it CANNOT do everything. Processing the indivdual behavour of 100 or more wagons takes a significant amount of steam to achieve. From my experiments in railway and terrain modelling (done on Linux and still in progress) doing highly detailed terrain and accurate modelling of long and/or many trains will difficult to achieve on a single system (Note 1). The obvious way around this is to do the mathematical modelling of the railway itself on another system something I am currently working on and as OR is designed to run on multiple systems something that the OR developers may be interested in.

Note 1: Calculating the effect of each wagon on the ones around it requires multiple passes through the mathematics down the string of wagons to work out if a particular section of the train has closed up or stretched out and requires a great deal of processing power to do correctly.

I am not an OR developer just an individual that is greatly interested in accurate and detailed railway simulation mathematics.

[conrail1]

XP 32, its only SP2 as it works OK and the machine is not and never will be on the net or any network for that matter. I run Linux (mainly Debian testing) on my network here, XP is ONLY used to run MSTS and now OR.

As far as performance is concerned I’ve tested XP/SP2 and XP/SP3 and there is no difference so completely updating the OS doesn’t hurt anything, then again it never has in my experience.

Not to get off the subject but I’m using various versions of Linux also, right now I’m using Ubuntu 10.10 (32-bit and 64-bit), and it’s without a doubt one of the best operating systems I’ve used. It’s amazing that something this good is free.

SAR Mid North and Barossa Valley route, activity is afternoon Angaston Passenger train , This is a South Australian Route. The frame rate given is in an early part of the route where the Frame rate in MSTS is in the mid teens (13 to 15 FPS)

Is there any way you can post some screen shot’s? I’m always interested in trying new routes if they accurately follow the prototype their modeled after.

A train sim will never be all things to all people as it depends what individuals are interested in. A single desk top machine has a limited amount of processing power availible and it CANNOT do everything. Processing the indivdual behavour of 100 or more wagons takes a significant amount of steam to achieve.

I’ve used professional train operation simulation programs at work that do a very good job at accurately simulating buff and draft forces as well as the air brake system on relatively low spec machines.

With a decent, up to date game engine and core program you’d be amazed what can be done when the software properly takes advantage of system resources.

OpenRails v128 was pretty much completely smooth with my quad and hex core based machines and the development team has acknowledged that there is a stuttering issue with v360 –

https://www.trainsim.com/vbts/showthread.php?295992-v360-thread/page15

[lindsayts]

As far as performance is concerned I’ve tested XP/SP2 and XP/SP3 and there is no difference so completely updating the OS doesn’t hurt anything, then again it never has in my experience.

Not to get off the subject but I’m using various versions of Linux also, right now I’m using Ubuntu 10.10 (32-bit and 64-bit), and it’s without a doubt one of the best operating systems I’ve used. It’s amazing that something this good is free.

Yes its not bad, I have been using it since 1993 kernel 0.96d.

Is there any way you can post some screen shot’s? I’m always interested in trying new routes if they accurately follow the prototype their modeled after.

I cannot speak for the accuracy of the Barossa Valley Route as its a fairly new route and I do not know that area of South Australia well. I used the route as its a known killer of MSTS having very low frame rates, runs well on OR though.
For an acknowledge very good SA route have a crack at................
http://msts.steam4me.net/routes/sar_adlv2_msts.html
This route has been acknowledged by quite a number of old time SAR drivers as turning back time very well. I do not know the whole area but the parts I do know are VERY well portrayed. The developer spent 6000 hours doing the route and made an enormous number of custom objects for it. There is a total of around 600 miles of main line including the entire 305 mile trip from Adelaide to Mt Gambier. Note on the site additional files required there is a link to a pdf file there that contains hot links to a large quantity of rolling stock required for the route.

I’ve used professional train operation simulation programs at work that do a very good job at accurately simulating buff and draft forces as well as the air brake system on relatively low spec machines.

With a decent, up to date game engine and core program you’d be amazed what can be done when the software properly takes advantage of system resources.

OpenRails v128 was pretty much completely smooth with my quad and hex core based machines and the development team has acknowledged that there is a stuttering issue with v360 –

https://www.trainsim.com/vbts/showth...-thread/page15

You likely have a point here, I am still doing much work on this issue but the combination of high detail graphics and half a dozen trains totalling 100's of vehicles being controlled accurately WILL be interesting. I am doing the terrain in quite a different way to allow more accurate depiction of earthworks and I have yet to try much in the way of optiomising the code instead at this stage going for complete accuracy.

[Erick_Cantu]

Brake triggers don't seem to work in terms of looped sounds (such as squealing brakes). This means that I can disengage the train brakes, but it most definitely sounds like I have wheels being dragged along... !

Some clips seem to want to play at rates they shouldn't, and this may be due to sampling rates above 11025Hz. I haven't been quite thorough in diagnosing it... I'll look at it later.

Lastly, more a suggestion than an observation. Like MSTS, diesel engines smoothly accelerate and decelerate at a constant rate with no respect to inertia. I think it would make a massive difference in the overall sound (and feel) of the sim if engine inertia were modelled. It could be based on simple derivations from power (effectively simplifying engine size and thus weight to a coefficient relating to the power output in the physics - it would be a 'close enough' kind of thing) and amount and speed of RPM change. Real diesel engines overspool and come down on the way up and underspool and come up on the way down.

[conrail1]

I cannot speak for the accuracy of the Barossa Valley Route as its a fairly new route and I do not know that area of South Australia well. I used the route as its a known killer of MSTS having very low frame rates, runs well on OR though.
For an acknowledge very good SA route have a crack at................
http://msts.steam4me.net/routes/sar_adlv2_msts.html

Thanks for the link; I’ll have to take a closer look at that route when I get a chance.

You likely have a point here, I am still doing much work on this issue but the combination of high detail graphics and half a dozen trains totalling 100's of vehicles being controlled accurately WILL be interesting.

If you page down to the “4 October 2010” section of this “What’s new?” page on X-Plane 10 development there’s some interesting info on what can be accomplished by taking advantage of multi-core CPU’s (16 cores with Hyper-threading enabled to be exact) –

http://www.x-plane.com/pg_news.html