How's your PC?

This is a great idea! Will help developers quickly rate their routes without having to ask people to try them out, so very efficient!

I don't agree. All it does is announce the fact that you happen to have a better/worse machine than the next man. It is of no relevance, because it is possible to assemble work on a Ferrari superPC which will still run on a Lada (just!!) You are missing the point entirely, if rating is to be meaningful it is the route software which needs quantifying. Hardware assessment does not do this.

done it, was a weird was to performance test my pc though! planes cars and women!

graymac wrote:I don't agree. All it does is announce the fact that you happen to have a better/worse machine than the next man. It is of no relevance, because it is possible to assemble work on a Ferrari superPC which will still run on a Lada (just!!) You are missing the point entirely, if rating is to be meaningful it is the route software which needs quantifying. Hardware assessment does not do this.

I agree too. Passmark is obviously not a real world test, while average framerates on routes are of more relevance. (I can for instance run FBT on 45 FPS, while any NYCT route coupled with the R46 causes 16 FPS thanks to AMD cards not liking Open's rendering mode)

By the way, the software is shareware (I think), so please make sure to
remove it when done with testing, or before the trial period expires.

It looks as if a payment is expected if the software is to be used permanently. It isn't really much use in the search for a simple route rating procedure.
Developers would be as well off indicating approximate system expectations for satisfactory running using deliberately broad terms.
For instance I could state that Ballyfeckin was "suitable for medium spec system" . . . with reasonable confidence that is true.

We should rather have a defined OpenBVE-specific benchmark like "Ballyfeckin, 600m sight, antialiasing 2x, sharp transparencies [...] XYZ fps at Knockmore"

Erm... It is good very much 'cause that's something I can compare directly. If somebody has 300fps on mentioned benchmark and states he can run his own route (with the same settings) at about 40fps, I know I don't even have to download it. HW data on the other hand is very incomparable. I've once had a 600MHz PC with 512MB RAM and some pretty standard (and pretty old) graphics card. German LokSim3D ran perfectly on it. Now I have my laptop with dual core 1.8GHz CPU, 2GB RAM and an Intel onboard graphics card, which while it is poor for today's standards still has better parameters in every aspect than the graphics card of my old PC. And guess what? LokSim3D doesn't run on it well, FPS are about 60 times lower. One of the reasons why I ran into OpenBVE.

I am serious. And if you read carefully you'll see I'm comparing a (near-)state-of-the-art-laptop with a >10 years old desktop PC. In this case, the laptop is far more powerful. Many programs which wouldn't run on the old one are running very well on this laptop. And yet this didn't help with LS3D being unable to run on the laptop. That's why the only meaningful benchmark for OpenBVE is OpenBVE itself. The PassMark rating tells, simply speaking, how well PassMark runs on your computer, not much more. OpenBVE can run totally different on two different computers having the same benchmark rating or comparably on two computers with totally different benchmark ratings.

We've visited this benchmarking question before. It isn't straightforward and we didn't find a satisfactory solution then. Quork is right about unpredictability in comparing apparently similar machines which (seemingly unexplained) end up performing completely at odds with expectations.
It's not very scientific, but my approach remains with getting a few good people I know who have "average" or "less than average" machines available for them to try the beta route for me. If they are happy it seems to work fine when I do release the stuff. I have had hardly any problems with huge numbers of downloads. The very few users who didn't get it to go first time either lacked a component (such as BRsigs) or ignored proper folder structure. No other faults have been found. BWR-2012 has had all textures optimised and kept to minimum size consistent with quality/performance balance.
The new Welsh work is rather less restrained. We will see if that causes problems before very much longer.

Quork wrote:That's why the only meaningful benchmark for OpenBVE is OpenBVE itself

Quork wrote:We should rather have a defined OpenBVE-specific benchmark like "Ballyfeckin, 600m sight, antialiasing 2x, sharp transparencies [...] XYZ fps at Knockmore"

I suppose, if there was an agreed route/train (with or without externals)/location/settings, it could be interesting to see what sort of variance there is, if for no other reason than to see how large the variance actually is, but at the end of the day, such a test is more use for a developer (to understand how their work may, or may not, perform), but less so for the user- my laptop is what it is, and with the wife, two children, a house, a car, a cat and a chocolate habit, it's staying that way.

Sounds reasonable! From which can I assume that your laptop is what you might call a "medium" performer? If that's true, do you have difficulty running any particular routes to a fair standard? And just out of curiosity I tried the test, not that I understand it all - it still seems like comparing willies to me !!

Well I think it's good for the user as well. Let's say we agree on route X and settings Y as a benchmark. Now if a route is published, the dev can say they have A fps on route X with settings Y and B fps on their released route with settings Y. This way I as a user can estimate how the route will run on my PC.
e.g.:
Derryck releases Brno and states he has 30FPS on it with standard settings, while he has 300FPS on the reference route with standard settings. I for myself know I have 30FPS on the reference route with standard settings - so I can estimate I'll have a slideshow only on Brno.

why dosn't everyone get together and choose 3 or 5 routes. decide the 'best performance settings' and then get the fps from each of the chosen routes, then do the math and get the average, and then from there create a new 'benchmark test'.

id choose probably the following routes.

NWM




2008 06.30 1A07
Express Maybank -Hammerwich via Mainline

Recommended traction:
Class 390

Weather types:
Sunny/Cloudy/Foggy

FPS achieved = ~ 60

RSR

Birmingham Cross-City South
Version 1.31.11 (December 2010)
[railsimroutes.net]

158_Spring_2005_1200_Rain_Overcast

FPS acheived ~ 60

Northern Line (visible prior train)

OpenBVE Northern Line. Edgware - Morden via Bank 1998

FPS acheived ~ 60



openbve settings on test

full screen mode 1024 x 786 x 32 bits per pixel

anisotropic filtering = 16 - sharp

viewing distance 600m with motion blur off

sounds - high - number of allowed sounds - 70

all detail of simulation with a tick in the box.






my average score for fps then would be 60 FPS, i put in roughly because the simulator wanted to give me a higher FPS but the 60hz on the monitor wouldn't allow it, all im trying to suggest is maybe we choose 3 routes, that are readily available to download, and then a certain route file from each route, ask users to them press f10 and get their fps from all the routes, then add them numbers up and divide them, that sould give them a rough estimate of their machines 'openbve' benchmark. maybe with people hitting an average of 60fps or more should be 100. and then work the math for individual numbers down. at least then its non of this downloading silly programs and such, nice and simple with a bit of simple math

Sounds reasonable enough to me. Though I'd suggest trying three more different (in detail, in dev strategy etc) routes, would probably be better comparable. e.g. Solyomvar-Kiralylighet, FEVF and Chashinai.

Hi all,

Ran the test and it almost made me sick with the spinning visuals, PC rated at 2410.



No idea what it all means , in real world I can run any route.

Bye for now,

Stephen.

It means you've got a decent machine there, lucky you!!!

Taht's quite a good result there, Stephen. Did you get your link to a banner?
@Graymac: What's wrong with your banner then? You converted it into a fig or something?

Hi all,

Ok, the PC is configured thus:

Intel i7 quad core 2.93GHz, Gigabyte MB, 16Gb RAM, Nvidia GTX460 1Gb, 240Gb SSD and 1Tb HDD, BluRay/DVD combo all in a Coolermaster case with a 750W PSU. This machine runs Windows 7 x64 and is primarily used for the creation of a magazine. This PC is now over 2 yrs old and still manages to handle all I throw at it. SSD was added recently and makes a big difference. It also has two (2) monitors attached the main being a 23" LG and second one a 21" Benq. This is great as you can have reference material, video, pictures or even creation tools running on main while RouteViewer sits on second screen and you can update it and see your changes.

The machine I do most the development on is my daughter's laptop as it runs WIndows 7 x32 OS and therefore most of the older utilities / tools for creating items for BVE run fine with resorting to DOSBOX etc. I often create a whole lot of objects on laptop and save them to the server drive or if on the move a usb stick then download them onto main PC from there. By the way most modern modem / routers and OSs can support the connection of an external drive to modem / router and all other computers on wireless network can share files once access has been granted.

Regardless of what spec your PC is, it is like giving me a very expensive camera and expecting it to make me a better photographer. As you know it won't and what is needed in the end is talent (and time). So to those talented few amongst us I tip my hat to you and look forward to seeing / drving your creations soon. No idea Dexter, I think I downloaded an older version but it was just an exercise and the software is no longer resides on my PC.

Bye for now,

Stephen.

......................After that there's any point in posting my laptop specs. The word Morris Minor springs to mind.............................!

But no problems, it does what it says on the tin!

Think i need to put another 50p in the meter..

.