[markus_GE]

Well, I can´t say I have the best System ever, but a dual core i5 at 2.27 GHz (non-overclocked Rating), 4 gigs of ram, ATI Radeon HD5650 onboard grphics (it´s a VAIO Laptop) at 2733 MB Vram should do a better Job than the above...

[copperpen]

Your processor is better than mine which is an E7500 slightly overclocked to 2.93Ghz and my video card only has 1024Mb of ram, but then it is an Nvidia GTX 650Ti, which while it is good, is actually near the bottom of the 600 series in pure performance.

[markus_GE]

Well, but I guess you have a stationary PC, not lap top - so cooling is no Limitation for your System, whereas some of that silly VAIO control Systems on mine always notch down GHz and so on to prevent overheating - actually good for my machine, but sometimes a bit... especially as that way I can´t get full performence anyway...

[copperpen]

Have you considered getting a laptop cooler?.

[markus_GE]

It´s already sitting on some pieces of Wood, 2 cm (about 1 Inch) above my Desktop (the one made of Wood ;) ). Yeah, have considered it, but in Austria These things aren´t cheap and at the time, I´m a bit short of the needed bucks... we´ll see, what about Christmas maybe ;)

[copperpen]

Know what you mean about short of the needed bucks. I put this rig together before I retired. :birthday: Upgrades will now take a bit longer if needed.

[James Ross]

AlFitz, on 12 June 2013 - 12:27 PM, said:

Thought I'd post these screens up to illustrate the variance in the frame rates I encounter. I run OR in a 1600x900 window.

Am I right in thinking that the low framerates in the second instance are caused by the large number of winter tree textures and the strain of rendering or blending the alpha channel in them rather than the actual total of graphic calls by the program?

In your second picture, the data shows 14,896 (2,737 + 12,159) render primitives. This is a lot and will absolutely be bogging down the CPU on the Render process with all the draw calls.

Without a log file to see your current settings, it's a bit of a guessing game, but some of the things which contribute to render primitives are: dynamic shadows (early morning/late evening - i.e. long shadows - more than middle of the day/night), ShadowAllShapes (hidden setting), viewing distance, distant mountains. The location and route also have a significant effect, since the main overall contributor is the route creator. :sign_thanks:

[disc]

Yeah, mostly the route creators who are causing the performance problems. Most of the MSTS content developers, obsessed with the polygon count, while it's the least problem. What the real problem are: the texture file quantity, node quantity in models, and too many instances of objects, and the lack of the usage of bulk objects.

What is better for performance is to use bigger, but fewer textures(texture atlas (i think this could be automatized)), try to model objects with as few nodes and materials as it's possible, and make and use bulk objects on routes (for example tree line in one model, city parts in one model, forest objects, tree groups...

[disc]

The problem is probably all of those trees and clutter are individual objects and generating to many draw calls, while the terrain is nearly flat so it would be perfect for grouped objects.